mox/smtpserver/server_test.go

1901 lines
63 KiB
Go
Raw Normal View History

2023-01-30 16:27:06 +03:00
package smtpserver
// todo: test delivery with failing spf/dkim/dmarc
// todo: test delivering a message to multiple recipients, and with some of them failing.
import (
"bytes"
"context"
"crypto/ed25519"
cryptorand "crypto/rand"
"crypto/tls"
"crypto/x509"
"encoding/base64"
"errors"
"fmt"
"log/slog"
2023-01-30 16:27:06 +03:00
"math/big"
"mime/quotedprintable"
"net"
"os"
"path/filepath"
"sort"
2023-01-30 16:27:06 +03:00
"strings"
"testing"
"time"
"github.com/mjl-/bstore"
"github.com/mjl-/mox/config"
"github.com/mjl-/mox/dkim"
"github.com/mjl-/mox/dmarcdb"
"github.com/mjl-/mox/dns"
"github.com/mjl-/mox/mlog"
2023-01-30 16:27:06 +03:00
"github.com/mjl-/mox/mox-"
"github.com/mjl-/mox/queue"
new feature: when delivering messages from the queue, make it possible to use a "transport" the default transport is still just "direct delivery", where we connect to the destination domain's MX servers. other transports are: - regular smtp without authentication, this is relaying to a smarthost. - submission with authentication, e.g. to a third party email sending service. - direct delivery, but with with connections going through a socks proxy. this can be helpful if your ip is blocked, you need to get email out, and you have another IP that isn't blocked. keep in mind that for all of the above, appropriate SPF/DKIM settings have to be configured. the "dnscheck" for a domain does a check for any SOCKS IP in the SPF record. SPF for smtp/submission (ranges? includes?) and any DKIM requirements cannot really be checked. which transport is used can be configured through routes. routes can be set on an account, a domain, or globally. the routes are evaluated in that order, with the first match selecting the transport. these routes are evaluated for each delivery attempt. common selection criteria are recipient domain and sender domain, but also which delivery attempt this is. you could configured mox to attempt sending through a 3rd party from the 4th attempt onwards. routes and transports are optional. if no route matches, or an empty/zero transport is selected, normal direct delivery is done. we could already "submit" emails with 3rd party accounts with "sendmail". but we now support more SASL authentication mechanisms with SMTP (not only PLAIN, but also SCRAM-SHA-256, SCRAM-SHA-1 and CRAM-MD5), which sendmail now also supports. sendmail will use the most secure mechanism supported by the server, or the explicitly configured mechanism. for issue #36 by dmikushin. also based on earlier discussion on hackernews.
2023-06-16 19:38:28 +03:00
"github.com/mjl-/mox/sasl"
2023-01-30 16:27:06 +03:00
"github.com/mjl-/mox/smtp"
"github.com/mjl-/mox/smtpclient"
"github.com/mjl-/mox/store"
"github.com/mjl-/mox/subjectpass"
"github.com/mjl-/mox/tlsrptdb"
)
var ctxbg = context.Background()
func init() {
// Don't make tests slow.
badClientDelay = 0
authFailDelay = 0
unknownRecipientsDelay = 0
}
2023-01-30 16:27:06 +03:00
func tcheck(t *testing.T, err error, msg string) {
if err != nil {
t.Helper()
t.Fatalf("%s: %s", msg, err)
}
}
var submitMessage = strings.ReplaceAll(`From: <mjl@mox.example>
To: <remote@example.org>
Subject: test
Message-Id: <test@mox.example>
test email
`, "\n", "\r\n")
var deliverMessage = strings.ReplaceAll(`From: <remote@example.org>
To: <mjl@mox.example>
Subject: test
Message-Id: <test@example.org>
test email
`, "\n", "\r\n")
var deliverMessage2 = strings.ReplaceAll(`From: <remote@example.org>
To: <mjl@mox.example>
Subject: test
Message-Id: <test2@example.org>
test email, unique.
`, "\n", "\r\n")
2023-01-30 16:27:06 +03:00
type testserver struct {
t *testing.T
acc *store.Account
switchStop func()
2023-01-30 16:27:06 +03:00
comm *store.Comm
cid int64
resolver dns.Resolver
auth func(mechanisms []string, cs *tls.ConnectionState) (sasl.Client, error)
2023-01-30 16:27:06 +03:00
user, pass string
submission bool
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
requiretls bool
2023-01-30 16:27:06 +03:00
dnsbls []dns.Domain
tlsmode smtpclient.TLSMode
implement outgoing tls reports we were already accepting, processing and displaying incoming tls reports. now we start tracking TLS connection and security-policy-related errors for outgoing message deliveries as well. we send reports once a day, to the reporting addresses specified in TLSRPT records (rua) of a policy domain. these reports are about MTA-STS policies and/or DANE policies, and about STARTTLS-related failures. sending reports is enabled by default, but can be disabled through setting NoOutgoingTLSReports in mox.conf. only at the end of the implementation process came the realization that the TLSRPT policy domain for DANE (MX) hosts are separate from the TLSRPT policy for the recipient domain, and that MTA-STS and DANE TLS/policy results are typically delivered in separate reports. so MX hosts need their own TLSRPT policies. config for the per-host TLSRPT policy should be added to mox.conf for existing installs, in field HostTLSRPT. it is automatically configured by quickstart for new installs. with a HostTLSRPT config, the "dns records" and "dns check" admin pages now suggest the per-host TLSRPT record. by creating that record, you're requesting TLS reports about your MX host. gathering all the TLS/policy results is somewhat tricky. the tentacles go throughout the code. the positive result is that the TLS/policy-related code had to be cleaned up a bit. for example, the smtpclient TLS modes now reflect reality better, with independent settings about whether PKIX and/or DANE verification has to be done, and/or whether verification errors have to be ignored (e.g. for tls-required: no header). also, cached mtasts policies of mode "none" are now cleaned up once the MTA-STS DNS record goes away.
2023-11-09 19:40:46 +03:00
tlspkix bool
2023-01-30 16:27:06 +03:00
}
const password0 = "te\u0301st \u00a0\u2002\u200a" // NFD and various unicode spaces.
const password1 = "tést " // PRECIS normalized, with NFC.
2023-01-30 16:27:06 +03:00
func newTestServer(t *testing.T, configPath string, resolver dns.Resolver) *testserver {
limitersInit() // Reset rate limiters.
ts := testserver{t: t, cid: 1, resolver: resolver, tlsmode: smtpclient.TLSOpportunistic}
2023-01-30 16:27:06 +03:00
if dmarcdb.EvalDB != nil {
dmarcdb.EvalDB.Close()
dmarcdb.EvalDB = nil
}
log := mlog.New("smtpserver", nil)
mox.Context = ctxbg
2023-01-30 16:27:06 +03:00
mox.ConfigStaticPath = configPath
mox.MustLoadConfig(true, false)
2023-01-30 16:27:06 +03:00
dataDir := mox.ConfigDirPath(mox.Conf.Static.DataDir)
os.RemoveAll(dataDir)
add aliases/lists: when sending to an alias, the message gets delivered to all members the members must currently all be addresses of local accounts. a message sent to an alias is accepted if at least one of the members accepts it. if no members accepts it (e.g. due to bad reputation of sender), the message is rejected. if a message is submitted to both an alias addresses and to recipients that are members of the alias in an smtp transaction, the message will be delivered to such members only once. the same applies if the address in the message from-header is the address of a member: that member won't receive the message (they sent it). this prevents duplicate messages. aliases have three configuration options: - PostPublic: whether anyone can send through the alias, or only members. members-only lists can be useful inside organizations for internal communication. public lists can be useful for support addresses. - ListMembers: whether members can see the addresses of other members. this can be seen in the account web interface. in the future, we could export this in other ways, so clients can expand the list. - AllowMsgFrom: whether messages can be sent through the alias with the alias address used in the message from-header. the webmail knows it can use that address, and will use it as from-address when replying to a message sent to that address. ideas for the future: - allow external addresses as members. still with some restrictions, such as requiring a valid dkim-signature so delivery has a chance to succeed. will also need configuration of an admin that can receive any bounces. - allow specifying specific members who can sent through the list (instead of all members). for github issue #57 by hmfaysal. also relevant for #99 by naturalethic. thanks to damir & marin from sartura for discussing requirements/features.
2024-04-24 20:15:30 +03:00
2023-01-30 16:27:06 +03:00
var err error
ts.acc, err = store.OpenAccount(log, "mjl")
2023-01-30 16:27:06 +03:00
tcheck(t, err, "open account")
err = ts.acc.SetPassword(log, password0)
2023-01-30 16:27:06 +03:00
tcheck(t, err, "set password")
add aliases/lists: when sending to an alias, the message gets delivered to all members the members must currently all be addresses of local accounts. a message sent to an alias is accepted if at least one of the members accepts it. if no members accepts it (e.g. due to bad reputation of sender), the message is rejected. if a message is submitted to both an alias addresses and to recipients that are members of the alias in an smtp transaction, the message will be delivered to such members only once. the same applies if the address in the message from-header is the address of a member: that member won't receive the message (they sent it). this prevents duplicate messages. aliases have three configuration options: - PostPublic: whether anyone can send through the alias, or only members. members-only lists can be useful inside organizations for internal communication. public lists can be useful for support addresses. - ListMembers: whether members can see the addresses of other members. this can be seen in the account web interface. in the future, we could export this in other ways, so clients can expand the list. - AllowMsgFrom: whether messages can be sent through the alias with the alias address used in the message from-header. the webmail knows it can use that address, and will use it as from-address when replying to a message sent to that address. ideas for the future: - allow external addresses as members. still with some restrictions, such as requiring a valid dkim-signature so delivery has a chance to succeed. will also need configuration of an admin that can receive any bounces. - allow specifying specific members who can sent through the list (instead of all members). for github issue #57 by hmfaysal. also relevant for #99 by naturalethic. thanks to damir & marin from sartura for discussing requirements/features.
2024-04-24 20:15:30 +03:00
ts.switchStop = store.Switchboard()
2023-01-30 16:27:06 +03:00
err = queue.Init()
tcheck(t, err, "queue init")
ts.comm = store.RegisterComm(ts.acc)
return &ts
}
func (ts *testserver) close() {
add webmail it was far down on the roadmap, but implemented earlier, because it's interesting, and to help prepare for a jmap implementation. for jmap we need to implement more client-like functionality than with just imap. internal data structures need to change. jmap has lots of other requirements, so it's already a big project. by implementing a webmail now, some of the required data structure changes become clear and can be made now, so the later jmap implementation can do things similarly to the webmail code. the webmail frontend and webmail are written together, making their interface/api much smaller and simpler than jmap. one of the internal changes is that we now keep track of per-mailbox total/unread/unseen/deleted message counts and mailbox sizes. keeping this data consistent after any change to the stored messages (through the code base) is tricky, so mox now has a consistency check that verifies the counts are correct, which runs only during tests, each time an internal account reference is closed. we have a few more internal "changes" that are propagated for the webmail frontend (that imap doesn't have a way to propagate on a connection), like changes to the special-use flags on mailboxes, and used keywords in a mailbox. more changes that will be required have revealed themselves while implementing the webmail, and will be implemented next. the webmail user interface is modeled after the mail clients i use or have used: thunderbird, macos mail, mutt; and webmails i normally only use for testing: gmail, proton, yahoo, outlook. a somewhat technical user is assumed, but still the goal is to make this webmail client easy to use for everyone. the user interface looks like most other mail clients: a list of mailboxes, a search bar, a message list view, and message details. there is a top/bottom and a left/right layout for the list/message view, default is automatic based on screen size. the panes can be resized by the user. buttons for actions are just text, not icons. clicking a button briefly shows the shortcut for the action in the bottom right, helping with learning to operate quickly. any text that is underdotted has a title attribute that causes more information to be displayed, e.g. what a button does or a field is about. to highlight potential phishing attempts, any text (anywhere in the webclient) that switches unicode "blocks" (a rough approximation to (language) scripts) within a word is underlined orange. multiple messages can be selected with familiar ui interaction: clicking while holding control and/or shift keys. keyboard navigation works with arrows/page up/down and home/end keys, and also with a few basic vi-like keys for list/message navigation. we prefer showing the text instead of html (with inlined images only) version of a message. html messages are shown in an iframe served from an endpoint with CSP headers to prevent dangerous resources (scripts, external images) from being loaded. the html is also sanitized, with javascript removed. a user can choose to load external resources (e.g. images for tracking purposes). the frontend is just (strict) typescript, no external frameworks. all incoming/outgoing data is typechecked, both the api request parameters and response types, and the data coming in over SSE. the types and checking code are generated with sherpats, which uses the api definitions generated by sherpadoc based on the Go code. so types from the backend are automatically propagated to the frontend. since there is no framework to automatically propagate properties and rerender components, changes coming in over the SSE connection are propagated explicitly with regular function calls. the ui is separated into "views", each with a "root" dom element that is added to the visible document. these views have additional functions for getting changes propagated, often resulting in the view updating its (internal) ui state (dom). we keep the frontend compilation simple, it's just a few typescript files that get compiled (combined and types stripped) into a single js file, no additional runtime code needed or complicated build processes used. the webmail is served is served from a compressed, cachable html file that includes style and the javascript, currently just over 225kb uncompressed, under 60kb compressed (not minified, including comments). we include the generated js files in the repository, to keep Go's easily buildable self-contained binaries. authentication is basic http, as with the account and admin pages. most data comes in over one long-term SSE connection to the backend. api requests signal which mailbox/search/messages are requested over the SSE connection. fetching individual messages, and making changes, are done through api calls. the operations are similar to imap, so some code has been moved from package imapserver to package store. the future jmap implementation will benefit from these changes too. more functionality will probably be moved to the store package in the future. the quickstart enables webmail on the internal listener by default (for new installs). users can enable it on the public listener if they want to. mox localserve enables it too. to enable webmail on existing installs, add settings like the following to the listeners in mox.conf, similar to AccountHTTP(S): WebmailHTTP: Enabled: true WebmailHTTPS: Enabled: true special thanks to liesbeth, gerben, andrii for early user feedback. there is plenty still to do, see the list at the top of webmail/webmail.ts. feedback welcome as always.
2023-08-07 22:57:03 +03:00
if ts.acc == nil {
return
}
2023-01-30 16:27:06 +03:00
ts.comm.Unregister()
queue.Shutdown()
ts.switchStop()
add webmail it was far down on the roadmap, but implemented earlier, because it's interesting, and to help prepare for a jmap implementation. for jmap we need to implement more client-like functionality than with just imap. internal data structures need to change. jmap has lots of other requirements, so it's already a big project. by implementing a webmail now, some of the required data structure changes become clear and can be made now, so the later jmap implementation can do things similarly to the webmail code. the webmail frontend and webmail are written together, making their interface/api much smaller and simpler than jmap. one of the internal changes is that we now keep track of per-mailbox total/unread/unseen/deleted message counts and mailbox sizes. keeping this data consistent after any change to the stored messages (through the code base) is tricky, so mox now has a consistency check that verifies the counts are correct, which runs only during tests, each time an internal account reference is closed. we have a few more internal "changes" that are propagated for the webmail frontend (that imap doesn't have a way to propagate on a connection), like changes to the special-use flags on mailboxes, and used keywords in a mailbox. more changes that will be required have revealed themselves while implementing the webmail, and will be implemented next. the webmail user interface is modeled after the mail clients i use or have used: thunderbird, macos mail, mutt; and webmails i normally only use for testing: gmail, proton, yahoo, outlook. a somewhat technical user is assumed, but still the goal is to make this webmail client easy to use for everyone. the user interface looks like most other mail clients: a list of mailboxes, a search bar, a message list view, and message details. there is a top/bottom and a left/right layout for the list/message view, default is automatic based on screen size. the panes can be resized by the user. buttons for actions are just text, not icons. clicking a button briefly shows the shortcut for the action in the bottom right, helping with learning to operate quickly. any text that is underdotted has a title attribute that causes more information to be displayed, e.g. what a button does or a field is about. to highlight potential phishing attempts, any text (anywhere in the webclient) that switches unicode "blocks" (a rough approximation to (language) scripts) within a word is underlined orange. multiple messages can be selected with familiar ui interaction: clicking while holding control and/or shift keys. keyboard navigation works with arrows/page up/down and home/end keys, and also with a few basic vi-like keys for list/message navigation. we prefer showing the text instead of html (with inlined images only) version of a message. html messages are shown in an iframe served from an endpoint with CSP headers to prevent dangerous resources (scripts, external images) from being loaded. the html is also sanitized, with javascript removed. a user can choose to load external resources (e.g. images for tracking purposes). the frontend is just (strict) typescript, no external frameworks. all incoming/outgoing data is typechecked, both the api request parameters and response types, and the data coming in over SSE. the types and checking code are generated with sherpats, which uses the api definitions generated by sherpadoc based on the Go code. so types from the backend are automatically propagated to the frontend. since there is no framework to automatically propagate properties and rerender components, changes coming in over the SSE connection are propagated explicitly with regular function calls. the ui is separated into "views", each with a "root" dom element that is added to the visible document. these views have additional functions for getting changes propagated, often resulting in the view updating its (internal) ui state (dom). we keep the frontend compilation simple, it's just a few typescript files that get compiled (combined and types stripped) into a single js file, no additional runtime code needed or complicated build processes used. the webmail is served is served from a compressed, cachable html file that includes style and the javascript, currently just over 225kb uncompressed, under 60kb compressed (not minified, including comments). we include the generated js files in the repository, to keep Go's easily buildable self-contained binaries. authentication is basic http, as with the account and admin pages. most data comes in over one long-term SSE connection to the backend. api requests signal which mailbox/search/messages are requested over the SSE connection. fetching individual messages, and making changes, are done through api calls. the operations are similar to imap, so some code has been moved from package imapserver to package store. the future jmap implementation will benefit from these changes too. more functionality will probably be moved to the store package in the future. the quickstart enables webmail on the internal listener by default (for new installs). users can enable it on the public listener if they want to. mox localserve enables it too. to enable webmail on existing installs, add settings like the following to the listeners in mox.conf, similar to AccountHTTP(S): WebmailHTTP: Enabled: true WebmailHTTPS: Enabled: true special thanks to liesbeth, gerben, andrii for early user feedback. there is plenty still to do, see the list at the top of webmail/webmail.ts. feedback welcome as always.
2023-08-07 22:57:03 +03:00
err := ts.acc.Close()
tcheck(ts.t, err, "closing account")
ts.acc.CheckClosed()
add webmail it was far down on the roadmap, but implemented earlier, because it's interesting, and to help prepare for a jmap implementation. for jmap we need to implement more client-like functionality than with just imap. internal data structures need to change. jmap has lots of other requirements, so it's already a big project. by implementing a webmail now, some of the required data structure changes become clear and can be made now, so the later jmap implementation can do things similarly to the webmail code. the webmail frontend and webmail are written together, making their interface/api much smaller and simpler than jmap. one of the internal changes is that we now keep track of per-mailbox total/unread/unseen/deleted message counts and mailbox sizes. keeping this data consistent after any change to the stored messages (through the code base) is tricky, so mox now has a consistency check that verifies the counts are correct, which runs only during tests, each time an internal account reference is closed. we have a few more internal "changes" that are propagated for the webmail frontend (that imap doesn't have a way to propagate on a connection), like changes to the special-use flags on mailboxes, and used keywords in a mailbox. more changes that will be required have revealed themselves while implementing the webmail, and will be implemented next. the webmail user interface is modeled after the mail clients i use or have used: thunderbird, macos mail, mutt; and webmails i normally only use for testing: gmail, proton, yahoo, outlook. a somewhat technical user is assumed, but still the goal is to make this webmail client easy to use for everyone. the user interface looks like most other mail clients: a list of mailboxes, a search bar, a message list view, and message details. there is a top/bottom and a left/right layout for the list/message view, default is automatic based on screen size. the panes can be resized by the user. buttons for actions are just text, not icons. clicking a button briefly shows the shortcut for the action in the bottom right, helping with learning to operate quickly. any text that is underdotted has a title attribute that causes more information to be displayed, e.g. what a button does or a field is about. to highlight potential phishing attempts, any text (anywhere in the webclient) that switches unicode "blocks" (a rough approximation to (language) scripts) within a word is underlined orange. multiple messages can be selected with familiar ui interaction: clicking while holding control and/or shift keys. keyboard navigation works with arrows/page up/down and home/end keys, and also with a few basic vi-like keys for list/message navigation. we prefer showing the text instead of html (with inlined images only) version of a message. html messages are shown in an iframe served from an endpoint with CSP headers to prevent dangerous resources (scripts, external images) from being loaded. the html is also sanitized, with javascript removed. a user can choose to load external resources (e.g. images for tracking purposes). the frontend is just (strict) typescript, no external frameworks. all incoming/outgoing data is typechecked, both the api request parameters and response types, and the data coming in over SSE. the types and checking code are generated with sherpats, which uses the api definitions generated by sherpadoc based on the Go code. so types from the backend are automatically propagated to the frontend. since there is no framework to automatically propagate properties and rerender components, changes coming in over the SSE connection are propagated explicitly with regular function calls. the ui is separated into "views", each with a "root" dom element that is added to the visible document. these views have additional functions for getting changes propagated, often resulting in the view updating its (internal) ui state (dom). we keep the frontend compilation simple, it's just a few typescript files that get compiled (combined and types stripped) into a single js file, no additional runtime code needed or complicated build processes used. the webmail is served is served from a compressed, cachable html file that includes style and the javascript, currently just over 225kb uncompressed, under 60kb compressed (not minified, including comments). we include the generated js files in the repository, to keep Go's easily buildable self-contained binaries. authentication is basic http, as with the account and admin pages. most data comes in over one long-term SSE connection to the backend. api requests signal which mailbox/search/messages are requested over the SSE connection. fetching individual messages, and making changes, are done through api calls. the operations are similar to imap, so some code has been moved from package imapserver to package store. the future jmap implementation will benefit from these changes too. more functionality will probably be moved to the store package in the future. the quickstart enables webmail on the internal listener by default (for new installs). users can enable it on the public listener if they want to. mox localserve enables it too. to enable webmail on existing installs, add settings like the following to the listeners in mox.conf, similar to AccountHTTP(S): WebmailHTTP: Enabled: true WebmailHTTPS: Enabled: true special thanks to liesbeth, gerben, andrii for early user feedback. there is plenty still to do, see the list at the top of webmail/webmail.ts. feedback welcome as always.
2023-08-07 22:57:03 +03:00
ts.acc = nil
2023-01-30 16:27:06 +03:00
}
add aliases/lists: when sending to an alias, the message gets delivered to all members the members must currently all be addresses of local accounts. a message sent to an alias is accepted if at least one of the members accepts it. if no members accepts it (e.g. due to bad reputation of sender), the message is rejected. if a message is submitted to both an alias addresses and to recipients that are members of the alias in an smtp transaction, the message will be delivered to such members only once. the same applies if the address in the message from-header is the address of a member: that member won't receive the message (they sent it). this prevents duplicate messages. aliases have three configuration options: - PostPublic: whether anyone can send through the alias, or only members. members-only lists can be useful inside organizations for internal communication. public lists can be useful for support addresses. - ListMembers: whether members can see the addresses of other members. this can be seen in the account web interface. in the future, we could export this in other ways, so clients can expand the list. - AllowMsgFrom: whether messages can be sent through the alias with the alias address used in the message from-header. the webmail knows it can use that address, and will use it as from-address when replying to a message sent to that address. ideas for the future: - allow external addresses as members. still with some restrictions, such as requiring a valid dkim-signature so delivery has a chance to succeed. will also need configuration of an admin that can receive any bounces. - allow specifying specific members who can sent through the list (instead of all members). for github issue #57 by hmfaysal. also relevant for #99 by naturalethic. thanks to damir & marin from sartura for discussing requirements/features.
2024-04-24 20:15:30 +03:00
func (ts *testserver) checkCount(mailboxName string, expect int) {
t := ts.t
t.Helper()
q := bstore.QueryDB[store.Mailbox](ctxbg, ts.acc.DB)
q.FilterNonzero(store.Mailbox{Name: mailboxName})
mb, err := q.Get()
tcheck(t, err, "get mailbox")
qm := bstore.QueryDB[store.Message](ctxbg, ts.acc.DB)
qm.FilterNonzero(store.Message{MailboxID: mb.ID})
qm.FilterEqual("Expunged", false)
n, err := qm.Count()
tcheck(t, err, "count messages in mailbox")
if n != expect {
t.Fatalf("messages in mailbox, found %d, expected %d", n, expect)
}
}
2023-01-30 16:27:06 +03:00
func (ts *testserver) run(fn func(helloErr error, client *smtpclient.Client)) {
add a webapi and webhooks for a simple http/json-based api for applications to compose/send messages, receive delivery feedback, and maintain suppression lists. this is an alternative to applications using a library to compose messages, submitting those messages using smtp, and monitoring a mailbox with imap for DSNs, which can be processed into the equivalent of suppression lists. but you need to know about all these standards/protocols and find libraries. by using the webapi & webhooks, you just need a http & json library. unfortunately, there is no standard for these kinds of api, so mox has made up yet another one... matching incoming DSNs about deliveries to original outgoing messages requires keeping history of "retired" messages (delivered from the queue, either successfully or failed). this can be enabled per account. history is also useful for debugging deliveries. we now also keep history of each delivery attempt, accessible while still in the queue, and kept when a message is retired. the queue webadmin pages now also have pagination, to show potentially large history. a queue of webhook calls is now managed too. failures are retried similar to message deliveries. webhooks can also be saved to the retired list after completing. also configurable per account. messages can be sent with a "unique smtp mail from" address. this can only be used if the domain is configured with a localpart catchall separator such as "+". when enabled, a queued message gets assigned a random "fromid", which is added after the separator when sending. when DSNs are returned, they can be related to previously sent messages based on this fromid. in the future, we can implement matching on the "envid" used in the smtp dsn extension, or on the "message-id" of the message. using a fromid can be triggered by authenticating with a login email address that is configured as enabling fromid. suppression lists are automatically managed per account. if a delivery attempt results in certain smtp errors, the destination address is added to the suppression list. future messages queued for that recipient will immediately fail without a delivery attempt. suppression lists protect your mail server reputation. submitted messages can carry "extra" data through the queue and webhooks for outgoing deliveries. through webapi as a json object, through smtp submission as message headers of the form "x-mox-extra-<key>: value". to make it easy to test webapi/webhooks locally, the "localserve" mode actually puts messages in the queue. when it's time to deliver, it still won't do a full delivery attempt, but just delivers to the sender account. unless the recipient address has a special form, simulating a failure to deliver. admins now have more control over the queue. "hold rules" can be added to mark newly queued messages as "on hold", pausing delivery. rules can be about certain sender or recipient domains/addresses, or apply to all messages pausing the entire queue. also useful for (local) testing. new config options have been introduced. they are editable through the admin and/or account web interfaces. the webapi http endpoints are enabled for newly generated configs with the quickstart, and in localserve. existing configurations must explicitly enable the webapi in mox.conf. gopherwatch.org was created to dogfood this code. it initially used just the compose/smtpclient/imapclient mox packages to send messages and process delivery feedback. it will get a config option to use the mox webapi/webhooks instead. the gopherwatch code to use webapi/webhook is smaller and simpler, and developing that shaped development of the mox webapi/webhooks. for issue #31 by cuu508
2024-04-15 22:49:02 +03:00
ts.t.Helper()
ts.runRaw(func(conn net.Conn) {
ts.t.Helper()
auth := ts.auth
if auth == nil && ts.user != "" {
auth = func(mechanisms []string, cs *tls.ConnectionState) (sasl.Client, error) {
return sasl.NewClientPlain(ts.user, ts.pass), nil
}
}
ourHostname := mox.Conf.Static.HostnameDomain
remoteHostname := dns.Domain{ASCII: "mox.example"}
opts := smtpclient.Opts{
Auth: auth,
RootCAs: mox.Conf.Static.TLS.CertPool,
}
log := pkglog.WithCid(ts.cid - 1)
client, err := smtpclient.New(ctxbg, log.Logger, conn, ts.tlsmode, ts.tlspkix, ourHostname, remoteHostname, opts)
if err != nil {
conn.Close()
} else {
defer client.Close()
}
fn(err, client)
})
}
func (ts *testserver) runRaw(fn func(clientConn net.Conn)) {
2023-01-30 16:27:06 +03:00
ts.t.Helper()
ts.cid += 2
serverConn, clientConn := net.Pipe()
defer serverConn.Close()
// clientConn is closed as part of closing client.
serverdone := make(chan struct{})
defer func() { <-serverdone }()
go func() {
tlsConfig := &tls.Config{
Certificates: []tls.Certificate{fakeCert(ts.t)},
}
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
serve("test", ts.cid-2, dns.Domain{ASCII: "mox.example"}, tlsConfig, serverConn, ts.resolver, ts.submission, false, 100<<20, false, false, ts.requiretls, ts.dnsbls, 0)
2023-01-30 16:27:06 +03:00
close(serverdone)
}()
fn(clientConn)
2023-01-30 16:27:06 +03:00
}
func (ts *testserver) smtpErr(err error, expErr *smtpclient.Error) {
t := ts.t
t.Helper()
add aliases/lists: when sending to an alias, the message gets delivered to all members the members must currently all be addresses of local accounts. a message sent to an alias is accepted if at least one of the members accepts it. if no members accepts it (e.g. due to bad reputation of sender), the message is rejected. if a message is submitted to both an alias addresses and to recipients that are members of the alias in an smtp transaction, the message will be delivered to such members only once. the same applies if the address in the message from-header is the address of a member: that member won't receive the message (they sent it). this prevents duplicate messages. aliases have three configuration options: - PostPublic: whether anyone can send through the alias, or only members. members-only lists can be useful inside organizations for internal communication. public lists can be useful for support addresses. - ListMembers: whether members can see the addresses of other members. this can be seen in the account web interface. in the future, we could export this in other ways, so clients can expand the list. - AllowMsgFrom: whether messages can be sent through the alias with the alias address used in the message from-header. the webmail knows it can use that address, and will use it as from-address when replying to a message sent to that address. ideas for the future: - allow external addresses as members. still with some restrictions, such as requiring a valid dkim-signature so delivery has a chance to succeed. will also need configuration of an admin that can receive any bounces. - allow specifying specific members who can sent through the list (instead of all members). for github issue #57 by hmfaysal. also relevant for #99 by naturalethic. thanks to damir & marin from sartura for discussing requirements/features.
2024-04-24 20:15:30 +03:00
var cerr smtpclient.Error
if expErr == nil && err != nil || expErr != nil && (err == nil || !errors.As(err, &cerr) || cerr.Permanent != expErr.Permanent || cerr.Code != expErr.Code || cerr.Secode != expErr.Secode) {
t.Fatalf("got err:\n%#v (%q)\nexpected:\n%#v", err, err, expErr)
add aliases/lists: when sending to an alias, the message gets delivered to all members the members must currently all be addresses of local accounts. a message sent to an alias is accepted if at least one of the members accepts it. if no members accepts it (e.g. due to bad reputation of sender), the message is rejected. if a message is submitted to both an alias addresses and to recipients that are members of the alias in an smtp transaction, the message will be delivered to such members only once. the same applies if the address in the message from-header is the address of a member: that member won't receive the message (they sent it). this prevents duplicate messages. aliases have three configuration options: - PostPublic: whether anyone can send through the alias, or only members. members-only lists can be useful inside organizations for internal communication. public lists can be useful for support addresses. - ListMembers: whether members can see the addresses of other members. this can be seen in the account web interface. in the future, we could export this in other ways, so clients can expand the list. - AllowMsgFrom: whether messages can be sent through the alias with the alias address used in the message from-header. the webmail knows it can use that address, and will use it as from-address when replying to a message sent to that address. ideas for the future: - allow external addresses as members. still with some restrictions, such as requiring a valid dkim-signature so delivery has a chance to succeed. will also need configuration of an admin that can receive any bounces. - allow specifying specific members who can sent through the list (instead of all members). for github issue #57 by hmfaysal. also relevant for #99 by naturalethic. thanks to damir & marin from sartura for discussing requirements/features.
2024-04-24 20:15:30 +03:00
}
}
2023-01-30 16:27:06 +03:00
// Just a cert that appears valid. SMTP client will not verify anything about it
// (that is opportunistic TLS for you, "better some than none"). Let's enjoy this
// one moment where it makes life easier.
func fakeCert(t *testing.T) tls.Certificate {
privKey := ed25519.NewKeyFromSeed(make([]byte, ed25519.SeedSize)) // Fake key, don't use this for real!
template := &x509.Certificate{
SerialNumber: big.NewInt(1), // Required field...
}
localCertBuf, err := x509.CreateCertificate(cryptorand.Reader, template, template, privKey.Public(), privKey)
if err != nil {
t.Fatalf("making certificate: %s", err)
}
cert, err := x509.ParseCertificate(localCertBuf)
if err != nil {
t.Fatalf("parsing generated certificate: %s", err)
}
c := tls.Certificate{
Certificate: [][]byte{localCertBuf},
PrivateKey: privKey,
Leaf: cert,
}
return c
}
// check expected dmarc evaluations for outgoing aggregate reports.
func checkEvaluationCount(t *testing.T, n int) []dmarcdb.Evaluation {
t.Helper()
l, err := dmarcdb.Evaluations(ctxbg)
tcheck(t, err, "get dmarc evaluations")
tcompare(t, len(l), n)
return l
}
2023-01-30 16:27:06 +03:00
// Test submission from authenticated user.
func TestSubmission(t *testing.T) {
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/mox.conf"), dns.MockResolver{})
2023-01-30 16:27:06 +03:00
defer ts.close()
// Set DKIM signing config.
dom, _ := mox.Conf.Domain(dns.Domain{ASCII: "mox.example"})
sel := config.Selector{
HashEffective: "sha256",
HeadersEffective: []string{"From", "To", "Subject"},
Key: ed25519.NewKeyFromSeed(make([]byte, ed25519.SeedSize)), // Fake key, don't use for real.
Domain: dns.Domain{ASCII: "mox.example"},
}
dom.DKIM = config.DKIM{
Selectors: map[string]config.Selector{"testsel": sel},
Sign: []string{"testsel"},
}
mox.Conf.Dynamic.Domains["mox.example"] = dom
testAuth := func(authfn func(user, pass string, cs *tls.ConnectionState) sasl.Client, user, pass string, expErr *smtpclient.Error) {
2023-01-30 16:27:06 +03:00
t.Helper()
new feature: when delivering messages from the queue, make it possible to use a "transport" the default transport is still just "direct delivery", where we connect to the destination domain's MX servers. other transports are: - regular smtp without authentication, this is relaying to a smarthost. - submission with authentication, e.g. to a third party email sending service. - direct delivery, but with with connections going through a socks proxy. this can be helpful if your ip is blocked, you need to get email out, and you have another IP that isn't blocked. keep in mind that for all of the above, appropriate SPF/DKIM settings have to be configured. the "dnscheck" for a domain does a check for any SOCKS IP in the SPF record. SPF for smtp/submission (ranges? includes?) and any DKIM requirements cannot really be checked. which transport is used can be configured through routes. routes can be set on an account, a domain, or globally. the routes are evaluated in that order, with the first match selecting the transport. these routes are evaluated for each delivery attempt. common selection criteria are recipient domain and sender domain, but also which delivery attempt this is. you could configured mox to attempt sending through a 3rd party from the 4th attempt onwards. routes and transports are optional. if no route matches, or an empty/zero transport is selected, normal direct delivery is done. we could already "submit" emails with 3rd party accounts with "sendmail". but we now support more SASL authentication mechanisms with SMTP (not only PLAIN, but also SCRAM-SHA-256, SCRAM-SHA-1 and CRAM-MD5), which sendmail now also supports. sendmail will use the most secure mechanism supported by the server, or the explicitly configured mechanism. for issue #36 by dmikushin. also based on earlier discussion on hackernews.
2023-06-16 19:38:28 +03:00
if authfn != nil {
ts.auth = func(mechanisms []string, cs *tls.ConnectionState) (sasl.Client, error) {
return authfn(user, pass, cs), nil
}
new feature: when delivering messages from the queue, make it possible to use a "transport" the default transport is still just "direct delivery", where we connect to the destination domain's MX servers. other transports are: - regular smtp without authentication, this is relaying to a smarthost. - submission with authentication, e.g. to a third party email sending service. - direct delivery, but with with connections going through a socks proxy. this can be helpful if your ip is blocked, you need to get email out, and you have another IP that isn't blocked. keep in mind that for all of the above, appropriate SPF/DKIM settings have to be configured. the "dnscheck" for a domain does a check for any SOCKS IP in the SPF record. SPF for smtp/submission (ranges? includes?) and any DKIM requirements cannot really be checked. which transport is used can be configured through routes. routes can be set on an account, a domain, or globally. the routes are evaluated in that order, with the first match selecting the transport. these routes are evaluated for each delivery attempt. common selection criteria are recipient domain and sender domain, but also which delivery attempt this is. you could configured mox to attempt sending through a 3rd party from the 4th attempt onwards. routes and transports are optional. if no route matches, or an empty/zero transport is selected, normal direct delivery is done. we could already "submit" emails with 3rd party accounts with "sendmail". but we now support more SASL authentication mechanisms with SMTP (not only PLAIN, but also SCRAM-SHA-256, SCRAM-SHA-1 and CRAM-MD5), which sendmail now also supports. sendmail will use the most secure mechanism supported by the server, or the explicitly configured mechanism. for issue #36 by dmikushin. also based on earlier discussion on hackernews.
2023-06-16 19:38:28 +03:00
} else {
ts.auth = nil
}
2023-01-30 16:27:06 +03:00
ts.run(func(err error, client *smtpclient.Client) {
t.Helper()
mailFrom := "mjl@mox.example"
rcptTo := "remote@example.org"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(submitMessage)), strings.NewReader(submitMessage), false, false, false)
2023-01-30 16:27:06 +03:00
}
var cerr smtpclient.Error
if expErr == nil && err != nil || expErr != nil && (err == nil || !errors.As(err, &cerr) || cerr.Code != expErr.Code || cerr.Secode != expErr.Secode) {
t.Fatalf("got err:\n%#v (%q)\nexpected:\n%#v", err, err, expErr)
2023-01-30 16:27:06 +03:00
}
checkEvaluationCount(t, 0)
2023-01-30 16:27:06 +03:00
})
}
ts.submission = true
new feature: when delivering messages from the queue, make it possible to use a "transport" the default transport is still just "direct delivery", where we connect to the destination domain's MX servers. other transports are: - regular smtp without authentication, this is relaying to a smarthost. - submission with authentication, e.g. to a third party email sending service. - direct delivery, but with with connections going through a socks proxy. this can be helpful if your ip is blocked, you need to get email out, and you have another IP that isn't blocked. keep in mind that for all of the above, appropriate SPF/DKIM settings have to be configured. the "dnscheck" for a domain does a check for any SOCKS IP in the SPF record. SPF for smtp/submission (ranges? includes?) and any DKIM requirements cannot really be checked. which transport is used can be configured through routes. routes can be set on an account, a domain, or globally. the routes are evaluated in that order, with the first match selecting the transport. these routes are evaluated for each delivery attempt. common selection criteria are recipient domain and sender domain, but also which delivery attempt this is. you could configured mox to attempt sending through a 3rd party from the 4th attempt onwards. routes and transports are optional. if no route matches, or an empty/zero transport is selected, normal direct delivery is done. we could already "submit" emails with 3rd party accounts with "sendmail". but we now support more SASL authentication mechanisms with SMTP (not only PLAIN, but also SCRAM-SHA-256, SCRAM-SHA-1 and CRAM-MD5), which sendmail now also supports. sendmail will use the most secure mechanism supported by the server, or the explicitly configured mechanism. for issue #36 by dmikushin. also based on earlier discussion on hackernews.
2023-06-16 19:38:28 +03:00
testAuth(nil, "", "", &smtpclient.Error{Permanent: true, Code: smtp.C530SecurityRequired, Secode: smtp.SePol7Other0})
authfns := []func(user, pass string, cs *tls.ConnectionState) sasl.Client{
func(user, pass string, cs *tls.ConnectionState) sasl.Client { return sasl.NewClientPlain(user, pass) },
func(user, pass string, cs *tls.ConnectionState) sasl.Client { return sasl.NewClientLogin(user, pass) },
func(user, pass string, cs *tls.ConnectionState) sasl.Client { return sasl.NewClientCRAMMD5(user, pass) },
func(user, pass string, cs *tls.ConnectionState) sasl.Client {
return sasl.NewClientSCRAMSHA1(user, pass, false)
},
func(user, pass string, cs *tls.ConnectionState) sasl.Client {
return sasl.NewClientSCRAMSHA256(user, pass, false)
},
func(user, pass string, cs *tls.ConnectionState) sasl.Client {
return sasl.NewClientSCRAMSHA1PLUS(user, pass, *cs)
},
func(user, pass string, cs *tls.ConnectionState) sasl.Client {
return sasl.NewClientSCRAMSHA256PLUS(user, pass, *cs)
},
new feature: when delivering messages from the queue, make it possible to use a "transport" the default transport is still just "direct delivery", where we connect to the destination domain's MX servers. other transports are: - regular smtp without authentication, this is relaying to a smarthost. - submission with authentication, e.g. to a third party email sending service. - direct delivery, but with with connections going through a socks proxy. this can be helpful if your ip is blocked, you need to get email out, and you have another IP that isn't blocked. keep in mind that for all of the above, appropriate SPF/DKIM settings have to be configured. the "dnscheck" for a domain does a check for any SOCKS IP in the SPF record. SPF for smtp/submission (ranges? includes?) and any DKIM requirements cannot really be checked. which transport is used can be configured through routes. routes can be set on an account, a domain, or globally. the routes are evaluated in that order, with the first match selecting the transport. these routes are evaluated for each delivery attempt. common selection criteria are recipient domain and sender domain, but also which delivery attempt this is. you could configured mox to attempt sending through a 3rd party from the 4th attempt onwards. routes and transports are optional. if no route matches, or an empty/zero transport is selected, normal direct delivery is done. we could already "submit" emails with 3rd party accounts with "sendmail". but we now support more SASL authentication mechanisms with SMTP (not only PLAIN, but also SCRAM-SHA-256, SCRAM-SHA-1 and CRAM-MD5), which sendmail now also supports. sendmail will use the most secure mechanism supported by the server, or the explicitly configured mechanism. for issue #36 by dmikushin. also based on earlier discussion on hackernews.
2023-06-16 19:38:28 +03:00
}
for _, fn := range authfns {
testAuth(fn, "mjl@mox.example", "test", &smtpclient.Error{Code: smtp.C535AuthBadCreds, Secode: smtp.SePol7AuthBadCreds8}) // Bad (short) password.
testAuth(fn, "mjl@mox.example", password0+"test", &smtpclient.Error{Code: smtp.C535AuthBadCreds, Secode: smtp.SePol7AuthBadCreds8}) // Bad password.
testAuth(fn, "mjl@mox.example", password0, nil)
testAuth(fn, "mjl@mox.example", password1, nil)
testAuth(fn, "móx@mox.example", password0, nil)
testAuth(fn, "móx@mox.example", password1, nil)
testAuth(fn, "mo\u0301x@mox.example", password0, nil)
testAuth(fn, "mo\u0301x@mox.example", password1, nil)
new feature: when delivering messages from the queue, make it possible to use a "transport" the default transport is still just "direct delivery", where we connect to the destination domain's MX servers. other transports are: - regular smtp without authentication, this is relaying to a smarthost. - submission with authentication, e.g. to a third party email sending service. - direct delivery, but with with connections going through a socks proxy. this can be helpful if your ip is blocked, you need to get email out, and you have another IP that isn't blocked. keep in mind that for all of the above, appropriate SPF/DKIM settings have to be configured. the "dnscheck" for a domain does a check for any SOCKS IP in the SPF record. SPF for smtp/submission (ranges? includes?) and any DKIM requirements cannot really be checked. which transport is used can be configured through routes. routes can be set on an account, a domain, or globally. the routes are evaluated in that order, with the first match selecting the transport. these routes are evaluated for each delivery attempt. common selection criteria are recipient domain and sender domain, but also which delivery attempt this is. you could configured mox to attempt sending through a 3rd party from the 4th attempt onwards. routes and transports are optional. if no route matches, or an empty/zero transport is selected, normal direct delivery is done. we could already "submit" emails with 3rd party accounts with "sendmail". but we now support more SASL authentication mechanisms with SMTP (not only PLAIN, but also SCRAM-SHA-256, SCRAM-SHA-1 and CRAM-MD5), which sendmail now also supports. sendmail will use the most secure mechanism supported by the server, or the explicitly configured mechanism. for issue #36 by dmikushin. also based on earlier discussion on hackernews.
2023-06-16 19:38:28 +03:00
}
2023-01-30 16:27:06 +03:00
}
// Test delivery from external MTA.
func TestDelivery(t *testing.T) {
resolver := dns.MockResolver{
A: map[string][]string{
"example.org.": {"127.0.0.10"}, // For mx check.
},
PTR: map[string][]string{},
}
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/mox.conf"), resolver)
2023-01-30 16:27:06 +03:00
defer ts.close()
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@127.0.0.10"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
2023-01-30 16:27:06 +03:00
}
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C550MailboxUnavail {
t.Fatalf("deliver to ip address, got err %v, expected smtpclient.Error with code %d", err, smtp.C550MailboxUnavail)
}
})
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@test.example" // Not configured as destination.
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
2023-01-30 16:27:06 +03:00
}
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C550MailboxUnavail {
t.Fatalf("deliver to unknown domain, got err %v, expected smtpclient.Error with code %d", err, smtp.C550MailboxUnavail)
}
})
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "unknown@mox.example" // User unknown.
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
2023-01-30 16:27:06 +03:00
}
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C550MailboxUnavail {
t.Fatalf("deliver to unknown user for known domain, got err %v, expected smtpclient.Error with code %d", err, smtp.C550MailboxUnavail)
}
})
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@mox.example"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
2023-01-30 16:27:06 +03:00
}
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C451LocalErr {
t.Fatalf("deliver from user without reputation, valid iprev required, got err %v, expected smtpclient.Error with code %d", err, smtp.C451LocalErr)
}
})
// Set up iprev to get delivery from unknown user to be accepted.
resolver.PTR["127.0.0.10"] = []string{"example.org."}
// Only ascii o@ is configured, not the greek and cyrillic lookalikes.
2023-01-30 16:27:06 +03:00
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "ο@mox.example" // omicron \u03bf, looks like the configured o@
msg := strings.ReplaceAll(deliverMessage, "mjl@mox.example", rcptTo)
2023-01-30 16:27:06 +03:00
if err == nil {
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(msg)), strings.NewReader(msg), false, true, false)
}
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C550MailboxUnavail {
t.Fatalf("deliver to omicron @ instead of ascii o @, got err %v, expected smtpclient.Error with code %d", err, smtp.C550MailboxUnavail)
}
})
ts.run(func(err error, client *smtpclient.Client) {
recipients := []string{
"mjl@mox.example",
"o@mox.example", // ascii o, as configured
"\u2126@mox.example", // ohm sign, as configured
"ω@mox.example", // lower-case omega, we match case-insensitively and this is the lowercase of ohm (!)
"\u03a9@mox.example", // capital omega, also lowercased to omega.
"móx@mox.example", // NFC
"mo\u0301x@mox.example", // not NFC, but normalized as móx@, see https://go.dev/blog/normalization
2023-01-30 16:27:06 +03:00
}
for _, rcptTo := range recipients {
// Ensure SMTP RCPT TO and message address headers are the same, otherwise the junk
// filter treats us more strictly.
msg := strings.ReplaceAll(deliverMessage, "mjl@mox.example", rcptTo)
mailFrom := "remote@example.org"
if err == nil {
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(msg)), strings.NewReader(msg), false, true, false)
}
tcheck(t, err, "deliver to remote")
changes := make(chan []store.Change)
go func() {
changes <- ts.comm.Get()
}()
timer := time.NewTimer(time.Second)
defer timer.Stop()
select {
case <-changes:
case <-timer.C:
t.Fatalf("no delivery in 1s")
}
2023-01-30 16:27:06 +03:00
}
})
checkEvaluationCount(t, 0)
2023-01-30 16:27:06 +03:00
}
func tinsertmsg(t *testing.T, acc *store.Account, mailbox string, m *store.Message, msg string) {
mf, err := store.CreateMessageTemp(pkglog, "queue-dsn")
2023-01-30 16:27:06 +03:00
tcheck(t, err, "temp message")
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
defer os.Remove(mf.Name())
defer mf.Close()
2023-01-30 16:27:06 +03:00
_, err = mf.Write([]byte(msg))
tcheck(t, err, "write message")
err = acc.DeliverMailbox(pkglog, mailbox, m, mf)
2023-01-30 16:27:06 +03:00
tcheck(t, err, "deliver message")
err = mf.Close()
tcheck(t, err, "close message")
}
func tretrain(t *testing.T, acc *store.Account) {
t.Helper()
// Fresh empty junkfilter.
basePath := mox.DataDirPath("accounts")
dbPath := filepath.Join(basePath, acc.Name, "junkfilter.db")
bloomPath := filepath.Join(basePath, acc.Name, "junkfilter.bloom")
os.Remove(dbPath)
os.Remove(bloomPath)
jf, _, err := acc.OpenJunkFilter(ctxbg, pkglog)
2023-01-30 16:27:06 +03:00
tcheck(t, err, "open junk filter")
defer jf.Close()
// Fetch messags to retrain on.
q := bstore.QueryDB[store.Message](ctxbg, acc.DB)
q.FilterEqual("Expunged", false)
2023-01-30 16:27:06 +03:00
q.FilterFn(func(m store.Message) bool {
return m.Flags.Junk || m.Flags.Notjunk
})
msgs, err := q.List()
tcheck(t, err, "fetch messages")
// Retrain the messages.
for _, m := range msgs {
ham := m.Flags.Notjunk
f, err := os.Open(acc.MessagePath(m.ID))
tcheck(t, err, "open message")
r := store.FileMsgReader(m.MsgPrefix, f)
jf.TrainMessage(ctxbg, r, m.Size, ham)
2023-01-30 16:27:06 +03:00
err = r.Close()
tcheck(t, err, "close message")
}
err = jf.Save()
tcheck(t, err, "save junkfilter")
}
// Test accept/reject with DMARC reputation and with spammy content.
func TestSpam(t *testing.T) {
resolver := &dns.MockResolver{
A: map[string][]string{
"example.org.": {"127.0.0.1"}, // For mx check.
},
TXT: map[string][]string{
"example.org.": {"v=spf1 ip4:127.0.0.10 -all"},
"_dmarc.example.org.": {"v=DMARC1;p=reject; rua=mailto:dmarcrpt@example.org"},
2023-01-30 16:27:06 +03:00
},
}
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/junk/mox.conf"), resolver)
2023-01-30 16:27:06 +03:00
defer ts.close()
// Insert spammy messages. No junkfilter training yet.
m := store.Message{
RemoteIP: "127.0.0.10",
RemoteIPMasked1: "127.0.0.10",
RemoteIPMasked2: "127.0.0.0",
RemoteIPMasked3: "127.0.0.0",
MailFrom: "remote@example.org",
MailFromLocalpart: smtp.Localpart("remote"),
MailFromDomain: "example.org",
RcptToLocalpart: smtp.Localpart("mjl"),
RcptToDomain: "mox.example",
MsgFromLocalpart: smtp.Localpart("remote"),
MsgFromDomain: "example.org",
MsgFromOrgDomain: "example.org",
MsgFromValidated: true,
MsgFromValidation: store.ValidationStrict,
Flags: store.Flags{Seen: true, Junk: true},
Size: int64(len(deliverMessage)),
2023-01-30 16:27:06 +03:00
}
for i := 0; i < 3; i++ {
nm := m
tinsertmsg(t, ts.acc, "Inbox", &nm, deliverMessage)
}
// Delivery from sender with bad reputation should fail.
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@mox.example"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
2023-01-30 16:27:06 +03:00
}
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C451LocalErr {
t.Fatalf("delivery by bad sender, got err %v, expected smtpclient.Error with code %d", err, smtp.C451LocalErr)
}
add aliases/lists: when sending to an alias, the message gets delivered to all members the members must currently all be addresses of local accounts. a message sent to an alias is accepted if at least one of the members accepts it. if no members accepts it (e.g. due to bad reputation of sender), the message is rejected. if a message is submitted to both an alias addresses and to recipients that are members of the alias in an smtp transaction, the message will be delivered to such members only once. the same applies if the address in the message from-header is the address of a member: that member won't receive the message (they sent it). this prevents duplicate messages. aliases have three configuration options: - PostPublic: whether anyone can send through the alias, or only members. members-only lists can be useful inside organizations for internal communication. public lists can be useful for support addresses. - ListMembers: whether members can see the addresses of other members. this can be seen in the account web interface. in the future, we could export this in other ways, so clients can expand the list. - AllowMsgFrom: whether messages can be sent through the alias with the alias address used in the message from-header. the webmail knows it can use that address, and will use it as from-address when replying to a message sent to that address. ideas for the future: - allow external addresses as members. still with some restrictions, such as requiring a valid dkim-signature so delivery has a chance to succeed. will also need configuration of an admin that can receive any bounces. - allow specifying specific members who can sent through the list (instead of all members). for github issue #57 by hmfaysal. also relevant for #99 by naturalethic. thanks to damir & marin from sartura for discussing requirements/features.
2024-04-24 20:15:30 +03:00
ts.checkCount("Rejects", 1)
checkEvaluationCount(t, 0) // No positive interactions yet.
})
// Delivery from sender with bad reputation matching AcceptRejectsToMailbox should
// result in accepted delivery to the mailbox.
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl2@mox.example"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage2)), strings.NewReader(deliverMessage2), false, false, false)
}
tcheck(t, err, "deliver")
add aliases/lists: when sending to an alias, the message gets delivered to all members the members must currently all be addresses of local accounts. a message sent to an alias is accepted if at least one of the members accepts it. if no members accepts it (e.g. due to bad reputation of sender), the message is rejected. if a message is submitted to both an alias addresses and to recipients that are members of the alias in an smtp transaction, the message will be delivered to such members only once. the same applies if the address in the message from-header is the address of a member: that member won't receive the message (they sent it). this prevents duplicate messages. aliases have three configuration options: - PostPublic: whether anyone can send through the alias, or only members. members-only lists can be useful inside organizations for internal communication. public lists can be useful for support addresses. - ListMembers: whether members can see the addresses of other members. this can be seen in the account web interface. in the future, we could export this in other ways, so clients can expand the list. - AllowMsgFrom: whether messages can be sent through the alias with the alias address used in the message from-header. the webmail knows it can use that address, and will use it as from-address when replying to a message sent to that address. ideas for the future: - allow external addresses as members. still with some restrictions, such as requiring a valid dkim-signature so delivery has a chance to succeed. will also need configuration of an admin that can receive any bounces. - allow specifying specific members who can sent through the list (instead of all members). for github issue #57 by hmfaysal. also relevant for #99 by naturalethic. thanks to damir & marin from sartura for discussing requirements/features.
2024-04-24 20:15:30 +03:00
ts.checkCount("mjl2junk", 1) // In ruleset rejects mailbox.
ts.checkCount("Rejects", 1) // Same as before.
checkEvaluationCount(t, 0) // This is not an actual accept.
2023-01-30 16:27:06 +03:00
})
// Mark the messages as having good reputation.
q := bstore.QueryDB[store.Message](ctxbg, ts.acc.DB)
q.FilterEqual("Expunged", false)
2023-01-30 16:27:06 +03:00
_, err := q.UpdateFields(map[string]any{"Junk": false, "Notjunk": true})
tcheck(t, err, "update junkiness")
// Message should now be accepted.
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@mox.example"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
2023-01-30 16:27:06 +03:00
}
tcheck(t, err, "deliver")
// Message should now be removed from Rejects mailboxes.
add aliases/lists: when sending to an alias, the message gets delivered to all members the members must currently all be addresses of local accounts. a message sent to an alias is accepted if at least one of the members accepts it. if no members accepts it (e.g. due to bad reputation of sender), the message is rejected. if a message is submitted to both an alias addresses and to recipients that are members of the alias in an smtp transaction, the message will be delivered to such members only once. the same applies if the address in the message from-header is the address of a member: that member won't receive the message (they sent it). this prevents duplicate messages. aliases have three configuration options: - PostPublic: whether anyone can send through the alias, or only members. members-only lists can be useful inside organizations for internal communication. public lists can be useful for support addresses. - ListMembers: whether members can see the addresses of other members. this can be seen in the account web interface. in the future, we could export this in other ways, so clients can expand the list. - AllowMsgFrom: whether messages can be sent through the alias with the alias address used in the message from-header. the webmail knows it can use that address, and will use it as from-address when replying to a message sent to that address. ideas for the future: - allow external addresses as members. still with some restrictions, such as requiring a valid dkim-signature so delivery has a chance to succeed. will also need configuration of an admin that can receive any bounces. - allow specifying specific members who can sent through the list (instead of all members). for github issue #57 by hmfaysal. also relevant for #99 by naturalethic. thanks to damir & marin from sartura for discussing requirements/features.
2024-04-24 20:15:30 +03:00
ts.checkCount("Rejects", 0)
ts.checkCount("mjl2junk", 1)
checkEvaluationCount(t, 1)
2023-01-30 16:27:06 +03:00
})
// Undo dmarc pass, mark messages as junk, and train the filter.
resolver.TXT = nil
q = bstore.QueryDB[store.Message](ctxbg, ts.acc.DB)
q.FilterEqual("Expunged", false)
2023-01-30 16:27:06 +03:00
_, err = q.UpdateFields(map[string]any{"Junk": true, "Notjunk": false})
tcheck(t, err, "update junkiness")
tretrain(t, ts.acc)
// Message should be refused for spammy content.
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@mox.example"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
2023-01-30 16:27:06 +03:00
}
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C451LocalErr {
t.Fatalf("attempt to deliver spamy message, got err %v, expected smtpclient.Error with code %d", err, smtp.C451LocalErr)
}
checkEvaluationCount(t, 1) // No new evaluation, this isn't a DMARC reject.
2023-01-30 16:27:06 +03:00
})
}
// Test accept/reject with forwarded messages, DMARC ignored, no IP/EHLO/MAIL
// FROM-based reputation.
func TestForward(t *testing.T) {
// Do a run without forwarding, and with.
check := func(forward bool) {
resolver := &dns.MockResolver{
A: map[string][]string{
"bad.example.": {"127.0.0.1"}, // For mx check.
"good.example.": {"127.0.0.1"}, // For mx check.
"forward.example.": {"127.0.0.10"}, // For mx check.
},
TXT: map[string][]string{
"bad.example.": {"v=spf1 ip4:127.0.0.1 -all"},
"good.example.": {"v=spf1 ip4:127.0.0.1 -all"},
"forward.example.": {"v=spf1 ip4:127.0.0.10 -all"},
"_dmarc.bad.example.": {"v=DMARC1;p=reject; rua=mailto:dmarc@bad.example"},
"_dmarc.good.example.": {"v=DMARC1;p=reject; rua=mailto:dmarc@good.example"},
"_dmarc.forward.example.": {"v=DMARC1;p=reject; rua=mailto:dmarc@forward.example"},
},
PTR: map[string][]string{
"127.0.0.10": {"forward.example."}, // For iprev check.
},
}
rcptTo := "mjl3@mox.example"
if !forward {
// For SPF and DMARC pass, otherwise the test ends quickly.
resolver.TXT["bad.example."] = []string{"v=spf1 ip4:127.0.0.10 -all"}
resolver.TXT["good.example."] = []string{"v=spf1 ip4:127.0.0.10 -all"}
rcptTo = "mjl@mox.example" // Without IsForward rule.
}
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/junk/mox.conf"), resolver)
defer ts.close()
totalEvaluations := 0
var msgBad = strings.ReplaceAll(`From: <remote@bad.example>
To: <mjl@mox.example>
Subject: test
Message-Id: <bad@example.org>
test email
`, "\n", "\r\n")
var msgOK = strings.ReplaceAll(`From: <remote@good.example>
To: <mjl@mox.example>
Subject: other
Message-Id: <good@example.org>
unrelated message.
`, "\n", "\r\n")
var msgOK2 = strings.ReplaceAll(`From: <other@forward.example>
To: <mjl@mox.example>
Subject: non-forward
Message-Id: <regular@example.org>
happens to come from forwarding mail server.
`, "\n", "\r\n")
// Deliver forwarded messages, then classify as junk. Normally enough to treat
// other unrelated messages from IP as junk, but not for forwarded messages.
ts.run(func(err error, client *smtpclient.Client) {
tcheck(t, err, "connect")
mailFrom := "remote@forward.example"
if !forward {
mailFrom = "remote@bad.example"
}
for i := 0; i < 10; i++ {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(msgBad)), strings.NewReader(msgBad), false, false, false)
tcheck(t, err, "deliver message")
}
totalEvaluations += 10
n, err := bstore.QueryDB[store.Message](ctxbg, ts.acc.DB).UpdateFields(map[string]any{"Junk": true, "MsgFromValidated": true})
tcheck(t, err, "marking messages as junk")
tcompare(t, n, 10)
// Next delivery will fail, with negative "message From" signal.
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(msgBad)), strings.NewReader(msgBad), false, false, false)
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C451LocalErr {
t.Fatalf("delivery by bad sender, got err %v, expected smtpclient.Error with code %d", err, smtp.C451LocalErr)
}
checkEvaluationCount(t, totalEvaluations)
})
// Delivery from different "message From" without reputation, but from same
// forwarding email server, should succeed under forwarding, not as regular sending
// server.
ts.run(func(err error, client *smtpclient.Client) {
tcheck(t, err, "connect")
mailFrom := "remote@forward.example"
if !forward {
mailFrom = "remote@good.example"
}
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(msgOK)), strings.NewReader(msgOK), false, false, false)
if forward {
tcheck(t, err, "deliver")
totalEvaluations += 1
} else {
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C451LocalErr {
t.Fatalf("delivery by bad ip, got err %v, expected smtpclient.Error with code %d", err, smtp.C451LocalErr)
}
}
checkEvaluationCount(t, totalEvaluations)
})
// Delivery from forwarding server that isn't a forward should get same treatment.
ts.run(func(err error, client *smtpclient.Client) {
tcheck(t, err, "connect")
mailFrom := "other@forward.example"
// Ensure To header matches.
msg := msgOK2
if forward {
msg = strings.ReplaceAll(msg, "<mjl@mox.example>", "<mjl3@mox.example>")
}
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(msg)), strings.NewReader(msg), false, false, false)
if forward {
tcheck(t, err, "deliver")
totalEvaluations += 1
} else {
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C451LocalErr {
t.Fatalf("delivery by bad ip, got err %v, expected smtpclient.Error with code %d", err, smtp.C451LocalErr)
}
}
checkEvaluationCount(t, totalEvaluations)
})
}
check(true)
check(false)
}
2023-01-30 16:27:06 +03:00
// Messages that we sent to, that have passing DMARC, but that are otherwise spammy, should be accepted.
func TestDMARCSent(t *testing.T) {
resolver := &dns.MockResolver{
A: map[string][]string{
"example.org.": {"127.0.0.1"}, // For mx check.
},
TXT: map[string][]string{
"example.org.": {"v=spf1 ip4:127.0.0.1 -all"},
"_dmarc.example.org.": {"v=DMARC1;p=reject;rua=mailto:dmarcrpt@example.org"},
2023-01-30 16:27:06 +03:00
},
}
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/junk/mox.conf"), resolver)
2023-01-30 16:27:06 +03:00
defer ts.close()
// First check that DMARC policy rejects message and results in optional evaluation.
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@mox.example"
if err == nil {
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
}
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C550MailboxUnavail {
t.Fatalf("attempt to deliver spamy message, got err %v, expected smtpclient.Error with code %d", err, smtp.C550MailboxUnavail)
}
l := checkEvaluationCount(t, 1)
tcompare(t, l[0].Optional, true)
})
// Update DNS for an SPF pass, and DMARC pass.
resolver.TXT["example.org."] = []string{"v=spf1 ip4:127.0.0.10 -all"}
2023-01-30 16:27:06 +03:00
// Insert spammy messages not related to the test message.
m := store.Message{
MailFrom: "remote@test.example",
RcptToLocalpart: smtp.Localpart("mjl"),
RcptToDomain: "mox.example",
Flags: store.Flags{Seen: true, Junk: true},
Size: int64(len(deliverMessage)),
2023-01-30 16:27:06 +03:00
}
for i := 0; i < 3; i++ {
nm := m
tinsertmsg(t, ts.acc, "Archive", &nm, deliverMessage)
}
tretrain(t, ts.acc)
// Baseline, message should be refused for spammy content.
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@mox.example"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
2023-01-30 16:27:06 +03:00
}
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C451LocalErr {
t.Fatalf("attempt to deliver spamy message, got err %v, expected smtpclient.Error with code %d", err, smtp.C451LocalErr)
}
checkEvaluationCount(t, 1) // No new evaluation.
2023-01-30 16:27:06 +03:00
})
// Insert a message that we sent to the address that is about to send to us.
sentMsg := store.Message{Size: int64(len(deliverMessage))}
2023-01-30 16:27:06 +03:00
tinsertmsg(t, ts.acc, "Sent", &sentMsg, deliverMessage)
err := ts.acc.DB.Insert(ctxbg, &store.Recipient{MessageID: sentMsg.ID, Localpart: "remote", Domain: "example.org", OrgDomain: "example.org", Sent: time.Now()})
2023-01-30 16:27:06 +03:00
tcheck(t, err, "inserting message recipient")
// Reject a message due to DMARC again. Since we sent a message to the domain, it
// is no longer unknown and we should see a non-optional evaluation that will
// result in a DMARC report.
resolver.TXT["example.org."] = []string{"v=spf1 ip4:127.0.0.1 -all"}
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@mox.example"
if err == nil {
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
}
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C550MailboxUnavail {
t.Fatalf("attempt to deliver spamy message, got err %v, expected smtpclient.Error with code %d", err, smtp.C550MailboxUnavail)
}
l := checkEvaluationCount(t, 2) // New evaluation.
tcompare(t, l[1].Optional, false)
})
2023-01-30 16:27:06 +03:00
// We should now be accepting the message because we recently sent a message.
resolver.TXT["example.org."] = []string{"v=spf1 ip4:127.0.0.10 -all"}
2023-01-30 16:27:06 +03:00
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@mox.example"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
2023-01-30 16:27:06 +03:00
}
tcheck(t, err, "deliver")
l := checkEvaluationCount(t, 3) // New evaluation.
tcompare(t, l[2].Optional, false)
2023-01-30 16:27:06 +03:00
})
}
// Test DNSBL, then getting through with subjectpass.
func TestBlocklistedSubjectpass(t *testing.T) {
// Set up a DNSBL on dnsbl.example, and get DMARC pass.
resolver := &dns.MockResolver{
A: map[string][]string{
"example.org.": {"127.0.0.10"}, // For mx check.
"2.0.0.127.dnsbl.example.": {"127.0.0.2"}, // For healthcheck.
"10.0.0.127.dnsbl.example.": {"127.0.0.10"}, // Where our connection pretends to come from.
},
TXT: map[string][]string{
"10.0.0.127.dnsbl.example.": {"blocklisted"},
"example.org.": {"v=spf1 ip4:127.0.0.10 -all"},
"_dmarc.example.org.": {"v=DMARC1;p=reject"},
},
PTR: map[string][]string{
"127.0.0.10": {"example.org."}, // For iprev check.
},
}
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/mox.conf"), resolver)
2023-01-30 16:27:06 +03:00
ts.dnsbls = []dns.Domain{{ASCII: "dnsbl.example"}}
defer ts.close()
// Message should be refused softly (temporary error) due to DNSBL.
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@mox.example"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
2023-01-30 16:27:06 +03:00
}
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C451LocalErr {
t.Fatalf("attempted deliver from dnsblocklisted ip, got err %v, expected smtpclient.Error with code %d", err, smtp.C451LocalErr)
}
})
// Set up subjectpass on account.
acc := mox.Conf.Dynamic.Accounts[ts.acc.Name]
acc.SubjectPass.Period = time.Hour
mox.Conf.Dynamic.Accounts[ts.acc.Name] = acc
// Message should be refused quickly (permanent error) due to DNSBL and Subjectkey.
var pass string
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@mox.example"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
2023-01-30 16:27:06 +03:00
}
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C550MailboxUnavail {
t.Fatalf("attempted deliver from dnsblocklisted ip, got err %v, expected smtpclient.Error with code %d", err, smtp.C550MailboxUnavail)
}
i := strings.Index(cerr.Line, subjectpass.Explanation)
if i < 0 {
t.Fatalf("got error line %q, expected error line with subjectpass", cerr.Line)
}
pass = cerr.Line[i+len(subjectpass.Explanation):]
})
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@mox.example"
passMessage := strings.Replace(deliverMessage, "Subject: test", "Subject: test "+pass, 1)
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(passMessage)), strings.NewReader(passMessage), false, false, false)
2023-01-30 16:27:06 +03:00
}
tcheck(t, err, "deliver with subjectpass")
})
}
// Test accepting a DMARC report.
func TestDMARCReport(t *testing.T) {
resolver := &dns.MockResolver{
A: map[string][]string{
"example.org.": {"127.0.0.10"}, // For mx check.
},
TXT: map[string][]string{
"example.org.": {"v=spf1 ip4:127.0.0.10 -all"},
"_dmarc.example.org.": {"v=DMARC1;p=reject; rua=mailto:dmarcrpt@example.org"},
2023-01-30 16:27:06 +03:00
},
PTR: map[string][]string{
"127.0.0.10": {"example.org."}, // For iprev check.
},
}
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/dmarcreport/mox.conf"), resolver)
2023-01-30 16:27:06 +03:00
defer ts.close()
run := func(report string, n int) {
t.Helper()
ts.run(func(err error, client *smtpclient.Client) {
t.Helper()
tcheck(t, err, "run")
mailFrom := "remote@example.org"
rcptTo := "mjl@mox.example"
msgb := &bytes.Buffer{}
_, xerr := fmt.Fprintf(msgb, "From: %s\r\nTo: %s\r\nSubject: dmarc report\r\nMIME-Version: 1.0\r\nContent-Type: text/xml\r\n\r\n", mailFrom, rcptTo)
tcheck(t, xerr, "write msg headers")
w := quotedprintable.NewWriter(msgb)
_, xerr = w.Write([]byte(strings.ReplaceAll(report, "\n", "\r\n")))
tcheck(t, xerr, "write message")
msg := msgb.String()
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(msg)), strings.NewReader(msg), false, false, false)
2023-01-30 16:27:06 +03:00
}
tcheck(t, err, "deliver")
records, err := dmarcdb.Records(ctxbg)
2023-01-30 16:27:06 +03:00
tcheck(t, err, "dmarcdb records")
if len(records) != n {
t.Fatalf("got %d dmarcdb records, expected %d or more", len(records), n)
}
})
}
run(dmarcReport, 0)
run(strings.ReplaceAll(dmarcReport, "xmox.nl", "mox.example"), 1)
// We always store as an evaluation, but as optional for reports.
evals := checkEvaluationCount(t, 2)
tcompare(t, evals[0].Optional, true)
tcompare(t, evals[1].Optional, true)
2023-01-30 16:27:06 +03:00
}
const dmarcReport = `<?xml version="1.0" encoding="UTF-8" ?>
<feedback>
<report_metadata>
<org_name>example.org</org_name>
<email>postmaster@example.org</email>
<report_id>1</report_id>
<date_range>
<begin>1596412800</begin>
<end>1596499199</end>
</date_range>
</report_metadata>
<policy_published>
<domain>xmox.nl</domain>
<adkim>r</adkim>
<aspf>r</aspf>
<p>reject</p>
<sp>reject</sp>
<pct>100</pct>
</policy_published>
<record>
<row>
<source_ip>127.0.0.10</source_ip>
<count>1</count>
<policy_evaluated>
<disposition>none</disposition>
<dkim>pass</dkim>
<spf>pass</spf>
</policy_evaluated>
</row>
<identifiers>
<header_from>xmox.nl</header_from>
</identifiers>
<auth_results>
<dkim>
<domain>xmox.nl</domain>
<result>pass</result>
<selector>testsel</selector>
</dkim>
<spf>
<domain>xmox.nl</domain>
<result>pass</result>
</spf>
</auth_results>
</record>
</feedback>
`
// Test accepting a TLS report.
func TestTLSReport(t *testing.T) {
// Requires setting up DKIM.
privKey := ed25519.NewKeyFromSeed(make([]byte, ed25519.SeedSize)) // Fake key, don't use this for real!
dkimRecord := dkim.Record{
Version: "DKIM1",
Hashes: []string{"sha256"},
Flags: []string{"s"},
PublicKey: privKey.Public(),
Key: "ed25519",
}
dkimTxt, err := dkimRecord.Record()
tcheck(t, err, "dkim record")
sel := config.Selector{
HashEffective: "sha256",
HeadersEffective: []string{"From", "To", "Subject", "Date"},
Key: privKey,
Domain: dns.Domain{ASCII: "testsel"},
}
dkimConf := config.DKIM{
Selectors: map[string]config.Selector{"testsel": sel},
Sign: []string{"testsel"},
}
resolver := &dns.MockResolver{
A: map[string][]string{
"example.org.": {"127.0.0.10"}, // For mx check.
},
TXT: map[string][]string{
"testsel._domainkey.example.org.": {dkimTxt},
"_dmarc.example.org.": {"v=DMARC1;p=reject;rua=mailto:dmarcrpt@example.org"},
2023-01-30 16:27:06 +03:00
},
PTR: map[string][]string{
"127.0.0.10": {"example.org."}, // For iprev check.
},
}
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/tlsrpt/mox.conf"), resolver)
2023-01-30 16:27:06 +03:00
defer ts.close()
run := func(rcptTo, tlsrpt string, n int) {
2023-01-30 16:27:06 +03:00
t.Helper()
ts.run(func(err error, client *smtpclient.Client) {
t.Helper()
mailFrom := "remote@example.org"
msgb := &bytes.Buffer{}
_, xerr := fmt.Fprintf(msgb, "From: %s\r\nTo: %s\r\nSubject: tlsrpt report\r\nMIME-Version: 1.0\r\nContent-Type: application/tlsrpt+json\r\n\r\n%s\r\n", mailFrom, rcptTo, tlsrpt)
tcheck(t, xerr, "write msg")
msg := msgb.String()
selectors := mox.DKIMSelectors(dkimConf)
headers, xerr := dkim.Sign(ctxbg, pkglog.Logger, "remote", dns.Domain{ASCII: "example.org"}, selectors, false, strings.NewReader(msg))
2023-01-30 16:27:06 +03:00
tcheck(t, xerr, "dkim sign")
msg = headers + msg
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(msg)), strings.NewReader(msg), false, false, false)
2023-01-30 16:27:06 +03:00
}
tcheck(t, err, "deliver")
records, err := tlsrptdb.Records(ctxbg)
2023-01-30 16:27:06 +03:00
tcheck(t, err, "tlsrptdb records")
if len(records) != n {
t.Fatalf("got %d tlsrptdb records, expected %d", len(records), n)
}
})
}
const tlsrpt = `{"organization-name":"Example.org","date-range":{"start-datetime":"2022-01-07T00:00:00Z","end-datetime":"2022-01-07T23:59:59Z"},"contact-info":"tlsrpt@example.org","report-id":"1","policies":[{"policy":{"policy-type":"no-policy-found","policy-domain":"xmox.nl"},"summary":{"total-successful-session-count":1,"total-failure-session-count":0}}]}`
run("mjl@mox.example", tlsrpt, 0)
run("mjl@mox.example", strings.ReplaceAll(tlsrpt, "xmox.nl", "mox.example"), 1)
run("mjl@mailhost.mox.example", strings.ReplaceAll(tlsrpt, "xmox.nl", "mailhost.mox.example"), 2)
// We always store as an evaluation, but as optional for reports.
evals := checkEvaluationCount(t, 3)
tcompare(t, evals[0].Optional, true)
tcompare(t, evals[1].Optional, true)
tcompare(t, evals[2].Optional, true)
}
func TestRatelimitConnectionrate(t *testing.T) {
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/mox.conf"), dns.MockResolver{})
defer ts.close()
// We'll be creating 300 connections, no TLS and reduce noise.
ts.tlsmode = smtpclient.TLSSkip
mlog.SetConfig(map[string]slog.Level{"": mlog.LevelInfo})
// We may be passing a window boundary during this tests. The limit is 300/minute.
// So make twice that many connections and hope the tests don't take too long.
for i := 0; i <= 2*300; i++ {
ts.run(func(err error, client *smtpclient.Client) {
t.Helper()
if err != nil && i < 300 {
t.Fatalf("expected smtp connection, got %v", err)
}
if err == nil && i == 600 {
t.Fatalf("expected no smtp connection due to connection rate limit, got connection")
}
if client != nil {
client.Close()
}
})
}
}
func TestRatelimitAuth(t *testing.T) {
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/mox.conf"), dns.MockResolver{})
defer ts.close()
ts.submission = true
ts.tlsmode = smtpclient.TLSSkip
ts.user = "bad"
ts.pass = "bad"
// We may be passing a window boundary during this tests. The limit is 10 auth
// failures/minute. So make twice that many connections and hope the tests don't
// take too long.
for i := 0; i <= 2*10; i++ {
ts.run(func(err error, client *smtpclient.Client) {
t.Helper()
if err == nil {
t.Fatalf("got auth success with bad credentials")
}
var cerr smtpclient.Error
badauth := errors.As(err, &cerr) && cerr.Code == smtp.C535AuthBadCreds
if !badauth && i < 10 {
t.Fatalf("expected auth failure, got %v", err)
}
if badauth && i == 20 {
t.Fatalf("expected no smtp connection due to failed auth rate limit, got other error %v", err)
}
if client != nil {
client.Close()
}
})
}
}
func TestRatelimitDelivery(t *testing.T) {
resolver := dns.MockResolver{
A: map[string][]string{
"example.org.": {"127.0.0.10"}, // For mx check.
},
PTR: map[string][]string{
"127.0.0.10": {"example.org."},
},
}
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/mox.conf"), resolver)
defer ts.close()
orig := limitIPMasked1MessagesPerMinute
limitIPMasked1MessagesPerMinute = 1
defer func() {
limitIPMasked1MessagesPerMinute = orig
}()
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@mox.example"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
}
tcheck(t, err, "deliver to remote")
2023-01-30 16:27:06 +03:00
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C452StorageFull {
t.Fatalf("got err %v, expected smtpclient error with code 452 for storage full", err)
}
})
limitIPMasked1MessagesPerMinute = orig
origSize := limitIPMasked1SizePerMinute
// Message was already delivered once. We'll do another one. But the 3rd will fail.
// We need the actual size with prepended headers, since that is used in the
// calculations.
msg, err := bstore.QueryDB[store.Message](ctxbg, ts.acc.DB).Get()
if err != nil {
t.Fatalf("getting delivered message for its size: %v", err)
}
limitIPMasked1SizePerMinute = 2*msg.Size + int64(len(deliverMessage)/2)
defer func() {
limitIPMasked1SizePerMinute = origSize
}()
ts.run(func(err error, client *smtpclient.Client) {
mailFrom := "remote@example.org"
rcptTo := "mjl@mox.example"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
}
tcheck(t, err, "deliver to remote")
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
var cerr smtpclient.Error
if err == nil || !errors.As(err, &cerr) || cerr.Code != smtp.C452StorageFull {
t.Fatalf("got err %v, expected smtpclient error with code 452 for storage full", err)
}
})
2023-01-30 16:27:06 +03:00
}
func TestNonSMTP(t *testing.T) {
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/mox.conf"), dns.MockResolver{})
defer ts.close()
ts.cid += 2
serverConn, clientConn := net.Pipe()
defer serverConn.Close()
serverdone := make(chan struct{})
defer func() { <-serverdone }()
go func() {
tlsConfig := &tls.Config{
Certificates: []tls.Certificate{fakeCert(ts.t)},
}
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
serve("test", ts.cid-2, dns.Domain{ASCII: "mox.example"}, tlsConfig, serverConn, ts.resolver, ts.submission, false, 100<<20, false, false, false, ts.dnsbls, 0)
close(serverdone)
}()
defer clientConn.Close()
buf := make([]byte, 128)
// Read and ignore hello.
if _, err := clientConn.Read(buf); err != nil {
t.Fatalf("reading hello: %v", err)
}
if _, err := fmt.Fprintf(clientConn, "bogus\r\n"); err != nil {
t.Fatalf("write command: %v", err)
}
n, err := clientConn.Read(buf)
if err != nil {
t.Fatalf("read response line: %v", err)
}
s := string(buf[:n])
if !strings.HasPrefix(s, "500 5.5.2 ") {
t.Fatalf(`got %q, expected "500 5.5.2 ...`, s)
}
if _, err := clientConn.Read(buf); err == nil {
t.Fatalf("connection not closed after bogus command")
}
}
// Test limits on outgoing messages.
func TestLimitOutgoing(t *testing.T) {
ts := newTestServer(t, filepath.FromSlash("../testdata/smtpserversendlimit/mox.conf"), dns.MockResolver{})
defer ts.close()
ts.user = "mjl@mox.example"
ts.pass = password0
ts.submission = true
err := ts.acc.DB.Insert(ctxbg, &store.Outgoing{Recipient: "a@other.example", Submitted: time.Now().Add(-24*time.Hour - time.Minute)})
tcheck(t, err, "inserting outgoing/recipient past 24h window")
testSubmit := func(rcptTo string, expErr *smtpclient.Error) {
t.Helper()
ts.run(func(err error, client *smtpclient.Client) {
t.Helper()
mailFrom := "mjl@mox.example"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(submitMessage)), strings.NewReader(submitMessage), false, false, false)
}
ts.smtpErr(err, expErr)
})
}
// Limits are set to 4 messages a day, 2 first-time recipients.
testSubmit("b@other.example", nil)
testSubmit("c@other.example", nil)
testSubmit("d@other.example", &smtpclient.Error{Code: smtp.C451LocalErr, Secode: smtp.SePol7DeliveryUnauth1}) // Would be 3rd recipient.
testSubmit("b@other.example", nil)
testSubmit("b@other.example", nil)
testSubmit("b@other.example", &smtpclient.Error{Code: smtp.C451LocalErr, Secode: smtp.SePol7DeliveryUnauth1}) // Would be 5th message.
}
// Test account size limit enforcement.
func TestQuota(t *testing.T) {
resolver := dns.MockResolver{
A: map[string][]string{
"other.example.": {"127.0.0.10"}, // For mx check.
},
PTR: map[string][]string{
"127.0.0.10": {"other.example."},
},
}
ts := newTestServer(t, filepath.FromSlash("../testdata/smtpserverquota/mox.conf"), resolver)
defer ts.close()
testDeliver := func(rcptTo string, expErr *smtpclient.Error) {
t.Helper()
ts.run(func(err error, client *smtpclient.Client) {
t.Helper()
mailFrom := "mjl@other.example"
if err == nil {
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
}
ts.smtpErr(err, expErr)
})
}
testDeliver("mjl@mox.example", &smtpclient.Error{Code: smtp.C452StorageFull, Secode: smtp.SeMailbox2Full2})
}
// Test with catchall destination address.
func TestCatchall(t *testing.T) {
resolver := dns.MockResolver{
A: map[string][]string{
"other.example.": {"127.0.0.10"}, // For mx check.
},
PTR: map[string][]string{
"127.0.0.10": {"other.example."},
},
}
ts := newTestServer(t, filepath.FromSlash("../testdata/smtpservercatchall/mox.conf"), resolver)
defer ts.close()
testDeliver := func(rcptTo string, expErr *smtpclient.Error) {
t.Helper()
ts.run(func(err error, client *smtpclient.Client) {
t.Helper()
mailFrom := "mjl@other.example"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(submitMessage)), strings.NewReader(submitMessage), false, false, false)
}
ts.smtpErr(err, expErr)
})
}
testDeliver("mjl@mox.example", nil) // Exact match.
testDeliver("mjl+test@mox.example", nil) // Domain localpart catchall separator.
testDeliver("MJL+TEST@mox.example", nil) // Again, and case insensitive.
testDeliver("unknown@mox.example", nil) // Catchall address, to account catchall.
n, err := bstore.QueryDB[store.Message](ctxbg, ts.acc.DB).Count()
tcheck(t, err, "checking delivered messages")
tcompare(t, n, 3)
acc, err := store.OpenAccount(pkglog, "catchall")
tcheck(t, err, "open account")
defer func() {
acc.Close()
acc.CheckClosed()
}()
n, err = bstore.QueryDB[store.Message](ctxbg, acc.DB).Count()
tcheck(t, err, "checking delivered messages to catchall account")
tcompare(t, n, 1)
}
// Test DKIM signing for outgoing messages.
func TestDKIMSign(t *testing.T) {
resolver := dns.MockResolver{
A: map[string][]string{
"mox.example.": {"127.0.0.10"}, // For mx check.
},
PTR: map[string][]string{
"127.0.0.10": {"mox.example."},
},
}
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/mox.conf"), resolver)
defer ts.close()
// Set DKIM signing config.
var gen byte
genDKIM := func(domain string) string {
dom, _ := mox.Conf.Domain(dns.Domain{ASCII: domain})
privkey := make([]byte, ed25519.SeedSize) // Fake key, don't use for real.
gen++
privkey[0] = byte(gen)
sel := config.Selector{
HashEffective: "sha256",
HeadersEffective: []string{"From", "To", "Subject"},
Key: ed25519.NewKeyFromSeed(privkey),
Domain: dns.Domain{ASCII: "testsel"},
}
dom.DKIM = config.DKIM{
Selectors: map[string]config.Selector{"testsel": sel},
Sign: []string{"testsel"},
}
mox.Conf.Dynamic.Domains[domain] = dom
pubkey := sel.Key.Public().(ed25519.PublicKey)
return "v=DKIM1;k=ed25519;p=" + base64.StdEncoding.EncodeToString(pubkey)
}
dkimtxt := genDKIM("mox.example")
dkimtxt2 := genDKIM("mox2.example")
// DKIM verify needs to find the key.
resolver.TXT = map[string][]string{
"testsel._domainkey.mox.example.": {dkimtxt},
"testsel._domainkey.mox2.example.": {dkimtxt2},
}
ts.submission = true
ts.user = "mjl@mox.example"
ts.pass = password0
n := 0
testSubmit := func(mailFrom, msgFrom string) {
t.Helper()
ts.run(func(err error, client *smtpclient.Client) {
t.Helper()
msg := strings.ReplaceAll(fmt.Sprintf(`From: <%s>
To: <remote@example.org>
Subject: test
Message-Id: <test@mox.example>
test email
`, msgFrom), "\n", "\r\n")
rcptTo := "remote@example.org"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(msg)), strings.NewReader(msg), false, false, false)
}
tcheck(t, err, "deliver")
add a webapi and webhooks for a simple http/json-based api for applications to compose/send messages, receive delivery feedback, and maintain suppression lists. this is an alternative to applications using a library to compose messages, submitting those messages using smtp, and monitoring a mailbox with imap for DSNs, which can be processed into the equivalent of suppression lists. but you need to know about all these standards/protocols and find libraries. by using the webapi & webhooks, you just need a http & json library. unfortunately, there is no standard for these kinds of api, so mox has made up yet another one... matching incoming DSNs about deliveries to original outgoing messages requires keeping history of "retired" messages (delivered from the queue, either successfully or failed). this can be enabled per account. history is also useful for debugging deliveries. we now also keep history of each delivery attempt, accessible while still in the queue, and kept when a message is retired. the queue webadmin pages now also have pagination, to show potentially large history. a queue of webhook calls is now managed too. failures are retried similar to message deliveries. webhooks can also be saved to the retired list after completing. also configurable per account. messages can be sent with a "unique smtp mail from" address. this can only be used if the domain is configured with a localpart catchall separator such as "+". when enabled, a queued message gets assigned a random "fromid", which is added after the separator when sending. when DSNs are returned, they can be related to previously sent messages based on this fromid. in the future, we can implement matching on the "envid" used in the smtp dsn extension, or on the "message-id" of the message. using a fromid can be triggered by authenticating with a login email address that is configured as enabling fromid. suppression lists are automatically managed per account. if a delivery attempt results in certain smtp errors, the destination address is added to the suppression list. future messages queued for that recipient will immediately fail without a delivery attempt. suppression lists protect your mail server reputation. submitted messages can carry "extra" data through the queue and webhooks for outgoing deliveries. through webapi as a json object, through smtp submission as message headers of the form "x-mox-extra-<key>: value". to make it easy to test webapi/webhooks locally, the "localserve" mode actually puts messages in the queue. when it's time to deliver, it still won't do a full delivery attempt, but just delivers to the sender account. unless the recipient address has a special form, simulating a failure to deliver. admins now have more control over the queue. "hold rules" can be added to mark newly queued messages as "on hold", pausing delivery. rules can be about certain sender or recipient domains/addresses, or apply to all messages pausing the entire queue. also useful for (local) testing. new config options have been introduced. they are editable through the admin and/or account web interfaces. the webapi http endpoints are enabled for newly generated configs with the quickstart, and in localserve. existing configurations must explicitly enable the webapi in mox.conf. gopherwatch.org was created to dogfood this code. it initially used just the compose/smtpclient/imapclient mox packages to send messages and process delivery feedback. it will get a config option to use the mox webapi/webhooks instead. the gopherwatch code to use webapi/webhook is smaller and simpler, and developing that shaped development of the mox webapi/webhooks. for issue #31 by cuu508
2024-04-15 22:49:02 +03:00
msgs, err := queue.List(ctxbg, queue.Filter{}, queue.Sort{})
tcheck(t, err, "listing queue")
n++
tcompare(t, len(msgs), n)
sort.Slice(msgs, func(i, j int) bool {
return msgs[i].ID > msgs[j].ID
})
f, err := queue.OpenMessage(ctxbg, msgs[0].ID)
tcheck(t, err, "open message in queue")
defer f.Close()
results, err := dkim.Verify(ctxbg, pkglog.Logger, resolver, false, dkim.DefaultPolicy, f, false)
tcheck(t, err, "verifying dkim message")
tcompare(t, len(results), 1)
tcompare(t, results[0].Status, dkim.StatusPass)
tcompare(t, results[0].Sig.Domain.ASCII, strings.Split(msgFrom, "@")[1])
})
}
testSubmit("mjl@mox.example", "mjl@mox.example")
testSubmit("mjl@mox.example", "mjl@mox2.example") // DKIM signature will be for mox2.example.
}
// Test to postmaster addresses.
func TestPostmaster(t *testing.T) {
resolver := dns.MockResolver{
A: map[string][]string{
"other.example.": {"127.0.0.10"}, // For mx check.
},
PTR: map[string][]string{
"127.0.0.10": {"other.example."},
},
}
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/postmaster/mox.conf"), resolver)
defer ts.close()
testDeliver := func(rcptTo string, expErr *smtpclient.Error) {
t.Helper()
ts.run(func(err error, client *smtpclient.Client) {
t.Helper()
mailFrom := "mjl@other.example"
if err == nil {
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(deliverMessage)), strings.NewReader(deliverMessage), false, false, false)
}
ts.smtpErr(err, expErr)
})
}
testDeliver("postmaster", nil) // Plain postmaster address without domain.
testDeliver("postmaster@host.mox.example", nil) // Postmaster address with configured mail server hostname.
testDeliver("postmaster@mox.example", nil) // Postmaster address without explicitly configured destination.
testDeliver("postmaster@unknown.example", &smtpclient.Error{Permanent: true, Code: smtp.C550MailboxUnavail, Secode: smtp.SeAddr1UnknownDestMailbox1})
}
// Test to address with empty localpart.
func TestEmptylocalpart(t *testing.T) {
resolver := dns.MockResolver{
A: map[string][]string{
"other.example.": {"127.0.0.10"}, // For mx check.
},
PTR: map[string][]string{
"127.0.0.10": {"other.example."},
},
}
make mox compile on windows, without "mox serve" but with working "mox localserve" getting mox to compile required changing code in only a few places where package "syscall" was used: for accessing file access times and for umask handling. an open problem is how to start a process as an unprivileged user on windows. that's why "mox serve" isn't implemented yet. and just finding a way to implement it now may not be good enough in the near future: we may want to starting using a more complete privilege separation approach, with a process handling sensitive tasks (handling private keys, authentication), where we may want to pass file descriptors between processes. how would that work on windows? anyway, getting mox to compile for windows doesn't mean it works properly on windows. the largest issue: mox would normally open a file, rename or remove it, and finally close it. this happens during message delivery. that doesn't work on windows, the rename/remove would fail because the file is still open. so this commit swaps many "remove" and "close" calls. renames are a longer story: message delivery had two ways to deliver: with "consuming" the (temporary) message file (which would rename it to its final destination), and without consuming (by hardlinking the file, falling back to copying). the last delivery to a recipient of a message (and the only one in the common case of a single recipient) would consume the message, and the earlier recipients would not. during delivery, the already open message file was used, to parse the message. we still want to use that open message file, and the caller now stays responsible for closing it, but we no longer try to rename (consume) the file. we always hardlink (or copy) during delivery (this works on windows), and the caller is responsible for closing and removing (in that order) the original temporary file. this does cost one syscall more. but it makes the delivery code (responsibilities) a bit simpler. there is one more obvious issue: the file system path separator. mox already used the "filepath" package to join paths in many places, but not everywhere. and it still used strings with slashes for local file access. with this commit, the code now uses filepath.FromSlash for path strings with slashes, uses "filepath" in a few more places where it previously didn't. also switches from "filepath" to regular "path" package when handling mailbox names in a few places, because those always use forward slashes, regardless of local file system conventions. windows can handle forward slashes when opening files, so test code that passes path strings with forward slashes straight to go stdlib file i/o functions are left unchanged to reduce code churn. the regular non-test code, or test code that uses path strings in places other than standard i/o functions, does have the paths converted for consistent paths (otherwise we would end up with paths with mixed forward/backward slashes in log messages). windows cannot dup a listening socket. for "mox localserve", it isn't important, and we can work around the issue. the current approach for "mox serve" (forking a process and passing file descriptors of listening sockets on "privileged" ports) won't work on windows. perhaps it isn't needed on windows, and any user can listen on "privileged" ports? that would be welcome. on windows, os.Open cannot open a directory, so we cannot call Sync on it after message delivery. a cursory internet search indicates that directories cannot be synced on windows. the story is probably much more nuanced than that, with long deep technical details/discussions/disagreement/confusion, like on unix. for "mox localserve" we can get away with making syncdir a no-op.
2023-10-14 11:54:07 +03:00
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/mox.conf"), resolver)
defer ts.close()
testDeliver := func(rcptTo string, expErr *smtpclient.Error) {
t.Helper()
ts.run(func(err error, client *smtpclient.Client) {
t.Helper()
mailFrom := `""@other.example`
msg := strings.ReplaceAll(deliverMessage, "To: <mjl@mox.example>", `To: <""@mox.example>`)
if err == nil {
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(msg)), strings.NewReader(msg), false, false, false)
}
ts.smtpErr(err, expErr)
})
}
testDeliver(`""@mox.example`, nil)
}
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
// Test handling REQUIRETLS and TLS-Required: No.
func TestRequireTLS(t *testing.T) {
resolver := dns.MockResolver{
A: map[string][]string{
"mox.example.": {"127.0.0.10"}, // For mx check.
},
PTR: map[string][]string{
"127.0.0.10": {"mox.example."},
},
}
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/mox.conf"), resolver)
defer ts.close()
ts.submission = true
ts.requiretls = true
ts.user = "mjl@mox.example"
ts.pass = password0
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
no := false
yes := true
msg0 := strings.ReplaceAll(`From: <mjl@mox.example>
To: <remote@example.org>
Subject: test
Message-Id: <test@mox.example>
TLS-Required: No
test email
`, "\n", "\r\n")
msg1 := strings.ReplaceAll(`From: <mjl@mox.example>
To: <remote@example.org>
Subject: test
Message-Id: <test@mox.example>
TLS-Required: No
TLS-Required: bogus
test email
`, "\n", "\r\n")
msg2 := strings.ReplaceAll(`From: <mjl@mox.example>
To: <remote@example.org>
Subject: test
Message-Id: <test@mox.example>
test email
`, "\n", "\r\n")
testSubmit := func(msg string, requiretls bool, expRequireTLS *bool) {
t.Helper()
ts.run(func(err error, client *smtpclient.Client) {
t.Helper()
rcptTo := "remote@example.org"
if err == nil {
err = client.Deliver(ctxbg, "mjl@mox.example", rcptTo, int64(len(msg)), strings.NewReader(msg), false, false, requiretls)
}
tcheck(t, err, "deliver")
add a webapi and webhooks for a simple http/json-based api for applications to compose/send messages, receive delivery feedback, and maintain suppression lists. this is an alternative to applications using a library to compose messages, submitting those messages using smtp, and monitoring a mailbox with imap for DSNs, which can be processed into the equivalent of suppression lists. but you need to know about all these standards/protocols and find libraries. by using the webapi & webhooks, you just need a http & json library. unfortunately, there is no standard for these kinds of api, so mox has made up yet another one... matching incoming DSNs about deliveries to original outgoing messages requires keeping history of "retired" messages (delivered from the queue, either successfully or failed). this can be enabled per account. history is also useful for debugging deliveries. we now also keep history of each delivery attempt, accessible while still in the queue, and kept when a message is retired. the queue webadmin pages now also have pagination, to show potentially large history. a queue of webhook calls is now managed too. failures are retried similar to message deliveries. webhooks can also be saved to the retired list after completing. also configurable per account. messages can be sent with a "unique smtp mail from" address. this can only be used if the domain is configured with a localpart catchall separator such as "+". when enabled, a queued message gets assigned a random "fromid", which is added after the separator when sending. when DSNs are returned, they can be related to previously sent messages based on this fromid. in the future, we can implement matching on the "envid" used in the smtp dsn extension, or on the "message-id" of the message. using a fromid can be triggered by authenticating with a login email address that is configured as enabling fromid. suppression lists are automatically managed per account. if a delivery attempt results in certain smtp errors, the destination address is added to the suppression list. future messages queued for that recipient will immediately fail without a delivery attempt. suppression lists protect your mail server reputation. submitted messages can carry "extra" data through the queue and webhooks for outgoing deliveries. through webapi as a json object, through smtp submission as message headers of the form "x-mox-extra-<key>: value". to make it easy to test webapi/webhooks locally, the "localserve" mode actually puts messages in the queue. when it's time to deliver, it still won't do a full delivery attempt, but just delivers to the sender account. unless the recipient address has a special form, simulating a failure to deliver. admins now have more control over the queue. "hold rules" can be added to mark newly queued messages as "on hold", pausing delivery. rules can be about certain sender or recipient domains/addresses, or apply to all messages pausing the entire queue. also useful for (local) testing. new config options have been introduced. they are editable through the admin and/or account web interfaces. the webapi http endpoints are enabled for newly generated configs with the quickstart, and in localserve. existing configurations must explicitly enable the webapi in mox.conf. gopherwatch.org was created to dogfood this code. it initially used just the compose/smtpclient/imapclient mox packages to send messages and process delivery feedback. it will get a config option to use the mox webapi/webhooks instead. the gopherwatch code to use webapi/webhook is smaller and simpler, and developing that shaped development of the mox webapi/webhooks. for issue #31 by cuu508
2024-04-15 22:49:02 +03:00
msgs, err := queue.List(ctxbg, queue.Filter{}, queue.Sort{})
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
tcheck(t, err, "listing queue")
tcompare(t, len(msgs), 1)
tcompare(t, msgs[0].RequireTLS, expRequireTLS)
_, err = queue.Drop(ctxbg, pkglog, queue.Filter{IDs: []int64{msgs[0].ID}})
implement "requiretls", rfc 8689 with requiretls, the tls verification mode/rules for email deliveries can be changed by the sender/submitter. in two ways: 1. "requiretls" smtp extension to always enforce verified tls (with mta-sts or dnssec+dane), along the entire delivery path until delivery into the final destination mailbox (so entire transport is verified-tls-protected). 2. "tls-required: no" message header, to ignore any tls and tls verification errors even if the recipient domain has a policy that requires tls verification (mta-sts and/or dnssec+dane), allowing delivery of non-sensitive messages in case of misconfiguration/interoperability issues (at least useful for sending tls reports). we enable requiretls by default (only when tls is active), for smtp and submission. it can be disabled through the config. for each delivery attempt, we now store (per recipient domain, in the account of the sender) whether the smtp server supports starttls and requiretls. this support is shown (after having sent a first message) in the webmail when sending a message (the previous 3 bars under the address input field are now 5 bars, the first for starttls support, the last for requiretls support). when all recipient domains for a message are known to implement requiretls, requiretls is automatically selected for sending (instead of "default" tls behaviour). users can also select the "fallback to insecure" to add the "tls-required: no" header. new metrics are added for insight into requiretls errors and (some, not yet all) cases where tls-required-no ignored a tls/verification error. the admin can change the requiretls status for messages in the queue. so with default delivery attempts, when verified tls is required by failing, an admin could potentially change the field to "tls-required: no"-behaviour. messages received (over smtp) with the requiretls option, get a comment added to their Received header line, just before "id", after "with".
2023-10-24 11:06:16 +03:00
tcheck(t, err, "deleting message from queue")
})
}
testSubmit(msg0, true, &yes) // Header ignored, requiretls applied.
testSubmit(msg0, false, &no) // TLS-Required header applied.
testSubmit(msg1, true, &yes) // Bad headers ignored, requiretls applied.
testSubmit(msg1, false, nil) // Inconsistent multiple headers ignored.
testSubmit(msg2, false, nil) // Regular message, no RequireTLS setting.
testSubmit(msg2, true, &yes) // Requiretls applied.
// Check that we get an error if remote SMTP server does not support the requiretls
// extension.
ts.requiretls = false
ts.run(func(err error, client *smtpclient.Client) {
t.Helper()
rcptTo := "remote@example.org"
if err == nil {
err = client.Deliver(ctxbg, "mjl@mox.example", rcptTo, int64(len(msg0)), strings.NewReader(msg0), false, false, true)
}
if err == nil {
t.Fatalf("delivered with requiretls to server without requiretls")
}
if !errors.Is(err, smtpclient.ErrRequireTLSUnsupported) {
t.Fatalf("got err %v, expected ErrRequireTLSUnsupported", err)
}
})
}
func TestSmuggle(t *testing.T) {
resolver := dns.MockResolver{
A: map[string][]string{
"example.org.": {"127.0.0.10"}, // For mx check.
},
PTR: map[string][]string{
"127.0.0.10": {"example.org."}, // For iprev check.
},
}
ts := newTestServer(t, filepath.FromSlash("../testdata/smtpsmuggle/mox.conf"), resolver)
ts.tlsmode = smtpclient.TLSSkip
defer ts.close()
test := func(data string) {
t.Helper()
ts.runRaw(func(conn net.Conn) {
t.Helper()
ourHostname := mox.Conf.Static.HostnameDomain
remoteHostname := dns.Domain{ASCII: "mox.example"}
opts := smtpclient.Opts{
RootCAs: mox.Conf.Static.TLS.CertPool,
}
log := pkglog.WithCid(ts.cid - 1)
_, err := smtpclient.New(ctxbg, log.Logger, conn, ts.tlsmode, ts.tlspkix, ourHostname, remoteHostname, opts)
tcheck(t, err, "smtpclient")
defer conn.Close()
write := func(s string) {
_, err := conn.Write([]byte(s))
tcheck(t, err, "write")
}
readPrefixLine := func(prefix string) string {
t.Helper()
buf := make([]byte, 512)
n, err := conn.Read(buf)
tcheck(t, err, "read")
s := strings.TrimRight(string(buf[:n]), "\r\n")
if !strings.HasPrefix(s, prefix) {
t.Fatalf("got smtp response %q, expected line with prefix %q", s, prefix)
}
return s
}
write("MAIL FROM:<remote@example.org>\r\n")
readPrefixLine("2")
write("RCPT TO:<mjl@mox.example>\r\n")
readPrefixLine("2")
write("DATA\r\n")
readPrefixLine("3")
write("\r\n") // Empty header.
write(data)
write("\r\n.\r\n") // End of message.
line := readPrefixLine("5")
if !strings.Contains(line, "smug") {
t.Errorf("got 5xx error with message %q, expected error text containing smug", line)
}
})
}
test("\r\n.\n")
test("\n.\n")
test("\r.\r")
test("\n.\r\n")
}
func TestFutureRelease(t *testing.T) {
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/mox.conf"), dns.MockResolver{})
ts.tlsmode = smtpclient.TLSSkip
ts.user = "mjl@mox.example"
ts.pass = password0
ts.submission = true
defer ts.close()
ts.auth = func(mechanisms []string, cs *tls.ConnectionState) (sasl.Client, error) {
return sasl.NewClientPlain(ts.user, ts.pass), nil
}
test := func(mailtoMore, expResponsePrefix string) {
t.Helper()
ts.runRaw(func(conn net.Conn) {
t.Helper()
ourHostname := mox.Conf.Static.HostnameDomain
remoteHostname := dns.Domain{ASCII: "mox.example"}
opts := smtpclient.Opts{Auth: ts.auth}
log := pkglog.WithCid(ts.cid - 1)
_, err := smtpclient.New(ctxbg, log.Logger, conn, ts.tlsmode, false, ourHostname, remoteHostname, opts)
tcheck(t, err, "smtpclient")
defer conn.Close()
write := func(s string) {
_, err := conn.Write([]byte(s))
tcheck(t, err, "write")
}
readPrefixLine := func(prefix string) string {
t.Helper()
buf := make([]byte, 512)
n, err := conn.Read(buf)
tcheck(t, err, "read")
s := strings.TrimRight(string(buf[:n]), "\r\n")
if !strings.HasPrefix(s, prefix) {
t.Fatalf("got smtp response %q, expected line with prefix %q", s, prefix)
}
return s
}
write(fmt.Sprintf("MAIL FROM:<mjl@mox.example>%s\r\n", mailtoMore))
readPrefixLine(expResponsePrefix)
if expResponsePrefix != "2" {
return
}
write("RCPT TO:<mjl@mox.example>\r\n")
readPrefixLine("2")
write("DATA\r\n")
readPrefixLine("3")
write("From: <mjl@mox.example>\r\n\r\nbody\r\n\r\n.\r\n")
readPrefixLine("2")
})
}
test(" HOLDFOR=1", "2")
test(" HOLDUNTIL="+time.Now().Add(time.Minute).UTC().Format(time.RFC3339), "2")
test(" HOLDUNTIL="+time.Now().Add(time.Minute).UTC().Format(time.RFC3339Nano), "2")
test(" HOLDFOR=0", "501") // 0 is invalid syntax.
test(fmt.Sprintf(" HOLDFOR=%d", int64((queue.FutureReleaseIntervalMax+time.Minute)/time.Second)), "554") // Too far in the future.
test(" HOLDUNTIL="+time.Now().Add(-time.Minute).UTC().Format(time.RFC3339), "554") // In the past.
test(" HOLDUNTIL="+time.Now().Add(queue.FutureReleaseIntervalMax+time.Minute).UTC().Format(time.RFC3339), "554") // Too far in the future.
test(" HOLDUNTIL=2024-02-10T17:28:00+00:00", "501") // "Z" required.
test(" HOLDUNTIL=24-02-10T17:28:00Z", "501") // Invalid.
test(" HOLDFOR=1 HOLDFOR=1", "501") // Duplicate.
test(" HOLDFOR=1 HOLDUNTIL="+time.Now().Add(time.Hour).UTC().Format(time.RFC3339), "501") // Duplicate.
}
// Test SMTPUTF8
func TestSMTPUTF8(t *testing.T) {
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/mox.conf"), dns.MockResolver{})
defer ts.close()
ts.user = "mjl@mox.example"
ts.pass = password0
ts.submission = true
test := func(mailFrom string, rcptTo string, headerValue string, filename string, clientSmtputf8 bool, expectedSmtputf8 bool, expErr *smtpclient.Error) {
t.Helper()
ts.run(func(_ error, client *smtpclient.Client) {
t.Helper()
msg := strings.ReplaceAll(fmt.Sprintf(`From: <%s>
To: <%s>
Subject: test
X-Custom-Test-Header: %s
MIME-Version: 1.0
Content-type: multipart/mixed; boundary="simple boundary"
--simple boundary
Content-Type: text/plain; charset=UTF-8;
Content-Disposition: attachment; filename="%s"
Content-Transfer-Encoding: base64
QW4gYXR0YWNoZWQgdGV4dCBmaWxlLg==
--simple boundary--
`, mailFrom, rcptTo, headerValue, filename), "\n", "\r\n")
err := client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(msg)), strings.NewReader(msg), true, clientSmtputf8, false)
ts.smtpErr(err, expErr)
if err != nil {
return
}
add a webapi and webhooks for a simple http/json-based api for applications to compose/send messages, receive delivery feedback, and maintain suppression lists. this is an alternative to applications using a library to compose messages, submitting those messages using smtp, and monitoring a mailbox with imap for DSNs, which can be processed into the equivalent of suppression lists. but you need to know about all these standards/protocols and find libraries. by using the webapi & webhooks, you just need a http & json library. unfortunately, there is no standard for these kinds of api, so mox has made up yet another one... matching incoming DSNs about deliveries to original outgoing messages requires keeping history of "retired" messages (delivered from the queue, either successfully or failed). this can be enabled per account. history is also useful for debugging deliveries. we now also keep history of each delivery attempt, accessible while still in the queue, and kept when a message is retired. the queue webadmin pages now also have pagination, to show potentially large history. a queue of webhook calls is now managed too. failures are retried similar to message deliveries. webhooks can also be saved to the retired list after completing. also configurable per account. messages can be sent with a "unique smtp mail from" address. this can only be used if the domain is configured with a localpart catchall separator such as "+". when enabled, a queued message gets assigned a random "fromid", which is added after the separator when sending. when DSNs are returned, they can be related to previously sent messages based on this fromid. in the future, we can implement matching on the "envid" used in the smtp dsn extension, or on the "message-id" of the message. using a fromid can be triggered by authenticating with a login email address that is configured as enabling fromid. suppression lists are automatically managed per account. if a delivery attempt results in certain smtp errors, the destination address is added to the suppression list. future messages queued for that recipient will immediately fail without a delivery attempt. suppression lists protect your mail server reputation. submitted messages can carry "extra" data through the queue and webhooks for outgoing deliveries. through webapi as a json object, through smtp submission as message headers of the form "x-mox-extra-<key>: value". to make it easy to test webapi/webhooks locally, the "localserve" mode actually puts messages in the queue. when it's time to deliver, it still won't do a full delivery attempt, but just delivers to the sender account. unless the recipient address has a special form, simulating a failure to deliver. admins now have more control over the queue. "hold rules" can be added to mark newly queued messages as "on hold", pausing delivery. rules can be about certain sender or recipient domains/addresses, or apply to all messages pausing the entire queue. also useful for (local) testing. new config options have been introduced. they are editable through the admin and/or account web interfaces. the webapi http endpoints are enabled for newly generated configs with the quickstart, and in localserve. existing configurations must explicitly enable the webapi in mox.conf. gopherwatch.org was created to dogfood this code. it initially used just the compose/smtpclient/imapclient mox packages to send messages and process delivery feedback. it will get a config option to use the mox webapi/webhooks instead. the gopherwatch code to use webapi/webhook is smaller and simpler, and developing that shaped development of the mox webapi/webhooks. for issue #31 by cuu508
2024-04-15 22:49:02 +03:00
msgs, _ := queue.List(ctxbg, queue.Filter{}, queue.Sort{Field: "Queued", Asc: false})
queuedMsg := msgs[0]
if queuedMsg.SMTPUTF8 != expectedSmtputf8 {
t.Fatalf("[%s / %s / %s / %s] got SMTPUTF8 %t, expected %t", mailFrom, rcptTo, headerValue, filename, queuedMsg.SMTPUTF8, expectedSmtputf8)
}
})
}
test(`mjl@mox.example`, `remote@example.org`, "header-ascii", "ascii.txt", false, false, nil)
test(`mjl@mox.example`, `remote@example.org`, "header-ascii", "ascii.txt", true, false, nil)
test(`mjl@mox.example`, `🙂@example.org`, "header-ascii", "ascii.txt", true, true, nil)
test(`mjl@mox.example`, `🙂@example.org`, "header-ascii", "ascii.txt", false, true, &smtpclient.Error{Permanent: true, Code: smtp.C553BadMailbox, Secode: smtp.SeMsg6NonASCIIAddrNotPermitted7})
test(`Ω@mox.example`, `remote@example.org`, "header-ascii", "ascii.txt", true, true, nil)
test(`Ω@mox.example`, `remote@example.org`, "header-ascii", "ascii.txt", false, true, &smtpclient.Error{Permanent: true, Code: smtp.C550MailboxUnavail, Secode: smtp.SeMsg6NonASCIIAddrNotPermitted7})
test(`mjl@mox.example`, `remote@example.org`, "header-utf8-😍", "ascii.txt", true, true, nil)
test(`mjl@mox.example`, `remote@example.org`, "header-utf8-😍", "ascii.txt", false, true, nil)
test(`mjl@mox.example`, `remote@example.org`, "header-ascii", "utf8-🫠️.txt", true, true, nil)
test(`Ω@mox.example`, `🙂@example.org`, "header-utf8-😍", "utf8-🫠️.txt", true, true, nil)
test(`mjl@mox.example`, `remote@xn--vg8h.example.org`, "header-ascii", "ascii.txt", true, false, nil)
}
add a webapi and webhooks for a simple http/json-based api for applications to compose/send messages, receive delivery feedback, and maintain suppression lists. this is an alternative to applications using a library to compose messages, submitting those messages using smtp, and monitoring a mailbox with imap for DSNs, which can be processed into the equivalent of suppression lists. but you need to know about all these standards/protocols and find libraries. by using the webapi & webhooks, you just need a http & json library. unfortunately, there is no standard for these kinds of api, so mox has made up yet another one... matching incoming DSNs about deliveries to original outgoing messages requires keeping history of "retired" messages (delivered from the queue, either successfully or failed). this can be enabled per account. history is also useful for debugging deliveries. we now also keep history of each delivery attempt, accessible while still in the queue, and kept when a message is retired. the queue webadmin pages now also have pagination, to show potentially large history. a queue of webhook calls is now managed too. failures are retried similar to message deliveries. webhooks can also be saved to the retired list after completing. also configurable per account. messages can be sent with a "unique smtp mail from" address. this can only be used if the domain is configured with a localpart catchall separator such as "+". when enabled, a queued message gets assigned a random "fromid", which is added after the separator when sending. when DSNs are returned, they can be related to previously sent messages based on this fromid. in the future, we can implement matching on the "envid" used in the smtp dsn extension, or on the "message-id" of the message. using a fromid can be triggered by authenticating with a login email address that is configured as enabling fromid. suppression lists are automatically managed per account. if a delivery attempt results in certain smtp errors, the destination address is added to the suppression list. future messages queued for that recipient will immediately fail without a delivery attempt. suppression lists protect your mail server reputation. submitted messages can carry "extra" data through the queue and webhooks for outgoing deliveries. through webapi as a json object, through smtp submission as message headers of the form "x-mox-extra-<key>: value". to make it easy to test webapi/webhooks locally, the "localserve" mode actually puts messages in the queue. when it's time to deliver, it still won't do a full delivery attempt, but just delivers to the sender account. unless the recipient address has a special form, simulating a failure to deliver. admins now have more control over the queue. "hold rules" can be added to mark newly queued messages as "on hold", pausing delivery. rules can be about certain sender or recipient domains/addresses, or apply to all messages pausing the entire queue. also useful for (local) testing. new config options have been introduced. they are editable through the admin and/or account web interfaces. the webapi http endpoints are enabled for newly generated configs with the quickstart, and in localserve. existing configurations must explicitly enable the webapi in mox.conf. gopherwatch.org was created to dogfood this code. it initially used just the compose/smtpclient/imapclient mox packages to send messages and process delivery feedback. it will get a config option to use the mox webapi/webhooks instead. the gopherwatch code to use webapi/webhook is smaller and simpler, and developing that shaped development of the mox webapi/webhooks. for issue #31 by cuu508
2024-04-15 22:49:02 +03:00
// TestExtra checks whether submission of messages with "X-Mox-Extra-<key>: value"
// headers cause those those key/value pairs to be added to the Extra field in the
// queue.
func TestExtra(t *testing.T) {
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/mox.conf"), dns.MockResolver{})
defer ts.close()
ts.user = "mjl@mox.example"
ts.pass = password0
ts.submission = true
extraMsg := strings.ReplaceAll(`From: <mjl@mox.example>
To: <remote@example.org>
Subject: test
X-Mox-Extra-Test: testvalue
X-Mox-Extra-a: 123
X-Mox-Extra-:
X-Mox-Extra-x-cANONICAL-z: ok
Message-Id: <test@mox.example>
test email
`, "\n", "\r\n")
ts.run(func(err error, client *smtpclient.Client) {
t.Helper()
tcheck(t, err, "init client")
mailFrom := "mjl@mox.example"
rcptTo := "mjl@mox.example"
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(extraMsg)), strings.NewReader(extraMsg), true, true, false)
tcheck(t, err, "deliver")
})
msgs, err := queue.List(ctxbg, queue.Filter{}, queue.Sort{})
tcheck(t, err, "queue list")
tcompare(t, len(msgs), 1)
tcompare(t, msgs[0].Extra, map[string]string{
"Test": "testvalue",
"A": "123",
"☺": "☹",
"X-Canonical-Z": "ok",
})
// note: these headers currently stay in the message.
}
// TestExtraDup checks for an error for duplicate x-mox-extra-* keys.
func TestExtraDup(t *testing.T) {
ts := newTestServer(t, filepath.FromSlash("../testdata/smtp/mox.conf"), dns.MockResolver{})
defer ts.close()
ts.user = "mjl@mox.example"
ts.pass = password0
ts.submission = true
extraMsg := strings.ReplaceAll(`From: <mjl@mox.example>
To: <remote@example.org>
Subject: test
X-Mox-Extra-Test: testvalue
X-Mox-Extra-Test: testvalue
Message-Id: <test@mox.example>
test email
`, "\n", "\r\n")
ts.run(func(err error, client *smtpclient.Client) {
tcheck(t, err, "init client")
mailFrom := "mjl@mox.example"
rcptTo := "mjl@mox.example"
err = client.Deliver(ctxbg, mailFrom, rcptTo, int64(len(extraMsg)), strings.NewReader(extraMsg), true, true, false)
ts.smtpErr(err, &smtpclient.Error{Permanent: true, Code: smtp.C554TransactionFailed, Secode: smtp.SeMsg6Other0})
add a webapi and webhooks for a simple http/json-based api for applications to compose/send messages, receive delivery feedback, and maintain suppression lists. this is an alternative to applications using a library to compose messages, submitting those messages using smtp, and monitoring a mailbox with imap for DSNs, which can be processed into the equivalent of suppression lists. but you need to know about all these standards/protocols and find libraries. by using the webapi & webhooks, you just need a http & json library. unfortunately, there is no standard for these kinds of api, so mox has made up yet another one... matching incoming DSNs about deliveries to original outgoing messages requires keeping history of "retired" messages (delivered from the queue, either successfully or failed). this can be enabled per account. history is also useful for debugging deliveries. we now also keep history of each delivery attempt, accessible while still in the queue, and kept when a message is retired. the queue webadmin pages now also have pagination, to show potentially large history. a queue of webhook calls is now managed too. failures are retried similar to message deliveries. webhooks can also be saved to the retired list after completing. also configurable per account. messages can be sent with a "unique smtp mail from" address. this can only be used if the domain is configured with a localpart catchall separator such as "+". when enabled, a queued message gets assigned a random "fromid", which is added after the separator when sending. when DSNs are returned, they can be related to previously sent messages based on this fromid. in the future, we can implement matching on the "envid" used in the smtp dsn extension, or on the "message-id" of the message. using a fromid can be triggered by authenticating with a login email address that is configured as enabling fromid. suppression lists are automatically managed per account. if a delivery attempt results in certain smtp errors, the destination address is added to the suppression list. future messages queued for that recipient will immediately fail without a delivery attempt. suppression lists protect your mail server reputation. submitted messages can carry "extra" data through the queue and webhooks for outgoing deliveries. through webapi as a json object, through smtp submission as message headers of the form "x-mox-extra-<key>: value". to make it easy to test webapi/webhooks locally, the "localserve" mode actually puts messages in the queue. when it's time to deliver, it still won't do a full delivery attempt, but just delivers to the sender account. unless the recipient address has a special form, simulating a failure to deliver. admins now have more control over the queue. "hold rules" can be added to mark newly queued messages as "on hold", pausing delivery. rules can be about certain sender or recipient domains/addresses, or apply to all messages pausing the entire queue. also useful for (local) testing. new config options have been introduced. they are editable through the admin and/or account web interfaces. the webapi http endpoints are enabled for newly generated configs with the quickstart, and in localserve. existing configurations must explicitly enable the webapi in mox.conf. gopherwatch.org was created to dogfood this code. it initially used just the compose/smtpclient/imapclient mox packages to send messages and process delivery feedback. it will get a config option to use the mox webapi/webhooks instead. the gopherwatch code to use webapi/webhook is smaller and simpler, and developing that shaped development of the mox webapi/webhooks. for issue #31 by cuu508
2024-04-15 22:49:02 +03:00
})
}