New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Do not rely on the version number to know if the tests are present and fixing test matrix #256
Conversation
Seems the tests have been failing for a bit now. |
I also added the ability to overwrite the path of the pylint test folder using the |
@aerostitch your patch looks good and TBH I was about to approve it and merge it. However the reason tests are failing now is that the classes we consume from I am +1 about the idea to be able to configure the additional import path but maybe you can make it a bit more flexible and try the import from 2 different places. While typing this I realize that the last paragraph can be a simple import within try/except block below the current changes. What do you think? Care to make Travis happy again ? |
The compat module is here for that very reason, the dependencies of pylint_django are very volatile so things like this crop up a lot, so if possible consider adding to: |
I know it sucks and is ugly but you gotta do what you gotta do. |
Thanks for the review guys. @atodorov I think we only use test_functional from pylint's test suite so I'll look into making the import in a try/catch to manage that and rename the environment variable to be more explicit. |
…ta do it the other way
…nc with the installed pkg
… to serve the transition
…int 2.4 when it should not
So I've been able to have the tests pass.
Right now I'm facing an issue when it seems that even with pylint 2.5 installed it looks for pylint 2.4 code which is weird, so still looking into it but having trouble reproducing locally. Seems like something related to caching on travis or something like that. |
So I placed the following line right after the print('DEBUG: pylint version: {}'.format(pylint.__version__)) And in the
See: https://travis-ci.org/PyCQA/pylint-django/jobs/616459641?utm_medium=notification |
e862fc5
to
7f52a76
Compare
7f52a76
to
da2146c
Compare
Alright, we have the tests installation setup properly now. Now we have some tests failing on the master branch of pylint, so I'll have a look at those now but at least we test with what we're expecting to be testing. |
I wonder if the error we get wouldn't be due to an issue in pylint:
|
Seems it's not really that.
|
Reviewing all the test cases that fail on travis with pylint 2.5 is basically all test cases that are not a |
@aerostitch thanks for your efforts. This appears to be more messier that it should. I will look at it in the next few days b/c I need to understand what's going on here before doing anything else. I was secretly hoping the change will be smaller. |
Thanks @atodorov Trying to do a summary so that's easier: 1. Lots of commits - can be squashed or separated in another PR There are a lot of commits mainly because I couldn't reproduce some of the issues locally and a this build was already breaking before my change so understand what was the origin of the breakages took a lot of tries. I can squash the commits if it's easier to understand (I was thinking this full PR could be squashed) or I can create a separate PR to fix the issue I've seen with the build matrix (not talking about the tests, just the build matrix). 2. Problems currently fixed in the PR The issues are divided in the following pools:
3. What's left to do? - Help needed on that The current status is that the tests overall are running with the expected version of pylint and only the ones related to the pylint master branch fail. Those tests fail due with an |
Hi guys,
Sometimes the distro re-add the files that were removed from the pip packages and in this case it will probably be the case in a few days, so maybe it would be better to check if the test directory exists instead of relying on the version. What do you think?
Thanks for your help,
Joseph