| Commit message (Collapse) | Author | Age | Files | Lines |
| | |
|
| |
|
|
| |
If the author of a patch series provides a base-commit tag, extract and store the hash.
|
| | |
|
| |
|
|
|
| |
Accept empty lines as part of a diff.
Fixes #6403
|
| |
|
|
|
|
|
|
|
|
|
| |
On LKML, users sometimes reply to a patch series while keeping the
original title. At the same time, sometimes new series are sent as a
reply to the previous versions. It all confuses the lore archive
parsing logic.
When reading the archive, remember if the patch were present. If not,
mark the series as corrupted and thus prevent is appearance in
syz-cluster.
|
| |
|
|
|
| |
Wrapping the email.Email object will let us add lore-specific fields to
it at a later point.
|
| |
|
|
|
| |
It's not done transparently by the email library.
Add a test that verifies the result.
|
| |
|
|
|
| |
Sometimes the patch series directly hint at the kernel tree they should
be applied to. Extract and remember this information.
|
| |
|
|
| |
./tools/syz-env bin/golangci-lint run ./... --fix
|
| |
|
|
| |
These will be necessary to properly route emails in dashboard.
|
| |
|
|
| |
This will allow us to easily identify them.
|
| |
|
|
|
|
| |
For some specified inboxes, forward the emails that contain syz
commands.
Add tests to verify the behavior.
|
| |
|
|
|
|
|
|
|
|
| |
Rename the method to LatestCommit and make it more flexible:
1) Return the commit date alongside the commit hash.
2) Rename the time filter to highlight that it's non-inclusive.
3) Make it possible to query the commits newer than the specified commit
hash.
It will let us poll lore archives more efficiently.
|
| | |
|
| |
|
|
|
| |
Refactor the code to make it more reusable.
Add a method to extract specifically the list of new patch series.
|
| |
|
|
| |
Support filtering by the commit date.
|
| |
|
|
|
|
| |
It changed recently and it broke our email discussion link extraction.
Closes #5431.
|
| |
|
|
| |
This is not needed for the actual operation.
|
| |
|
|
|
| |
Fix checking of Logf, it has string in 0-th arg.
Add checking of t.Errorf/Fatalf.
|
| |
|
|
|
| |
We used to support this only for external reportings, but let's
auto-guess these parameters for all incoming patch testing commands.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Let bug labels come in three flavours:
1) Bug labels with multiple values (e.g. `subsystems`).
2) Bug labels with only one value (e.g. `prio`).
3) Flags.
Let users configure bug labels via email by issuing the following
commands:
#syz set subsystems: abc, def
#syz set no-reminders
#syz unset no-reminders
Also let users set tags for invididual bugs in reported bug lists:
#syz set <1> some-tag
|
| |
|
|
|
|
|
|
|
|
| |
It's not always convenient that we can receive only one command per
email.
Update pkg/email parsing code to extract everything which there is.
Update reporting_email.go to handle multiple email commands. Set
different limits for bug email commands and bug list commands.
|
| | |
|
| |
|
|
|
|
| |
Patches are not always spelled upper case. Also, looking for just PATCH
may lead to false positives, so first put the subject to lower case and
then look for the "[patch" substring.
|
| |
|
|
|
|
| |
DiscussionReport are the discussions started by syzbot.
All other discussions that Cc syzbot or mention its bugs are
DiscussionMention.
|
| |
|
|
|
| |
It seems to be a common pattern in LKML: bug fixes are sent as replies
to syzbot reports. Recognize such replies as separate discussions.
|
| |
|
|
|
|
|
|
|
|
|
| |
Now that too much logic seems to be duplicated in tools/syz-lore and
dahsboard/app/discussion.go, it's time to refactor the code.
Factor out the code that decides whether an incoming email message
should start a new thread or be appended to the previous one.
Restructure pkg/email/lore so that email processing matches the
one-by-one approach of reporting_email.go.
|
| |
|
|
| |
Adjust tests.
|
| |
|
|
|
| |
We anyway already extract this information, there's no sense in
extracting it again in other places.
|
| |
|
|
|
| |
Even though the standard seems to prohibit it, there are real world
cases of messages with multiple IDs in an "In-Reply-To" header.
|
| |
|
|
|
| |
bufio.Scanner has a limit on the maximum line size. Don't panic in this
case.
|
| |
|
|
| |
There's no need to re-initialize them every time.
|
| |
|
|
|
|
| |
We need to
1) Query raw messages from an LKML git archive.
2) Group raw messages into threads.
|
| |
|
|
|
|
|
|
|
| |
Extract Date and InReplyTo.
Extract not just one, but multiple BugIDs from multiple sources:
1) Email addresses in From/To/Cc.
2) Email addresses in Body.
3) Dashboard links in Body.
|
| |
|
|
|
|
| |
If the `#syz regenerate` command is sent in response to a bug list,
dashboard will schedule its regeneration the next time the corresponding
cron job is run.
|
| |
|
|
| |
For now this command will be used to manually set bug's subsystems.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
Currently the function returns the values that are not being used
anywhere. Refactor it.
|
| |
|
|
| |
It look like email clients can insert newlines there if the line is too long.
|
| |
|
|
|
|
| |
Handle two types of mailing list responses:
* From contains the original author.
* From is replaced by the mailing list.
|
| |
|
|
|
|
|
| |
Sometimes we call it before accessing emails, sometimes we do not. This
is prone to errors.
Always canonicalize From and Cc, this should somewhat reduce the problem.
|
| | |
|
| |
|
|
|
|
| |
There was a precedent of using:
syz fix: repo \t commit
This was rejected as error. Support \t between tokens.
|
| |
|
|
|
| |
Add "#syz unfix" command that undoes "#syz fix" effects
and resets any existing fixing commits.
|
| | |
|
| |
|
|
|
|
|
|
| |
Some terms are normalised on the technical level
but may be oppressive on a societal level.
Replace them with more technically neutral terms.
See the following doc for a longer version:
https://tools.ietf.org/id/draft-knodel-terminology-00.html
|
| |
|
|
|
| |
Some users spell the command as "#syz-dup:".
Support this and few more variations.
|