Re: avoiding warnings: one last patch

[Jim Meyering]

Paolo Bonzini wrote:
> On 03/23/2010 10:05 AM, Jim Meyering wrote:
>> Testing with other versions of gcc, I encountered these
>> warnings from gcc-4.1.2 (mapped to error with -Werror):
>>
>>      cc1: warnings being treated as errors
>>      dfa.c: In function 'lex':
>>      dfa.c:467: warning: 'wc' may be used uninitialized in this function
>>      dfa.c:463: warning: 'c' may be used uninitialized in this function
>>      dfa.c:467: warning: 'wc2' may be used uninitialized in this function
>>      dfa.c:467: warning: 'wc1' may be used uninitialized in this function
>>      dfa.c:463: warning: 'c2' may be used uninitialized in this function
>>      make[2]: *** [dfa.o] Error 1
>>
>> And in a way, they seem legitimate, assuming that older gcc doesn't
>> recognize "error (EXIT_TROUBLE, ..." as something that resolves to
>> exit (EXIT_TROUBLE).
>>
>> The clang static analyzer makes the same "mistake" in reporting a
>> false-positive "used-uninitialized" warning, so I wrote the following
>> patch, which is sufficient to teach both clang and gcc-4.1.2 enough
>> that they no longer give FP warnings.
>>
>> It's risky to do *anything* this close to release, but I've confirmed
>> this solves the problem and everything still builds on numerous systems,
>> so unless someone points out a problem very soon, I'll push it and it
>> will be included in grep-2.6.
>
> Yes, good idea.

Thanks for the confirmation.