the description sounds neat but I can't think of a use case. Can you give me one?
Eldar Abusalimov
@abusalimov
I encountered few complains on pytest-capturelog's performance overhead when I was researching pytest mailing list archives. And I have some ideas on how to make it faster, but one can't just claim smth. is "faster" without proofs on his hands :smile:
Oh, and just for fun, of course!
Arthur Skowronek
@eisensheng
Still reading through the changeset but it looks good to me and I'm in favor of merging this. I find it a little bit amusing that we're back to tox on travis with this PR :wink2:
Eldar Abusalimov
@abusalimov
Yeah, but I resisted, TBH :smile:
Arthur Skowronek
@eisensheng
is the before_install script a template you found elsewhere?
and I'm wondering a little bit which python version will be chosen by tox to run the tests. Will it be the version that's being used with the container?
Hum
I'll comment this on the PR.
ah, this is what the script does, neat. Nevermind what I was babbling above.
Eldar Abusalimov
@abusalimov
No, I came up with it myself. It infers TOXENVfrom a Python version running inside the container, yes. And tox respects TOXENV
I decided to make it more explicit and verbose (+ announce() stuff)
Arthur Skowronek
@eisensheng
yes, this is also a favorable approach
Arthur Skowronek
@eisensheng
other than this I can't find anything else aside from the evil usage of eval but oh well. :shipit:
Eldar Abusalimov
@abusalimov
:smile: I'm not who invented this, pytest itself uses a similar eval trick within its marker handling logic.
Thank you!
Eldar Abusalimov
@abusalimov
OK, everything seems to work, I'll rebase and merge then
Arthur Skowronek
@eisensheng
ok. :+1:
Eldar Abusalimov
@abusalimov
Hi, could someone take a look at #42, please? Nothing interesting there, but it hinders some more useful changes...
Michael Aquilina
@MichaelAquilina
Hi, is pytest-catchlog basically a near drop in replacement for pytest-capturelog? I have a pytest warning about a deprecated argument. Would changing to this package remove that?
Eldar Abusalimov
@abusalimov
@MichaelAquilina pretty much. That deprecation warning is fixed, though catchlog itself might warn about legacy usage of caplogfixture, like caplog.records() should be replaced with plain caplog.records, etc.
_
Tibo Beijen
@TBeijen
Discovered there's a gitter right after creating a PR, otherwise I'd discuss here first.
Anyway, I just added eisensheng/pytest-catchlog#54 , thoughts most welcome.
Johann Kellerman
@kellerza
Hi, is there any way to stop logger messages from the std-out print portion? All messages seem to be duplicated between std-out and the catchlog output
Hi, did you plan to release a new version of this wonderful plugin pytest-catchlog? There is lots of new features but not yet available in pypi.. :(
Florian Bruhin
@The-Compiler
Hey @abusalimov and @eisensheng, we haven't talked in ages!
I took the freedom of closing some issues and PRs now that the pytest-catchlog functionality has been merged to the pytest core, I hope that's fine with you