Closed engineerjoe440 closed 1 year ago
Thanks, @engineerjoe440!
Let me take a look into those test failures. (Actually I'll have to get back to you next week, sorry 'bout that...)
Don't be sorry! That's how this stuff goes sometime. Let me know if there's anything you find that I can change to help get fixes in place all in one shot. :)
@engineerjoe440 much appreciated 👍
I did some CI fixes in #49 - Python 3.8 failing is a bit of a mystery, but we're going to ignore it for now. Would you mind rebasing this PR?
Absolutely! Will do! Thanks @klauer
Merging #48 (ea32b88) into master (336831c) will increase coverage by
0.2%
. The diff coverage is100.0%
.
Well, I can see the same failures when I don't have apischema
installed. I do see everything load properly when I run the tests locally (with apischema
installed, mind you). So, I wonder if it might be beneficial to have that module installed in the 3.8 test? Maybe that could be broken out as a separate test for its uniqueness?
Thoughts?
Thanks again, @klauer :tada: !!!!
Well, I can see the same failures when I don't have
apischema
installed.
Hmm, can you paste the failure output? What I see in CI is it failing on import of apischema
- which is installed there, by the way, as a dev dependency in requirements-dev.txt
Edit: also, I'm going to push one additional test and then merge this. It looks good to me. PLC programmers may balk at the idea of "CONTINUE" outside of a loop, but I think that check could/should happen at a higher level and not fail during parsing outright, I think.
Sorry, I might have made that even more confusing. I was seeing the apischema
errors when I didn't have that package installed. Notably, I'm running on Python 3.11, so it's a bit odd.
Thank you, @klauer!!! This is very exciting, and will free up some of my work on a linter! :heart: :tada:
Your continued contributions are very much appreciated! Glad it has been of use to you. 👍
Ah, I see - the errors are a bit different from what you pasted.
The test suite requires apischema to be run completely, so those errors you pasted above are legitimate. The docs are very likely lacking, but requirements-dev.txt
is considered a necessity for running the full test suite.
The weird/unexpected failure is the Python 3.8 job which shows:
So, an on-import failure inside of apischema before anything blark gets run. That makes it extremely likely to be an upstream issue, but it's difficult to confirm without being able to reproduce just yet.
Related Issues:
47
This should add support of the
CONTINUE
statement.Notes: Seems that the tests are failing, but they appear to have been failing prior to changes, as well. Open to thoughts, there!