This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: _GLIBCXX_ASSERT lossage
- From: Benjamin Kosnik <bkoz at redhat dot com>
- To: Paolo Carlini <pcarlini at suse dot de>
- Cc: peturr02 at ru dot is, libstdc++ at gcc dot gnu dot org
- Date: Tue, 2 Mar 2004 13:54:35 -0600
- Subject: Re: _GLIBCXX_ASSERT lossage
- Organization: Red Hat / Chicago
- References: <07D05A69A3D0C14FAEA60C3ACE8E5564028F561B@mail.ru.is><4023AA90.2030402@suse.de>
>>However, I haven't seen any test case that uses VERIFY, but requires
>>that _GLIBCXX_ASSERT should not be defined (I may have missed some
>>though) so this seems pretty bogus.
>>
>Yes, seems bogus to me too, now. Probably, time ago had an actual
>use... dunno...
>
>Let's wait a bit for other opinions: I vote in favor of simply removing
>the thing and unconditionally having VERIFY equivalent to an assert.
I'm ok with this.
VERIFY was done with the assert conditional so that the large testsuite
files of the day could be navigated simply by putting a gdb watch point
on the bool variable test. When it changed to false, that's were the
errors were located. Multiple errors could be examined without going to core.
Now, this is less useful, since the test cases are of a manageable size.
Hopefully this explains what was being attempted.
-benjamin