This is the mail archive of the
mailing list for the GCC project.
Re: GCC 3.2
- From: Jeff Law <law at porcupine dot slc dot redhat dot com>
- To: Jakub Jelinek <jakub at redhat dot com>
- Cc: Franz Sirl <Franz dot Sirl-kernel at lauterbach dot com>, Mark Mitchell <mark at codesourcery dot com>, gcc at gcc dot gnu dot org, "Kevin B. Hendricks" <kevin dot hendricks at sympatico dot ca>
- Date: Mon, 12 Aug 2002 14:32:22 -0600
- Subject: Re: GCC 3.2
- Reply-to: law at redhat dot com
In message <20020812180910.K18963@sunsite.ms.mff.cuni.cz>, Jakub Jelinek
>You mean gcc/testsuite/consistency.vlad stuff?
Nope. This was something rather different. We had hooks in the compiler
itself to dump out structure layouts for use by assembly coders.
It didn't even occur to us to use that infrastructure to solve the problem
that Vlad's layout checker solves. The offset-info stuff would need some
beefing up, but it shouldn't be a terrible task.
>IMHO it would be better to write a layout test generator, using all
>combinations for the simpler tests and pick some hundreds tests at random for
>more complicated tests.
You could certainly do this too. The idea behind having it done in libstdc++
is that we can verify that sizes/layouts of any externally visible classes
and structures don't change from one rev of libstdc++ to another. That
in conjunction with testing the signatures of every externally visible
function gives us a much better chance of keeping libstdc++ compatible
from one rev to the next.