This feature is off by default and can can be configured with the `email_total_attachment_size_limit_kb` site setting.
Co-authored-by: Maja Komel <maja.komel@gmail.com>
This reduces chances of errors where consumers of strings mutate inputs
and reduces memory usage of the app.
Test suite passes now, but there may be some stuff left, so we will run
a few sites on a branch prior to merging
Minor fixes to add Rails 6 support to Discourse, we now will boot
with RAILS_MASTER=1, all specs pass
Only one tiny deprecation left
Largest change was the way ActiveModel:Errors changed interface a
bit but there is a simple backwards compat way of working it
We had quite a few cases in core where inputs are being mutated as a side
effect of calling a method.
This handles all the cases where specs caught this.
Mutating inputs makes code harder to reason about. Eg:
```
frog = "frog"
jump(frog)
puts frog
"fly" # ?????
```
This commit is part of a followup commit that adds # frozen_string_literal
to all our specs.
Includes support for flags, reviewable users and queued posts, with REST API
backwards compatibility.
Co-Authored-By: romanrizzi <romanalejandro@gmail.com>
Co-Authored-By: jjaffeux <j.jaffeux@gmail.com>
- The test_email job is removed, because it was always being run synchronously (not in sidekiq)
- 34b29f62 added a bypass for critical emails, to match the spec. This removes the bypass, and removes the spec.
- This adapts the specs for 72ffabf6, so that they check for emails being sent
- This reimplements c2797921, allowing test emails to be sent even when emails are disabled
Migrates email user options to a new data structure, where `email_always`, `email_direct` and `email_private_messages` are replace by
* `email_messages_level`, with options: `always`, `only_when_away` and `never` (defaults to `always`)
* `email_level`, with options: `always`, `only_when_away` and `never` (defaults to `only_when_away`)
We can only be sure that an email is sent when we get a mailer in
`ActionMailer::Deliveries`. A couple of tests were actually incorrect
because it didn't flow through our email sender where there are more
conditions in determining whether an email is sent or not.
If you reply to an email with the word "mute" a topic will be muted
If you reply to an email with the word "track" a topic will be tracked
If you reply to an email with the word "watch" a topic will be watched
These ninja command can help advanced mailing list ex-users, saves a trip
to the website
* Enable user email PM when posting to group or replying to topic via email
* remove extra line
* Add test and fix snake_case
* Only reenable email_private_messages for PM replies
* FEATURE: add indication if incoming email attachment was rejected and inform sender about it
* include errors for rejected attachments in email
* don't send warning email to staged users
* use user object instead of user_id in add_attachments method
* Add possibility to add hidden posts with PostCreator
* FEATURE: Create hidden posts for received spam emails
Spamchecker usually have 3 results: HAM, SPAM and PROBABLY_SPAM
SPAM gets usually directly rejected and needs no further handling.
HAM is good message and usually gets passed unmodified.
PROBABLY_SPAM gets an additional header to allow further processing.
This change addes processing capabilities for such headers and marks
new posts created as hidden when received via email.
This updates tests to use latest rails 5 practice
and updates ALL dependencies that could be updated
Performance testing shows that performance has not regressed
if anything it is marginally faster now.
Ignores the site setting "find_related_post_with_key" and always tries to honor the `In-Reply-To` and `References` header for emails sent to a group.
The senders email address must be included in the `To` or `CC` header of a previous email sent to the group and the `Message-ID` of that email must be included in the current email's `In-Reply-To` or `References` header.
* `rescue nil` is a really bad pattern to use in our code base.
We should rescue errors that we expect the code to throw and
not rescue everything because we're unsure of what errors the
code would throw. This would reduce the amount of pain we face
when debugging why something isn't working as expexted. I've
been bitten countless of times by errors being swallowed as a
result during debugging sessions.
This change allows email-clients to show threaded views of mails as
expected. Apparently most algorithms expect the message ids of mails
in the Reference-header-field to be sorted such that they build a
traversal through the thread, so the oldest (original) message being
first, then its child, grandchild and so on until it arrives at the
message id that the "new" mail (that is to be sent) is the reply to.
MSGA [1]
+- Re: MSGA [1-1]
| +- Re: Re: MSGA [1-2-1]
| +- Re: Re: MSGA [1-2-2]
+- Re: MSGA [1-1]
If the stuff in brackets would be the message ID, the References-Header
field of a message that is a reply to [1-2-1] should look like:
References: 1, 1-1, 1-2-1
Discussion took place in:
https://meta.discourse.org/t/e-mail-threading-in-ml-mode-does-not-work-in-thunderbird
Main information taken from:
https://www.jwz.org/doc/threading.html
* store time it took to index message in DB (to find performance issues)
* ignore listserv specific files
* better examples for split_regex
* first email in mbox shouldn't contain the split string
* always lock the DB in exclusive mode
* save email within transaction
* messages can be grouped by subject and use original order (for Listserv)
* adds option to index emails without running the import
Although 407a23663d will send rejection
messages for unrecognized errors, sometimes processing the email will
raise an error which has a blank message.
This commit:
1. Shows rejected emails which have already been processed and contain
a blank error in /admin/email/rejected
2. Replaces new blank error messages with the error type
The mail class seems to handle mails sent with Content-Transfer-Encoding: 8bit
somewhat weirdly: It decodes them (to utf-8), changes the raw source to base64,
and does not modify the Content-Type:charset= header.
This leads to Discourse trying the message encoding (in my example ISO-8859-1)
first, and if that does not contain any unparseable characters, it uses that.
Sadly, in ISO-8859-1, every byte sequence is valid.
Fix this by always trying to decode as UTF-8 first. The probability of someone
using another encoding that cleanly (but wrongly) decodes as UTF-8 should be
fairly low.
