[Python-Dev] Cleanup of abstract.h header

[Benjamin Peterson]

I think it looks better. Thank you.

On Thu, Dec 15, 2016, at 02:22, Victor Stinner wrote:
> Hi,
> 
> I made multiple changes to the Include/abstract.h header file, because
> it was inconsistent in different manners:
> 
> * Parameter names of functions of the PyObject_Call family were
> inconsistent: "func" versus "callable" for a Python callable object
> for example (sometimes, .c and .h files were inconsistent too)
> 
> * Only this header file used an indentation of 4 spaces in the whole
> space
> 
> * Only this header used comments *after* the function declaration and
> with a newline between the comment and the declaration. Other headers
> use a comment *before* the declaration and no newline.
> 
> * Some comments were far (100 lines) from function declaration, so it
> wasn't possible anymore to understand these comments.
> 
> * Other tiny changes
> 
> Before:
> https://hg.python.org/cpython/file/f692dafe6797/Include/abstract.h
> 
> Now:
> https://hg.python.org/cpython/file/c4bcca326c0a/Include/abstract.h
> 
> (Not sure that it's easy to compare such long header file, ~1 200
> lines, in a web browser.)
> 
> See http://bugs.python.org/issue28838 for more information.
> 
> 
> Serhiy Storchaka was opposed to these changes, some extract of his
> comments:
> 
> * "Isn't this just a lot of churn for not a lot of benefit?" - Eric V.
> Smith asked the same question
> * "It breaks "hg annotation" and makes harder researching the history
> of the file"
> 
> I already pushed all my changes anyway, but Serhiy asked me to discuss
> these changes on Python-Dev, so here I am.
> 
> 
> I decided to cleanup abstract.h because I had to modify it multiple
> times last 2 years, especially when I added new functions for fast
> calls, and I was always surprised and confused by the style of header.
> I didn't know how to format my code, and it seems like I also
> introduced some inconsistent coding style in the newly added code.
> 
> For the first change that I made recently, normalizing parameter
> names, I first pushed directly a change without review, but Serhiy
> asked me on this list to revert it. So I reverted the change and
> continued the discussion on the issue #28838. We agreed on better
> names, and so I pushed a different change.
> 
> Victor
> _______________________________________________
> Python-Dev mailing list
> Python-Dev at python.org
> https://mail.python.org/mailman/listinfo/python-dev
> Unsubscribe:
> https://mail.python.org/mailman/options/python-dev/benjamin%40python.org