bug#35062: [PATCH v2] min_cols/rows is always 0, remove noop actions

[Konstantin Kharlamov <hi-angel@yandex.ru>]

В Сб, апр 6, 2019 at 10:09, Eli Zaretskii <eliz@gnu.org> 
написал:
>>  Date: Sat, 06 Apr 2019 01:18:38 +0300
>>  From: Konstantin Kharlamov <hi-angel@yandex.ru>
>>  Cc: 35062@debbugs.gnu.org
>> 
>>  >>  * src/gtkutil.c:
>>  >>      remove "always false" comparison
>>  >>      remove multiplication by zero, it's a noop anyway
>>  >>      remove min_cols and min_rows as it's now unused
>>  >>
>>  >>  Fixes LGTM warnings:
>>  >>      * Comparison is always false because min_cols <= 0.
>>  >>      * Comparison is always false because min_rows <= 0.
>>  >>  ---
>>  >>  v2: remove the min_rows/min_cols whatsoever
>>  >
>>  > Thanks, pushed.  Please in the future write the commit log 
>> messages in
>>  > the ChangeLog-like style, see CONTRIBUTE for the details.  I did 
>> that
>>  > this time, you can look at what I pushed to see an example of that
>>  > style in action.
>> 
>>  Thanks, so, I'm looking right now to modify the rest of series and
>>  resend it here, and… I think I miss something: the only 
>> difference of
>>  my commit from CONTRIBUTE file is that I forgot to list the 
>> functions
>>  changed.
> 
> That's right.  And also the lack of the bug number, see below.
> 
>>  I think I misunderstand something, because you completely rewrote 
>> the
>>  commit messages compared to mine. Titles now are more vague, and for
>>  some reason you removed a note that one of patches fixes warnings in
>>  the code.
> 
> That's just my personal preferences of style, as there's always a
> judgment call regarding the description of the changes.  We don't have
> to agree about that, and if your log message was in the right format,
> I probably wouldn't have bothered changing their text.
> 
> I can explain my preference: your text described in a very detailed
> form something that was acutely visible from the diffs themselves,
> whereas my text only describes the motivation, leaving the rest to the
> diffs which speak for themselves.

Thanks!

> The part about LGTM warnings didn't seem important enough to me: LGTM
> is not a compiler, and without a reference to what it is, the text is
> a small riddle in itself.  More importantly, I don't think we need any
> justification for removing variables initialized to zero and never
> changed.

Well, one can find what LGTM through a search engine, or for that 
matter by using "BUG#XXX" to find this discussion. But whatever.

> Again, these are my personal preferences, not something we require
> from each contributor.  So if you disagree, you can keep your style,
> and only adjust the form.
> 
>>  And then you also added "Bug#35062", even though there's no
>>  bug (the report existence is incidental, it's because bugtracker is
>>  used to track patches in Emacs.
> 
> References to bug numbers should always be in the log message, because
> that allows an easy way of reading relevant discussions recorded by
> the bug tracker.  This _is_ in CONTRIBUTE, btw.

Oh, thanks, I missed that.

>>  But writing "Bug" will only confuse readers — there's no bug).
> 
> "Bug" here doesn't mean a literal bug, it means an issue recorded by
> the bug tracker.

I see, still it doesn't mean arbitrary readers won't be confused. 
git-logs are being read by lots of different people, most of them are 
not even contributors, could be e.g. users wondering whether it's worth 
to re-build their Emacs-from-git.