[FIX] queue_job: repoint default_env to SUPERUSER in runjob by dannyadair · Pull Request #923 · OCA/queue

dannyadair · 2026-04-22T03:58:29Z

runjob is declared auth="none", so _auth_method_none pins request.env and transaction.default_env to a uid=None env. env = http.request.env( user=SUPERUSER_ID) only creates a local superuser env, leaving the default_env as uid=None. Any flush that goes through Transaction.flush() -> default_env.flush_all() then recomputes stored fields as uid=None and fails on anything dereferencing self.env.user.

http.request.update_env(user=SUPERUSER_ID) additionally sets transaction.default_env to the superuser env.

Closes#922

OCA-git-bot · 2026-04-22T04:35:29Z

Hi @guewen, @sbidoul,
some modules you are maintaining are being modified, check this out!

runjob is declared auth="none", so _auth_method_none pins request.env and transaction.default_env to a uid=None env. `env = http.request.env( user=SUPERUSER_ID)` only creates a local superuser env, leaving the default_env as uid=None. Any flush that goes through Transaction.flush() -> default_env.flush_all() then recomputes stored fields as uid=None and fails on anything dereferencing self.env.user. `http.request.update_env(user=SUPERUSER_ID)` additionally sets transaction.default_env to the superuser env. ClosesOCA#922

dannyadair · 2026-06-29T04:27:48Z

Rebased after #924 was merged.
That PR removed the db/session.db handling but not the uid=None default_env issue this fixes, so the change is unchanged.

amh-mw

Seems reasonable after reading the source here and https://github.com/odoo/odoo/blob/19.0/odoo/http.py#L1856-L1867:

defupdate_env(self, user=None, context=None, su=None):
""" Update the environment of the current request. :param user: optional user/user id to change the current user :type user: int or :class:`res.users record<~odoo.addons.base.models.res_users.ResUsers>` :param dict context: optional context dictionary to change the current context :param bool su: optional boolean to change the superuser mode """cr=None# None is a sentinel, it keeps the same cursorself.env=self.env(cr, user, context, su)
self.env.transaction.default_env=self.envthreading.current_thread().uid=self.env.uid

~~That said, why make the change? What problem does this avoid?~~ ☕ Unit test that problem?

dannyadair · 2026-06-29T11:40:25Z

The problem is described in #922 and solved by this PR.

amh-mw · 2026-06-29T11:45:16Z

The problem is described in #922 and solved by this PR.

Yep, sorry! Struck through much of my comment already! Still desire a unit test.

dannyadair · 2026-07-01T10:58:14Z

The problem is described in #922 and solved by this PR.
Yep, sorry! Struck through much of my comment already! Still desire a unit test.

ok with the existing tests this one would be weak and almost meaningless. (test_run_job_controller just calling class methods so we'd be mocking and verify that our two line change calls the two-line change), I 'll add something using HttpCase.

dannyadair · 2026-07-01T11:47:35Z

ugh, sorry for the noise. _acquire_job commits which is forbidden in tests. Have to get tricky.

Adds test_run_job_controller_http.py covering OCA#922: - test_runjob_pins_default_env_to_superuser asserts that when _acquire_job is entered, env.transaction.default_env.uid is SUPERUSER_ID -- the state a job would inherit at perform time. - test_runjob_without_updating_default_env_leaves_uid_none patches Request.update_env to a no-op and verifies default_env.uid stays None, isolating default_env repointing as the load-bearing behavior of runjob. Tests use HttpCase so _auth_method_none, session handling, and update_env run for real. _acquire_job is patched to return None so the endpoint exits cleanly without performing DB work (its internal commit is forbidden inside tests).

dannyadair · 2026-07-01T12:02:58Z

added test

amh-mw

Code review only, LGTM.

guewen

Thanks @dannyadair, can you fix the license to match the one of the module please?

Ricardoalso

Thanks !

OCA-git-bot · 2026-07-15T15:06:33Z

This PR has the approved label and has been created more than 5 days ago. It should therefore be ready to merge by a maintainer (or a PSC member if the concerned addon has no declared maintainer). 🤖

Co-authored-by: Guewen Baconnier <guewen@baconnier.me>

sbidoul · 2026-07-20T14:30:45Z

/ocabot merge patch

OCA-git-bot · 2026-07-20T14:30:52Z

Hey, thanks for contributing! Proceeding to merge this for you.
Prepared branch 19.0-ocabot-merge-pr-923-by-sbidoul-bump-patch, awaiting test results.

OCA-git-bot · 2026-07-20T15:25:02Z

Congratulations, your PR was merged at d2c1759. Thanks a lot for contributing to OCA. ❤️

OCA-git-bot added series:19.0 mod:queue_job Module queue_job labels Apr 22, 2026

dannyadair force-pushed the 19.0-fix-runjob-env-user branch from 4a36fc9 to 415420fCompare June 29, 2026 04:25

amh-mw suggested changes Jun 29, 2026
View reviewed changes

dannyadair force-pushed the 19.0-fix-runjob-env-user branch from 588919a to eeea3eeCompare July 1, 2026 11:28

dannyadair force-pushed the 19.0-fix-runjob-env-user branch from eeea3ee to fd214cbCompare July 1, 2026 11:54

amh-mw approved these changes Jul 2, 2026
View reviewed changes

sbidoul mentioned this pull request Jul 2, 2026
[16.0][FIX] queue_job: rebind args/kwargs to inner cursor in in_temporary_env #926
Open

guewen reviewed Jul 9, 2026
View reviewed changes

Comment threadqueue_job/tests/test_run_job_controller_http.py Outdated

guewen reviewed Jul 9, 2026
View reviewed changes

sbidoul mentioned this pull request Jul 15, 2026
[19.0][FIX] queue_job: repair default_env in runjob #947
Closed

Ricardoalso approved these changes Jul 15, 2026
View reviewed changes

OCA-git-bot added approved ready to merge labels Jul 15, 2026

Adjust license from AGPL to LGPL
ffed861
Co-authored-by: Guewen Baconnier <guewen@baconnier.me>

SilvioC2C approved these changes Jul 16, 2026
View reviewed changes

grindtildeath approved these changes Jul 16, 2026
View reviewed changes

MiquelRForgeFlow approved these changes Jul 17, 2026
View reviewed changes

MiquelRForgeFlow mentioned this pull request Jul 17, 2026
[19.0][FIX] edi_core_oca: use savepoint to prevent state corruption on errors OCA/edi-framework#316
Merged

OCA-git-bot added the bot is merging ⏳ label Jul 20, 2026

OCA-git-bot merged commit d6e7702 into OCA:19.0Jul 20, 2026
7 checks passed

OCA-git-bot added merged 🎉 and removed bot is merging ⏳ labels Jul 20, 2026

sbidoul mentioned this pull request Jul 20, 2026
[19.0] env.user empty during flush/recompute inside queue jobs #922
Closed

Uh oh!

Uh oh!

Conversation

dannyadair commented Apr 22, 2026

Uh oh!

OCA-git-bot commented Apr 22, 2026

Uh oh!

dannyadair commented Jun 29, 2026

Uh oh!

amh-mw left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dannyadair commented Jun 29, 2026

Uh oh!

amh-mw commented Jun 29, 2026

Uh oh!

dannyadair commented Jul 1, 2026

Uh oh!

dannyadair commented Jul 1, 2026

Uh oh!

dannyadair commented Jul 1, 2026

Uh oh!

amh-mw left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

guewen left a comment

Choose a reason for hiding this comment

Uh oh!

Ricardoalso left a comment

Choose a reason for hiding this comment

Uh oh!

OCA-git-bot commented Jul 15, 2026

Uh oh!

sbidoul commented Jul 20, 2026

Uh oh!

OCA-git-bot commented Jul 20, 2026

Uh oh!

Uh oh!

OCA-git-bot commented Jul 20, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

9 participants

amh-mw left a comment •
edited
Loading