requirements-test.txt: Drop no longer needed test dependency
It's already pulled from swh.core[db] which is a runtime dependency.
Related to swh/infra/sysadm-environment#2746 (closed)
Test Plan
tox
Migrated from D4554 (view on Phabricator)
Merge request reports
Activity
Build is green
Patch application report for D4554 (id=16159)
Could not rebase; Attempt merge onto 29db82c6...
Updating 29db82c..42ac003 Fast-forward requirements-test.txt | 2 +- swh/vault/tests/test_server.py | 16 ++++++++-------- 2 files changed, 9 insertions(+), 9 deletions(-)Changes applied before test
commit 42ac00365bdc9495de52a4af4b5d7196ec90b796 Author: Antoine R. Dumont (@ardumont) <ardumont@softwareheritage.org> Date: Mon Nov 23 09:56:27 2020 +0100 requirements-test.txt: Use correct test dependency swh.core[db] Instead of pytest-postgresql. Related to swh/infra/sysadm-environment#2746 commit 516e5956c9f5d47f979ecc17b98a529ea58542cf Author: Antoine R. Dumont (@ardumont) <ardumont@softwareheritage.org> Date: Mon Nov 23 09:48:48 2020 +0100 test_server: Fix exception structure swh.core (0.9.0) changed the remote exception serialization handling. So the current build [1] is broken as it needs adaptation to retrieve the exception information at its first level. This adapts accordingly the exception handling in those failing tests. - [1] https://jenkins.softwareheritage.org/job/DVAU/job/tests/881/See https://jenkins.softwareheritage.org/job/DVAU/job/tests-on-diff/54/ for more details.
! In !74 (closed), @douardda wrote:
swh.core[db]is already a dep inrequirements-swh.txt, why would is be needed here?But more importantly, I disagree with the validity of the diff because pytest-posgtgresql is used //directly// in this package (in
conftest.py), not only by the mean of theswh.core.tests.dbpackage.(I wrote a bit too fast, I missed the imported postgresql factory from
swh.core.pytest_pluginhowever) the fixture from pytest-postgresql is also used as is. So the dependency is not (only) transitive, so IMHO the dependency should stay.
But more importantly, I disagree with the validity of the diff because pytest-posgtgresql is used directly in this package (in conftest.py),
no, it's not used directly.
(I wrote a bit too fast, I missed the imported postgresql factory from swh.core.pytest_plugin however)
heh, that ^ for the previous point, ok.
the fixture from pytest-postgresql is also used as is.
no, that fixture adds the
resets those tables behaviorwhich is not what's the original fixture does.So the dependency is not (only) transitive, so IMHO the dependency should stay.
I disagree because see my previous point.
I'm trying hard to make sense of our spaghetti deps here... And i'm trying to make it have some sense both here and in the debian package dimension...
swh.core[db]allows, in the python dimension to expose the swh.core.db.pytest_plugin test fixture, thus why it's declared explicitely here.In the debian dimension, the build dependency that needs to be used for my last sentence is python3-swh.core.db.pytestplugin (swh-scheduler!160 (closed) for an example on scheduler).
Note that for the requirement tests, i hesitated to add the
swh.core[testing]instead ofswh.core[db]but i'm unsure that it makes more sense... [1]-
Also, debian package wise, if I don't drop that dependency here, it makes my previous work on the debian package from core moot. As it will pull again the python3-pytest-postgresql (and thus postgresql). This time, not from the core package but by the vault package...
And following the current logic, this ^ wrong behavior will be true for other packages i started fixing (scheduler, ...).
-
Also, debian package wise, if I don't drop that dependency here, it makes my previous work on the debian package from core moot. As it will pull again the python3-pytest-postgresql (and thus postgresql). This time, not from the core package but by the vault package... And following the current logic, this ^ wrong behavior will be true for other packages i started fixing (scheduler, ...).
Ah no, great, it's not to a problem because it's a test dependency here, not a runtime one As per swh-scheduler!160 (closed)'s test plan shows.
My previous comment still holds though.
Some references in the commit message have been migrated:
- T2746 is now swh/infra/sysadm-environment#2746 (closed)
-
Build is green
Patch application report for D4554 (id=16178)
Rebasing onto 516e5956...
Current branch diff-target is up to date.Changes applied before test
commit db8f2ab792ab2802a2d0e0ac13dce7f407a4f936 Author: Antoine R. Dumont (@ardumont) <ardumont@softwareheritage.org> Date: Mon Nov 23 09:56:27 2020 +0100 requirements-test.txt: Drop no longer needed test dependency It's already pulled from swh.core[db] which is a runtime dependency. Related to swh/infra/sysadm-environment#2746See https://jenkins.softwareheritage.org/job/DVAU/job/tests-on-diff/55/ for more details.
-
Build is green
Patch application report for D4554 (id=16179)
Rebasing onto 516e5956...
Current branch diff-target is up to date.Changes applied before test
commit 96fb83830a2c0e3f42a0c47c0e17c85e81ed752e Author: Antoine R. Dumont (@ardumont) <ardumont@softwareheritage.org> Date: Mon Nov 23 09:56:27 2020 +0100 requirements-test.txt: Drop no longer needed test dependency It's already pulled from swh.core[db] which is a runtime dependency. Related to swh/infra/sysadm-environment#2746See https://jenkins.softwareheritage.org/job/DVAU/job/tests-on-diff/56/ for more details.
