Re: PATCH RFA: -Wc++-compat warning about ++/-- with enum value

[Ian Lance Taylor <iant at google dot com>]

Steve Kargl <sgk@troutmask.apl.washington.edu> writes:

> Sure, -Wc++-compat prevents someone from committing code that,
> for example, uses ++ on an enum-specifer entity.  That someone,
> now needs to find a C++ reference that explains the issue and
> how to code around the problem.

I don't see this as equivalent to a requirement to learn C++.  If you
have any questions -Wc++-compat warnings, please feel free to ask me or
the gcc@ mailing list.  There are many things that a gcc maintainer
needs to know.


>> > More to the point, including -Wc++-compat in CFLAGS can make it
>> > impossible to do a binary search to determine what revision 
>> > caused a regression.  Try backing out only the Fortran portion of
>> > r146855 and rebuild. 
>> 
>> I don't understand how this is different from any other compiler
>> warning.  Warnings do change over time.  For example, the same thing
>> could have occurred when we toughened up -Wparentheses.
>> 
>> Note that -Wc++-compat has been in CFLAGS since at least gcc 4.2, so
>> what is new is not -Wc++-compat, but the fact that -Wc++-compat is
>> getting tougher.
>
> Here's the problem.  There was a regression in fortran io,
> and one recent change that touched trans-io.c was r146855.
> It was found that r146855 was the cause, but one can't 
> determine this from reverting the change to trans-io.c.

I understand the problem.  I'm just saying that there is nothing
specific to -Wc++-compat about this problem.  It is a general problem
when warning options change.  Despite the existence of this problem, we
do change warning options.

I apologize if I broken fortran io with r146855.  This is the first I've
heard of the problem.

I will note that if you fully reverted r146855, which affected many
files in the compiler, then bootstrap would have succeeded.

Ian