At Tue, 21 Jun 2022 12:49:24 +0200, Peter Eisentraut <peter.eisentraut@enterprisedb.com> wrote in
> I suggest that we add the gcc (also clang) option -ftabstop=4.
>
> This has two effects: First, it produces more accurate column numbers
> in compiler errors and correctly indents the code excerpts that the
> compiler shows with those. Second, it enables the compiler's
> detection of confusingly indented code to work more correctly. (But
> the latter is only a potential problem for code that does not use tabs
> for indentation, so it would be mostly a help during development with
> sloppy editor setups.)
>
> Attached is a patch to discover the option in configure.
>
> One bit of trickery not addressed yet is that we might want to strip
> out the option and not expose it through PGXS, since we don't know
> what whitespacing rules external code uses.
>
> Thoughts?
There's no strong reason for everyone to accept that change, but I
myself don't mind whether the option is applied or not.
There was a related discussion.
https://www.postgresql.org/message-id/BDE54C55-438C-48E9-B2A3-08EB3A0CBB9F%40yesql.se
> The -ftabstop option is (to a large extent, not entirely) to warn when tabs and
> spaces are mixed creating misleading indentation that the author didn't even
> notice due to tabwidth settings? ISTM we are better of getting these warnings
> than suppressing them.
regards.
--
Kyotaro Horiguchi
NTT Open Source Software Center