[PATCH] D22725: [clang-tidy] Add check 'modernize-use-algorithm'

Piotr Padlewski via cfe-commits cfe-commits at lists.llvm.org
Tue Aug 9 16:34:24 PDT 2016


2016-08-09 5:49 GMT-07:00 Aaron Ballman <aaron.ballman at gmail.com>:
>
> I think this boils down to personal preference, which is why I'm
> concerned about the check. Either mechanism is correct, so this is
> purely a stylistic check in many regards.
>
> About warnings - well, if someone choose this check to be run, then he
> > probably wants to get warnings instead of notes.
>
> The problem is that people don't always choose this check to be run,
> they choose to run clang-tidy and this check is enabled by default. Or
> they choose to run modernize and this check is enabled by default.
>
> As with most checks. We can either be "perfect" and add only checks that
we consider very useful, or we can be open
and let people choose checks that they want from bigger set. After I used
clang-tidy first time, I know what checks are not something that I want, so
I just didn't use it.
I think we can work out on other solution. Maybe we should make some levels
of recomendation, so the user would use some option like --level=recommend
or something,
and we would have to somehow split the checks into groups. This would be
still problematic, but I know that it would be nice to have analyzer-alpha
checks into not recommend mode
(I don't know why they are used on default. They have so many bugs).

Anyway as you can see the matter of usefulness and very subjective. I can
argue that this is should belong to modernize, because I think that this is
modernization (changing code from
bugprone C ways to slightly better C++ ways), but of course everyone will
have own opinion.


Piotr

https://reviews.llvm.org/D22725
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20160809/17ef4df8/attachment.html>


More information about the cfe-commits mailing list