Re: 21.4.19 with gcc 4.1
Hi Ville,
Ville Skyttä <scop(a)xemacs.org> wrote:
Is anyone having success with running 21.4.19 built with gcc 4.1?
The package currently in Fedora Extras 5 isn't working too impressively;
among the reported things are at least: non-ASCII chars in buffers being
rendered somewhat larger than ASCII; various crashes, for example in M-x
man netstat if built with -fstack-protector, and others with gnus;
random infinite loops in gnus, crashes at startup using certain locales
(not reproducible everywhere).
Rebuilding the same package with gcc 3.2.3 appears to avoid most if not
all of the above problems on the exact same systems.
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=183643
FYI, attached is the warning output from the build when built with gcc
4.1 on a x86_64.
I just upgraded my machines to Fedora Core 5 over the last couple of
days, so I'll give this a try. Just a random shot in the dark, though:
does compiling with -fno-strict-aliasing have any effect?
I have a couple of x86_64 machines now and noticed that we don't have an
m/ file for them. On the other hand, I'm not sure anything would have
to be in an m/x86_64.h file, other than possibly definitions for
LOAD_AVE_TYPE and LOAD_AVE_CVT. Emacs now has an m/amd64.h, which is
slightly misnamed in light of Intel's EM64T. Should we get one of
those, too?
--
Jerry James, Assistant Professor james(a)xemacs.org
Computer Science Department http://www.cs.usu.edu/~jerry/
Utah State University