useful for new accounts. we don't want to start rejecting incoming messages for
having a score near 0.5 because of too little training material. we err on the
side of allowing messages in. the user will mark them as junk, training the
filter. once enough non-junk has come in, we'll start the actual filtering.
for issue #64 by x8x, and i've also seen this concern on matrix
we'll now get a coverage file artifact for each of the builds. we do two
builds, and the last was likely overwriting the coverage file "artifact" of the
first.
hopefully fixes the test. can't test it locally...
this should be relatively common with setups involving NAT.
so we do warn about it when NAT isn't active since it could highlight potential
misconfiguration.
for issue #239 by exander77
Intended for future use with chatmail servers. Standard email ports may be
blocked on some networks, while the HTTPS port may be accessible.
This is a squashed commit of PR #255 by s0ph0s-dog.
may help users understand when /admin/ isn't enabled on a hostname but the
account web interface is at /. the error will now say: no session for "account"
web interface. it hopefully tells users that their request isn't going to an
admin interface, but ends up at the account web interface.
for issue #268
and mention in a tooltip too that "empty mailbox" only affects messages in the
mailbox, not submailboxes or their messages.
prompted by a question on matrix/irc
mail clients will use these message from addresses also for smtp mail from, so
sending over smtp would fail for these cases. for the webmail and webapi they
already succeeded since we just took the "message from" address as "smtp mail
from" address.
for issue #266 by Robby-, thanks for reporting!
by not trying to parse the full message for the MessageItem, but only reading
headers when needed.
before previous commit, we wouldn't try reading such messages in full either.
i've seen the error a few times:
msgheaderElem.children[(msgheaderElem.children.length - 1)] is undefined
i've seen it happen after sending a reply (with the "answered" flag added).
the updateKeywords callback would render the message again, but the code for
rendering the "additional headers" table rows again was making invalid
assumptions.
the approach is now changed. the backend now just immediately sends the
additional headers to the frontend. before, the frontend would first render the
base message, then render again once the headers came in for the parsed
message. this also prevents a reflow for the (quite common) case that one of
the additional headers are present in the message.
tls servers send a list of certificates for the connection. the first is the
leaf certificate. that's the one for the server itself. that's the one we want
to verify. the others are intermediate CA's. and possibly even the root CA
certificate that it hopes is trusted at the client (though sending it doesn't
make it trusted). with dane-ta, the public key of an intermediate or root CA
certificate is listed in the TSLA record. when verifying, we add any
intermediate/root CA that matches a dane-ta tlsa record to the trusted root CA
certs. we should also have added CA certs that didn't match a TLSA record to
the "intermediates" of x509.VerifyOptions. because we didn't,
x509.Certificate.Verify couldn't verify the chain from the trusted dane-ta ca
cert to the leaf cert. we would only properly verify a dane-ta connection
correctly if the dane-ta-trusted ca cert was the one immediately following the
leaf cert. not when there were one or more additional intermediate certs.
this showed when connecting to mx.runbox.com.
problem reported by robbo5000 on matrix, thanks!
quite some output is printed. you could remember to tee it all to a file. but
that's probably often realized only after having run the quickstart. you can
also copy/paste it all from the terminal, but that's sometimes annoying to do.
writing to a file is more helpful to users.
this has been requested a few times in the past on irc/matrix (i forgot who).
after seeing some junk messages pass the filter, i investigated word counts in
junkfilter.db. i had seen suspicious counts that were just around powers of
two. did not make sense at the time. more investigating makes it clear: instead
of setting new word counts when updating the junk filter, we were adding the
new value to the current value (instead of just setting the new value). so the
counts got approximately doubled when being updated.
users should retrain the junk filter after this update using the "retrain"
subcommand.
this also adds logging for the hypothetical case where numbers would get
decreased below zero (which would wrap around due to uints).
and this fixes junk filter tests that were passing wrong parameters to
train/untrain...
the types in webhook should be subjected to apidiff'ing, this was a shared
function. it is better off in package queue. also change the apidiff script so
it leaves apidiff/next.txt empty when there aren't any changes. makes it easier
to rotate the files after releases where nothing changed (a common occurrence).
these settings are applied anywhere the webmail is open. the settings are for
showing keyboard shortcuts in the lower right after a mouse interaction, and
showing additional headers. the shorcuts were configurable in the "help" popup
before. the additional headers were only configurable through the developer
console before.
the "mailto:" (un)register buttons are now in the settings popup too.
ctrl-l is commonly "focus on browser address bar".
ctrl-u is commonly "view source".
ctrl-I (shift i) is commonly "open developer console".
these keys are more useful to leave for the browser. ctrl-l and ctrl-u (moving
to a message without opening it) can still be had by using also pressing shift.
the previous ctrl-shift-i (show all headers) is now just ctrl-i.
this has been requested in the past on irc/matrix (i forgot who).
the field is optional. if absent, the default behaviour is currently to disable
session tickets. users can set the option if they want to try if delivery from
microsoft is working again. in a future version, we can switch the default to
enabling session tickets.
the previous fix was to disable session tickets for all tls connections,
including https. that was a bit much.
for issue #237
for each message part. The ContentDisposition value is the base value without
header key/value parameters. the Filename field is the likely filename of the
part. the different email clients encode filenames differently. there is a
standard mime mechanism from rfc 2231. and there is the q/b-word-encoding from
rfc 2047. instead of letting users of the webhook api deal with those
differences, we provide just the parsed filename.
for issue #258 by morki, thanks for reporting!
the imap & smtp servers now allow logging in with tls client authentication and
the "external" sasl authentication mechanism. email clients like thunderbird,
fairemail, k9, macos mail implement it. this seems to be the most secure among
the authentication mechanism commonly implemented by clients. a useful property
is that an account can have a separate tls public key for each device/email
client. with tls client cert auth, authentication is also bound to the tls
connection. a mitm cannot pass the credentials on to another tls connection,
similar to scram-*-plus. though part of scram-*-plus is that clients verify
that the server knows the client credentials.
for tls client auth with imap, we send a "preauth" untagged message by default.
that puts the connection in authenticated state. given the imap connection
state machine, further authentication commands are not allowed. some clients
don't recognize the preauth message, and try to authenticate anyway, which
fails. a tls public key has a config option to disable preauth, keeping new
connections in unauthenticated state, to work with such email clients.
for smtp (submission), we don't require an explicit auth command.
both for imap and smtp, we allow a client to authenticate with another
mechanism than "external". in that case, credentials are verified, and have to
be for the same account as the tls client auth, but the adress can be another
one than the login address configured with the tls public key.
only the public key is used to identify the account that is authenticating. we
ignore the rest of the certificate. expiration dates, names, constraints, etc
are not verified. no certificate authorities are involved.
users can upload their own (minimal) certificate. the account web interface
shows openssl commands you can run to generate a private key, minimal cert, and
a p12 file (the format that email clients seem to like...) containing both
private key and certificate.
the imapclient & smtpclient packages can now also use tls client auth. and so
does "mox sendmail", either with a pem file with private key and certificate,
or with just an ed25519 private key.
there are new subcommands "mox config tlspubkey ..." for
adding/removing/listing tls public keys from the cli, by the admin.
needed for upcoming changes, where (now) package admin needs to import package
store. before, because package store imports mox- (for accessing the active
config), that would lead to a cyclic import. package mox- keeps its active
config, package admin has the higher-level config-changing functions.
this allows removing some ugly instantiations of an rng based on the current
time.
Intn is now IntN for our concurrency-safe prng wrapper to match the randv2 api.
v2 exists since go1.22, which we already require.
and add an alerting rule if the failure rate becomes >10% (e.g. expired
certificate).
the prometheus metrics includes a reason, including potential tls alerts, if
remote smtp clients would send those (openssl s_client -starttls does).
inspired by issue #237, where incoming connections were aborted by remote. such
errors would show up as "eof" in the metrics.
