apparent gcc4 regression in xprint on powerpc

[Drew Parsons <dparsons@debian.org>]

Dear powerpc and gcc developers,

some Xprint users have reported a segfault in Xprt ( bug #337570,
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=337570). running on
powerpc.

They report it occurs with xprint 0.1.0.alpha1-12, but not with -11 or
-10.

Checking the build logs at
http://buildd.debian.org/build.php?pkg=xprint, I see -11 was built for
powerpc with gcc-3.3 (1:3.3.6-7), while -12 was built with gcc-4.0
(4.0.1-6).

So there appears to be some regression in gcc4.

Current gcc is 4.0.2-1. Is the segfault in xprint -12 something you
might expect to occur, and would a rebuilt with gcc 4.0.2 be likely to
repair it?

One of the users (Felix, cc:d here) is attempting to build Xprt himself
so we'll hear what his success is (he needs a fresh binary to get
unstripped symbols so we can obtain a backtrace from gdb).

Thanks,

Drew Parsons