Remove @NoEffect annotations#1677

EliahKagan

Along with removing all @NoEffect annotations (there were only a handful), this:

Adds missing asserts, where an expression statement was by itself that was intended as an assertion. This turned out to be the case in the places @NoEffect appeared in rendered documentation, making it so no per-file-ignores or other broadened suppressions were needed.
Fixes misspellings (including one case affecting documentation).
Adds a FIXME comment for investigating a free-standing expression statement with no obvious side effects that may have been meant as an assertion but would fail if turned into one.

There are some other commented @-style suppression annotations throughout the project. This PR deliberately does not touch those. Some of them express useful intent that should eventually be stated in a different way but currently could be a breaking change to express better. For example, some of them express that a name that is imported but not used is intended to be present, which is better expressed with a static __all__ listing exactly the names (imports and otherwise) meant to be accessed through the module, but to usefully express this an __all__ needs to omit names not intended to be accessed that way, which may be a breaking change.

Byron

Thanks a lot!

I wonder why @NoEffect was chosen back then over making it an assertion in the first place. It's good that this is settled now though.

If made into an assertion it would fail, I wonder if this means that there is a bug in the submodule implementation or the assertion is simply wrong to begin with. Maybe it's an assertion that doesn't work similarly on all platforms?

I'm not fully clear on what is expected to happen. In context, we have:

GitPython/test/test_submodule.py

Lines 107 to 114 in b27a89f

    
           # some commits earlier we still have a submodule, but its at a different commit 
        
           smold = next(Submodule.iter_items(rwrepo, self.k_subm_changed)) 
        
           assert smold.binsha != sm.binsha 
        
           assert smold != sm  # the name changed 
        
           # force it to reread its information 
        
           del smold._url 
        
           smold.url == sm.url  # noqa: B015  # FIXME: Should this be an assertion?

When it is made into an assertion, pytest shows:

E AssertionError: assert 'git://gitorious.org/git-python/gitdb.git' == 'https://github.com/gitpython-developers/gitdb.git' E - https://github.com/gitpython-developers/gitdb.git E + git://gitorious.org/git-python/gitdb.git test/test_submodule.py:114: AssertionError

Is the different remote URL part of what this intends to test? Or is this something that broke at some point (or would have broken, if it were an assertion) as a result of moving the remote to GitHub?

It's surprising to see a gitorious URL there - where would that be coming from?

When looking at this confused, I'd think it's definitely not suitable as tutorial of any kind. Maybe it's better to either revamp it into something more useful, or remove it entirely.

I don't think this was part of the tutorial. In this PR, I removed @NoEffect everywhere in the tests, not just in lines of code that are included in the generated documentation. (This occurrence was in the submodule tests.)

Oh, thanks for the clarification, I should have known by looking at the filename in the provided code excerpt.

Since it's already a FIXME, I presume that when trying to improve the GitPython package layout and maybe make submodule tests independent of their containing repository, this will naturally be resolved.

I think making the tests independent would entail fixing it. Whether or not improving the package layout would depends in more details of how that is done. It could also probably be fixed directly, but this would require figuring out where that old URL came from and what, exactly, the bounds are of what the test intends (or should intend) to test.

Byron added this to the v3.1.37 - Bugfixes milestone Sep 22, 2023

Byron approved these changes Sep 22, 2023

View reviewed changes

Byron merged commit 1774f1e into gitpython-developers:main Sep 22, 2023

EliahKagan deleted the no-noeffect branch September 22, 2023 07:03

data-sync-user mentioned this pull request Feb 9, 2025

Bump gitpython from 3.1.32 to 3.1.37 mozilla/opmon#168

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Remove @NoEffect annotations#1677