Age | Commit message (Collapse) | Author |
|
We do reach that timeout on production for large series. Give a bit more
leeway.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Patches are sometimes cross posted on mailing-lists handled by the same
patchwork instance, so msgids aren't unique across projects anymore.
This fixes:
Traceback (most recent call last):
File "patchwork/bin/parsemail.py", line 883, in main
return parse_mail(mail)
File "patchwork/bin/parsemail.py", line 761, in parse_mail
content = find_content(project, mail)
File "patchwork/bin/parsemail.py", line 423, in find_content
ret.patch_order, n, refs)
File "patchwork/bin/parsemail.py", line 533, in find_series_for_mail
previous_patch = find_previous_patch(revision, order, refs)
File "patchwork/bin/parsemail.py", line 468, in find_previous_patch
patch = Patch.objects.get(msgid=ref)
File "django/db/models/manager.py", line 127, in manager_method
return getattr(self.get_queryset(), name)(*args, **kwargs)
File "django/db/models/query.py", line 338, in get
(self.model._meta.object_name, num)
MultipleObjectsReturned: get() returned more than one Patch -- it returned 2!
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Fedora package is GitPython, not python-GitPython
|
|
|
|
Just like in the series case, we'll use the filtering from the web page
and git-pw.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
To be consistent with what has been done for series, even it's not
entirely pure.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
We're going to have more than one list of patches (just like series).
Extract common properties into a mixin.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Just like with series, it's handy to be able to query patches that have
changed since the last time we looked.
The migration is a bit complicated as we need to populate the initial
value of that field, ensuring the invariant that this field cannot be
NULL.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
This field is also shared between the SeriesList viewsets.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
SeriesListMixin already has that field, no need to override it.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
We'll need that for admin support.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
We've been caught a few times with pushing commits where the state of
manage.py makemigrations wasn't clean and had pending migrations. Make
sure it doesn't happen again.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
The migration was created with a different order of the TestState
fields. Everything worked as intented, but makemigrations was yelling
that we had a pending migration because of a different order in the
choice field.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Unfortunately we cannot apply patches from the getpatchwork branch, so I
needed to redo that one.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
v2: Port to fdo's patchwork (Damien)
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
|
|
patches which descripes renamed files show different hashes with git
show and pwclient. Use git show -C (Detect copies as well as renames.)
Signed-off-by: Jan Remmet <j.remmet@phytec.de>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
It's already obvious that the user has logged out, so don't bother
displaying a message saying as much. Redirect to the home page instead.
v2: Port to fdo's patchwork (Damien)
Signed-off-by: Stephen Finucane <stephen.finucane@intel.com>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Since the link to TODOs (reviews pending) can already be seen there,
it makes sense to link to bundles also.
A divider is added to logically group these elements.
v2: Port to fdo's patchwork (Damien)
Signed-off-by: Stephen Finucane <stephen.finucane@intel.com>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
While we don't have an API to delete a test result at the moment, it's
future proof to fix this straight away. That means we can also use it
when tinkering with the db from manage.py shell for instance.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
We don't really want a big cross for info test results being sent out as
emails, a circle seems neutral enough to me.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Nothing major, the most annoying thing was the argument parsing
rejecting the 'info' state for filtering.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Signed-off-by: Michael Wood <michael.g.wood@intel.com>
|
|
Adds the front end parts to display a test which has a result as 'info'.
Signed-off-by: Michael Wood <michael.g.wood@intel.com>
|
|
Signed-off-by: Michael Wood <michael.g.wood@intel.com>
|
|
Sometimes tests return information which is neither a fail or success for
example statistics or linters which are useful to manually review the output.
Signed-off-by: Michael Wood <michael.g.wood@intel.com>
|
|
People are asking our CI team to re-trigger tests when something goes
wrong (for instance when a test slave was online during the previous
test run) or to double check that the results are stable.
This action allows maintainers to re-trigger test runs by inserting
series-new-revision events into the event stream.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Depending on what we want to do, we need POST and PATCH requests.
Introduce a POST one, rewriting the patch_data() as a function of
post_data() (we use a header to override the POST method to a PATCH).
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
We're doing a PATCH request, not a POST one!
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
For some API calls, we not only need the series id but also the latest
revision.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
From a checkbox (that's the element we receive the event from), it's
handy to be able to retrieve the row data. That function does it, given
the column header we want.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
It's useful to have more metadata about the table columns in one place
so we can query various things about them. This time, I need the rank of
the column so I can query data from the name of column header.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
We'll need the version for API calls on the correct revision, so store
it in the table.
We'll also be able, later, to let people select which column to
show/hide to personalize which date one would like to see.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
We're going to have another type of action that doesn't use a menu/form,
but we want to reuse that code.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
People would like to be able to re-trigger testing through the web
interface. Create a new series-new-revision event should be enough to
trick the testing systems to re-trigger a test run.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
git-pw: fix issue with python-requests
|
|
|
|
git-pw: print error messages on stderr
|
|
docs: Update default fixtures list s/default_actions/default_events
|
|
$ git pw list -s `date +%F -d '-1 day'` --json
No series found!
Non-json message is unexpected in this case.
|
|
Signed-off-by: Michael Wood <michael.g.wood@intel.com>
|
|
Ander noticed we didn't actually recurse, and that
commit 74039877a755711bea370e200984acbb2334b07c
Author: Damien Lespiau <damien.lespiau@intel.com>
Date: Wed Nov 11 12:42:05 2015 +0000
git-pw: Make git-pw work with older versions of GitPython
was too hopeful in its expectations (I swear it was true at the time?
but may have been fooled).
v2: Use git.__version__ to decide when to include the
search_parent_directories named argument.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
|
|
We can now add Patchwork links to commit messages!
$ git am -s -l $series_id
$ git am -s -l --link-name Link $series_id
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Fixes: https://github.com/dlespiau/patchwork/issues/176
|