fix issue where passing two styles would fail with index error

[jspaezp]

Fixes the first section of:

Closes #160

[codecov]

Codecov Report

Merging #162 (edf4976) into main (9268fd6) will not change coverage.
The diff coverage is 100.00%.

@@            Coverage Diff            @@
##              main      #162   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           21        21           
  Lines          591       592    +1     
=========================================
+ Hits           591       592    +1     

Impacted Files	Coverage Δ
pydocstringformatter/_formatting/base.py	100.00% <100.00%> (ø)
...stringformatter/_formatting/formatters_numpydoc.py	100.00% <100.00%> (ø)

[jspaezp]

Tests fail due to the auto-fix from pre-commit.ci 0c6795b
... which raises the question ... why is it generating those trailing whitespaces?

• Should a "two pass" approach be done?
• Should there be an "Extras" where everything that was not placed in a section would go?

[DanielNoord]

• Should a "two pass" approach be done?

I have been thinking about this as well, but not sure if we should do that as we can create infinite loops.

Those trailing whitespaces seem to match the previous indentation. I think that is a bug and should only be a new line without the extra spacing.

[jspaezp]

@DanielNoord What about this behavior? Where there is a section separator but no name. I think it makes clear that the parser was not able to assign what is under it to any section and not that it is part of the former.

Those trailing whitespaces seem to match the previous indentation. I think that is a bug and should only be a new line without the extra spacing.

[DanielNoord]

Oh this is becoming more difficult than I hoped. It seems both styles are somewhat incompatible with each other. This is the missing section name, which is now being used as a summary/title...

I'll have to think about this. Perhaps we should fix this in another PR, but I like to check if this is easily fixable.

[jspaezp]

Well ... if you believe that the issue is that the styles are mutually exclusive in some instances, enforcing that is not hard internally.

[DanielNoord]

I think the issue that they are currently mutually exclusive. Perhaps we can fix this by changing the order in which formatters are called. I hope to have time somewhere this week to look into this.

[jspaezp]

@DanielNoord

• Should a "two pass" approach be done?

I have been thinking about this as well, but not sure if we should do that as we can create infinite loops.

Those trailing whitespaces seem to match the previous indentation. I think that is a bug and should only be a new line without the extra spacing.

Thinking about this ... I started working a branch that does a two pass to check the stability of the tokens.
I will reference it in a new issue.

[Pierre-Sassoulas]

Was this replaced by #169 @jspaezp ? (Should we close ?)

[jspaezp]

Hello @Pierre-Sassoulas! sorry for not closing this issue before (just for the future, it was fixed here #163)

[Pierre-Sassoulas]

No problem, thank you for answering :) !

[DanielNoord]

I do think this test still fails on main? At least it did last time I checked after we merged #163.

Edit: this does indeed still reproduce.

[github-actions]

According to the primer, this change has no effect on the checked open source code. 🤖🎉