Re: Some XEmacs notes
Ar an t-ochtú lá is fiche de mí na Nollaig, scríobh Jerry James:
Adding Jeff Sparkes to the conversation. One of you please clue me
in
if there is an active mailing list I need to re-sign up for.
See other communication, https://lists.xemacs.org/ to sign up.
On Thu, Dec 28, 2017 at 5:45 PM, Jerry James
<loganjerry(a)gmail.com> wrote:
> Well, doggone, it only fixed ppc64le. Both big endian architectures,
> ppc64 and s390x, now exhibit some kind of memory failure on startup,
> so I appear to have introduced an assumption of little endianness in
> something I just did. I'll sort that out next....
No, it wasn't me. Building on big endian architectures was broken by
this commit:
changeset: 6384:c6e0b59281cb
user: Jeff Sparkes <jsparkes(a)gmail.com>
date: Mon Apr 26 07:15:04 2010 -0400
summary: Use G_BIG_ENDIAN instead of WORDS_BIGENDIAN.
Specifically, src/lisp-union.h requires WORDS_BIGENDIAN to layout a
Lisp object properly, which is the cause of the breakage I am seeing
now. It looks like somebody merged a GTK branch into master, is that
what I'm seeing? Because this commit is from going on 8 years ago.
A grep shows that there are more instances of WORDS_BIGENDIAN in
src/md5.c, src/glyphs-x.c, and src/redisplay-x.c.
Was that merge supposed to make building with either gtk2 or gtk3
work? I get build failures with both. I'm wondering if that is
expected or not. Let me know if it is not expected and I will provide
details.
Yes, it was. Building on GTK2 and GTK3 work with warnings but without
problems for me. Please give details if it doesn’t work for you.
I have undone Jeff’s change to config.h.in, and now WORS_BIGENDIAN is
available again. Thanks for reporting this!
--
‘As I sat looking up at the Guinness ad, I could never figure out /
How your man stayed up on the surfboard after forty pints of stout’
(C. Moore)