preventing writing out a domains.conf that is invalid and can't be parsed
again. this happens when the last address was removed from an account. just a
click in the admin web interface.
accounts without email address cannot log in.
for issue #133 by ally9335
by explaining (in the titles/hovers) what the concepts and requirements are, by
using selects/dropdowns or datalist suggestions where we have a known list, by
automatically suggesting a good account name, and putting the input fields in a
more sensible order.
based on issue #132 by ally9335
before, an error about an invalid address was not used, causing a delivery
attempt to an empty address (empty localpart/domain). delivery to that address
would fail, but we should've prevented that message from being queued at all.
additionally, an error in adding the message to the queue was ignored too.
an é (e with accent) can also be written as e+\u0301. the first form is NFC,
the second NFD. when logging in, we transform usernames (email addresses) to
NFC. so both forms will be accepted. if a client is using NFD, they can log
in too.
for passwords, we apply the PRECIS "opaquestring", which (despite the name)
transforms the value too: unicode spaces are replaced with ascii spaces. the
string is also normalized to NFC. PRECIS may reject confusing passwords when
you set a password.
both when parsing our configs, and for incoming on smtp or in messages.
so we properly compare things like é and e+accent as equal, and accept the
different encodings of that same address.
by mentioning the dnssec root keys, mentioning which unbound version has EDE,
giving a "dig" invocation to check for dnssec results.
based on issue #131 by romner-set, thanks for reporting
check the authentic data bit for the NS records of "com.", not for ".": some
dnssec-verifying resolvers return unauthentic data for ".".
for issue #139 by triatic, thanks!
transferring the data only once. we only do this when the recipient domains
are the same. when queuing, we now take care to set the same NextAttempt
timestamp, so queued messages are actually eligable for combined delivery.
this adds a DeliverMultiple to the smtp client. for pipelined requests, it will
send all RCPT TO (and MAIL and DATA) in one go, and handles the various
responses and error conditions, returning either an overal error, or per
recipient smtp responses. the results of the smtp LIMITS extension are also
available in the smtp client now.
this also takes the "LIMITS RCPTMAX" smtp extension into account: if the server
only accepts a single recipient, we won't send multiple.
if a server doesn't announce a RCPTMAX limit, but still has one (like mox does
for non-spf-verified transactions), we'll recognize code 452 and 552 (for
historic reasons) as temporary error, and try again in a separate transaction
immediately after. we don't yet implement "LIMITS MAILMAX", doesn't seem likely
in practice.
the current Intn calls in queue could be called concurrently, found by the race
detector with upcoming new tests. best to just prevent any possible concurrent
access.
so you can still know when someone has put you on their blocklist (which may
affect delivery), without using them.
also query dnsbls for our ips more often when we do more outgoing connections
for delivery: once every 100 messages, but at least 5 mins and at most 3 hours
since the previous check.
embarrassingly, we didn't accept all reports we generated. after the changed
handling of reports about mx/mail host vs recipient domains, would send reports
to mail hosts about multiple recipient domains + the mail host. and we included
a policy domain of the mail host when sending to a recipient domain. we were
still being strict in what we accepted: only a single domain in total in the
entire report, and we still enforced that a report sent to the mx host tlsrpt
address only contained the mx host as policy domain. and likewise for recipient
domains and their tls reporting addresses. those checks would reject reports
generated by a mox instance. this probably only happens with dane configured,
probably most users haven't seen it because of that.
somewhat related to issue #125
smtp & imap can only indicate which mechanisms the server software supports.
individual accounts may not have derived secrets for all those mechanisms. imap
& smtp cannot indicate that a client should try another (specific) mechanism.
but at least we shouldn't slow the connection down due to failed auth attempts
in that case.
heard from ben that this is a common source for trouble when setting up email
accounts.
actually, this fix can reduce focus changes for more operations. withStatus is
often used to show an operation in progress in the status bar, only when the
operation isn't done within 1 second. we would restore focus to the element
before the operation started. that was done because we disable elements
sometimes (preventing duplicate form submission). for things like the
autocomplete, with the tab key, which also moves focus to the next element, we
don't want that focus switched back again.
have the full smtp reply in the Diagnostic-Code field, not something that
resembles it but isn't quite the same.
include any additional error message in the Status field as comment.
before, we ended up having an Diagnostic-Code that didn't include the original
smtp code. it only had the enhanced error code.
we weren't properly tracking the cr's and lf's when being strict about message
lines when sending data.
we are reading buffered data from a Reader. if that chunk happens to start with
a newline, we weren't looking at the previously written data, which could be a
cr. instead, in that case, we would always claim the cr/lf wasn't correct.
the new test case triggered the behaviour before having the fix.
should solve issue #129 by x8x, thanks for the report!
we now keep track of the full smtp error responses, potentially multi-line. and
we include it in a dsn in the first free-form human-readable text.
it can have multiple lines in practice, e.g. when a destination mail server
tries to be helpful in explaining what the problem is.
the smtp extension, rfc 4865.
also implement in the webmail.
the queueing/delivery part hardly required changes: we just set the first
delivery time in the future instead of immediately.
still have to find the first client that implements it.
to start composing a message.
the help popup now has a button to register the "mailto:" links with the mox
webmail (typically only works over https, not all browsers support it).
the mailto links are specified in 6068. we support the to/cc/bcc/subject/body
parameters. other parameters should be seen as custom headers, but we don't
support messages with custom headers at all at the moment, so we ignore them.
we now also turn text of the form "mailto:user@host" into a clickable link
(will not be too common). we could be recognizing any "x@x.x" as email address
and make them clickable in the future.
thanks to Hans-Jörg for explaining this functionality.
before, we showed the xn-- ascii names, along with the unicode name. but users
of internationalized email don't want to see any xn-- names. we now put those
in an html title attribute for some cases, so you can still see them if you
really want to, by hovering.
after talking to arnt at fosdem.
broken in previous update. the tricky part keeps being about when browsers fire
'load' and 'hashchange' events for the outer and two inner documents. the
previous change attempted to prevent a history item being set on the first
load. that behaviour seems to be kept.
