Quality Issues on mainline

[Gerald Pfeifer]

On Thu, 21 Aug 2003, Andrew Pinski wrote:
>>   template <class T>
>>   void f( const char *s ) {
>>     std::cout << s;            // line 5
>>     }
>>
>>   % gccvs -c -W bug.cc
>>   bug.cc: In function `void f(const char*)':
>>   bug.cc:5: warning: statement has no effect
> This is caused by Nathan's work which tried to fix a regression.
> Please file a bug so this will not happen again and also cc nathan
> on the bug because it is his bug.

The point of my message was less to point out this specific bug, but to
note that we have some quality problems (which might indicate serious
weaknesses in our testsuites or the default options used there).

I know from others, who volunteer as guinea pigs by building GCC and
testing it against their sources on a daily base, that recently the
situation has been worse than during the last months (apart from a
few days after the merger of the new C++ parser where this was to be
expected).

As I said, this may be partly due to testsuite problems.

Gerald