This is the mail archive of the
mailing list for the GCC project.
Re: Regressions on mainline\
Richard Henderson wrote:
On Mon, Jun 21, 2004 at 03:44:30PM -0700, Mark Mitchell wrote:The latter. Nobody is likely to worry that they caused a problem if
they see an XPASS; at worst, they'll investigate to see if they really
solved the problem. With a FAIL, people have to do extra work to
confirm that they didn't cause the problem before checking in, for example.
OK. Would you please make sure that the PR is marked for 3.5.0 and
XFAIL the test?
The problem is that it doesn't fail everywhere. Not on i686-linux
for instance. So: do you prefer FAIL on some targets, or XPASS on
the opposite set?