[prev in list] [next in list] [prev in thread] [next in thread] 

List:       gcc-patches
Subject:    Re: [PATCH v2] New C++ warning option '-Waccess-specifiers'
From:       Franz Sirl <Franz.Sirl-kernel () lauterbach ! com>
Date:       2017-07-24 12:26:00
Message-ID: 47ec4552-7940-8860-34c6-5da679764eaf () lauterbach ! com
[Download RAW message or body]

Am 2017-07-24 um 00:19 schrieb Volker Reichelt:
> On 23 Jul, Eric Gallager wrote:
>> On 7/23/17, Volker Reichelt <v.reichelt@netcologne.de> wrote:
>>> Hi again,
>>>
>>> here is an updated patch for a new warning about redundant
>>> access-specifiers. It takes Dave's various comments into account.
>>>
>>> The main changes w.r.t. to the previous versions are:
>>>
>>> * The warning is now a two-level warning with a slightly shorter name:
>>>    -Waccess-specifiers=1, -Waccess-specifiers=2
>>>    with -Waccess-specifiers defaulting to -Waccess-specifiers=1.
>>
>> Just a more generalized comment as a user, but I don't really like
>> this trend that new warning options are so often given numeric levels
>> these days. A warning option with different levels requires special
>> handling in configure scripts or Makefiles, which is harder than just
>> toggling different names (i.e. how things work without numeric
>> levels).
> 
> Fair point.

Another point is the handling of -Werror=. AFAIK it would be impossible 
right now to have "-Werror=access-specifiers=1 -Waccess-specifiers=2", 
with a combined meaning of "error for level 1 + warning for level 2".

Actually, are the intended semantics for the existing cases (eg. 
-Warray-bounds=) vs. -Werror= even documented somewhere?

Franz
[prev in list] [next in list] [prev in thread] [next in thread]