September 1998 - XEmacs-Beta - XEmacs Mailing List Archives

Olivier Galibert

In lrecord.h : extern int gc_in_progress; #ifdef USE_INDEXED_LRECORD_IMPLEMENTATION # define MARKED_RECORD_P(obj) (gc_in_progress && XRECORD_LHEADER (obj)->mark) #else # define MARKED_RECORD_P(obj) (gc_in_progress && \ XRECORD_LHEADER (obj)->implementation->finalizer == \ this_marks_a_marked_record) #endif Especially in the USE_INDEXED_LRECORD_IMPLEMENTATION, what is this gc_in_progress test for? The last stage of gc clears all marks (or the next gc would not work right), and anyway one should not test for marking outside of gc, no? OG.

1 participants
0 comments

Re: XEmacs 21.2 won't build on Solaris 25 years, 7 months

Michael Sperber

>>>>> "BP" == William M Perry <wmperry(a)aventail.com> writes: BP> Hrvoje Niksic <hniksic(a)srce.hr> writes: >> David Bush <david.bush(a)adn.alcatel.com> writes: >> >> > I haven't attempted 21.2 yet, but AFAIK there is no need for a -R if >> > the library is in the same place for both build and execution. >> >> There certainly is, under Solaris. Try it. BP> Under just about _any_ system. Although didn't some versions of gcc BP> automatically stick a -Rfoo for every -Lfoo they saw on the link line? BP> Ick. :) AIX executables store a path to dynamically linked libraries always. -- Cheers =8-} Chipsy Friede, Völkerverständigung und überhaupt blabla

1 participants
0 comments

Re: XEmacs 21.2 won't build on Solaris 25 years, 7 months

William M. Perry

Hrvoje Niksic <hniksic(a)srce.hr> writes: > David Bush <david.bush(a)adn.alcatel.com> writes: > > > I haven't attempted 21.2 yet, but AFAIK there is no need for a -R if > > the library is in the same place for both build and execution. > > There certainly is, under Solaris. Try it. Under just about _any_ system. Although didn't some versions of gcc automatically stick a -Rfoo for every -Lfoo they saw on the link line? Ick. :) -Bill P.

1 participants
0 comments

Re: Yet another SUMO tarball... 25 years, 7 months

Jan Vroonhof

Jonathan Harris <jhar(a)tardis.ed.ac.uk> writes: > install_dir/XEmacs-21.0/etc > i386-pc-win32 <-- binaries here > info > lisp > lock > site-packages/... > xemacs-packages/... and ../mule-packages for MULE enabled emacsen. > Instead, XEmacs finds > packages using configure-package-path defined at build time and/or the > EMACSPACKAGEPATH variable defined at runtime. > use to people on Windows who are building from source; the Windows > native binary bundle includes its own version of the sumo bundle. Does that mean 1. The Windows version is not as easily relocatable as the Unix version? 2. on Windows NT you basically revert to the old distribution method, i.e. just include everything?. Why? Jan

2 participants
1 comments

Re: XEmacs 21.2 won't build on Solaris 25 years, 7 months

Hrvoje Niksic

David Bush <david.bush(a)adn.alcatel.com> writes: > I haven't attempted 21.2 yet, but AFAIK there is no need for a -R if > the library is in the same place for both build and execution. There certainly is, under Solaris. Try it. -- Hrvoje Niksic <hniksic(a)srce.hr> | Student at FER Zagreb, Croatia --------------------------------+-------------------------------- Your lucky number today is 29842924728. Look for it everywhere.

1 participants
0 comments

Re: XEmacs 21.2 won't build on Solaris 25 years, 7 months

Hrvoje Niksic

Didier Verna <verna(a)inf.enst.fr> writes: > Hrvoje Niksic <hniksic(a)srce.hr> writes: > > > In 21.2, -L/usr/dt/lib is used, but without the appropriate -R. This > > makes conftests fail and ultimately configure to abort. Note that > > -R/usr/openwin/lib is there. > > ?? I didn't have any problem with this version (right out of the CVS > archive). What's the problem exactly ? Well, I said it. -R/usr/dt/lib is not added to the link path, so conftests die with "killed: cannot locate libXm.so.3" or the like. Do you have LD_LIBRARY_PATH set in your environment? This might explain why it works for you. -- Hrvoje Niksic <hniksic(a)srce.hr> | Student at FER Zagreb, Croatia --------------------------------+-------------------------------- - Now what did we learn from this? - I learned what my liver looks like!

2 participants
1 comments

Re: XEmacs 21.2 won't build on Solaris 25 years, 7 months

Didier Verna

Hrvoje Niksic <hniksic(a)srce.hr> writes: > In 21.2, -L/usr/dt/lib is used, but without the appropriate -R. This > makes conftests fail and ultimately configure to abort. Note that > -R/usr/openwin/lib is there. ?? I didn't have any problem with this version (right out of the CVS archive). What's the problem exactly ? --didier

1 participants
0 comments

Re: Yet another SUMO tarball... 25 years, 7 months

Jan Vroonhof

Jonathan Harris <jhar(a)tardis.ed.ac.uk> writes: > Please don't do that. I was expecting this (however not for this reason). > The appropriate path on native Windows systems > doesn't include the lib/xemacs bit. Why not. Where do the packages go under NT? Does it include just the "xemacs/..." part > A brief note in the README advising > the user to unpack into lib/xemacs would probably be as good. I really hate it when tarfiles unpack in more than one directory, it also very annoying when you have to unpack the binary, the common and the sumo tarball in different directories. My feeling is that as long as XEmacs thinks bindir/../xemacs is special we better make sure this exists. It is easier to persuade capable people to just rename the tree if they need it somewhere different. Alternatively we could just unpack everything into some xemacs-bla.bla directory and have the sysadmin move/copy the whole tree. Jan

3 participants
2 comments

Dynodump hosed in 21.2? 25 years, 7 months

Hrvoje Niksic

Either something is seriously hosed in 21.2, or I am missing something really obvious. I am unable to build 21.2, obtained by CVS checkout, on my usual system. First the -R problem, then Dynodump: make[1]: Entering directory `/home/srce/hniksic/work/emacs/xemacs-21.2/mybin/dynodump' gcc -c -g -O3 -msupersparc -funroll-loops -R/usr/dt/lib -R/usr/local/lib -R/home/srce/hniksic/work/lib -fpic -I/home/srce/hniksic/work/emacs/xemacs-21.2/dynodump -I/home/srce/hniksic/work/emacs/xemacs-21.2/dynodump/sparc /home/srce/hniksic/work/emacs/xemacs-21.2/dynodump/sparc/_relocate.c In file included from /usr/include/sys/elf.h:18, from /usr/include/libelf.h:14, from /home/srce/hniksic/work/emacs/xemacs-21.2/dynodump/sparc/_relocate.c:40: /usr/include/sys/elftypes.h:53: #error "64 bit integer types not supported on this platform" make[1]: *** [_relocate.o] Error 1 make[1]: Leaving directory `/home/srce/hniksic/work/emacs/xemacs-21.2/mybin/dynodump' make: *** [dynodump] Error 2 Anybody has a clue what's wrong here? -- Hrvoje Niksic <hniksic(a)srce.hr> | Student at FER Zagreb, Croatia --------------------------------+-------------------------------- The meta-Turing test counts a thing as intelligent if it seeks to devise and apply Turing tests to objects of its own creation.

1 participants
0 comments

XEmacs 21.2 won't build on Solaris 25 years, 7 months

Hrvoje Niksic

In 21.2, -L/usr/dt/lib is used, but without the appropriate -R. This makes conftests fail and ultimately configure to abort. Note that -R/usr/openwin/lib is there. This problem didn't exist in 21.0. -- Hrvoje Niksic <hniksic(a)srce.hr> | Student at FER Zagreb, Croatia --------------------------------+-------------------------------- Try to use "ad nauseam" at least once per flame. It doesn't mean anything; but it gives that polished feel to your postings.

1 participants
0 comments

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

XEmacs-Beta September 1998