Declutter infoview tests.

[rish987]

Should also make testing infoview-specific stuff much more deliberate and complete!

[codecov]

Codecov Report

Merging #104 (05fcb4d) into main (ba7562e) will increase coverage by 0.28%.
The diff coverage is 98.13%.

@@            Coverage Diff             @@
##             main     #104      +/-   ##
==========================================
+ Coverage   92.15%   92.43%   +0.28%     
==========================================
  Files          25       27       +2     
  Lines        1109     1243     +134     
==========================================
+ Hits         1022     1149     +127     
- Misses         87       94       +7     

Impacted Files	Coverage Δ
lua/tests/helpers.lua	95.40% <96.21%> (+1.20%)	⬆️
lua/tests/infoview/autoopen_false_spec.lua	100.00% <100.00%> (ø)
lua/tests/infoview/autoopen_spec.lua	100.00% <100.00%> (ø)
lua/tests/infoview/close_all_spec.lua	100.00% <100.00%> (ø)
lua/tests/infoview/run_spec.lua	100.00% <100.00%> (ø)
lua/tests/infoview/toggle_spec.lua	100.00% <100.00%> (ø)
lua/lean/abbreviations.lua	65.92% <0.00%> (-8.15%)	⬇️
lua/tests/abbreviations/snippets_nvim_spec.lua
... and 3 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update ba7562e...05fcb4d. Read the comment docs.

[Julian]

This seems quite fragile to me because it adds global mutable state to all infoview-touching test runs. Would really prefer if we avoided doing that, we're already struggling with #94 showing us we have too much mutable state.

[rish987]

This is something that you'd "opt-in" to on certain tests, and should be backwards compatible with the way we did things before (see e.g. infoview/lsp_spec.lua). ISTM that mutable state is pretty much the only way to get this kind of coverage, and the mutable state before was just littered throughout the tests themselves rather than centralized in a helper (not to mention being hopelessly incomplete).

Did you perhaps mean #93? I'm pretty sure at this point the best way to handle it would be to:

1. Split up infoview/run_spec.lua.
2. Add a feature to nvim-lua/plenary.nvim that allows tests to early-abort as soon as an it() fails, so only the first error is shown.

[rish987]

@Julian Thanks for the nudge. What I had before was actually a pretty bad take, sorry for making you read through it. I pretty much overhauled this in the recent commit, it should hopefully be much easier to follow now! And at this point I think it gives 100% coverage of what I wanted to test.

[rish987]

Will add more assert messages and split up infoview/run_spec.lua tommorrow!

[rish987]

Okay, should be good for a review now! The only last thing I want to do is add a few helpful assertion failure messages, will do soon. Also regarding #93, see aa8b0db, nvim-lua/plenary.nvim#196 and nvim-lua/plenary.nvim#197.

[Julian]

Nice! Thanks for more fun upstream work, well done. Yeah sounds good to merge here then to me!