[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: gnulib and automake
From: |
Thomas Treichl |
Subject: |
Re: gnulib and automake |
Date: |
Thu, 29 Oct 2009 22:31:30 +0100 |
User-agent: |
Thunderbird 2.0.0.23 (Macintosh/20090812) |
John W. Eaton schrieb:
| I would be surprised if this has any effect on the way .oct files are
| created now, because they are constructed with libtool (as described
| in the post that started this thread). The configure script still has
| the code that sets up Octave's old shared-library creation variables,
| but they are no longer used to compile anything. The plan is to
| remove this code from the configure script once mkoctfile has also
| been converted to use use libtool.
Sorry, maybe I misunderstood what the problem was. Looking at your
patch (duh, should have done that before I replied the first time) it
doesn't change the way .oct files are created, but does change the way
they are opened, correct? If so, then yes, please try this and see if
it works to use the dlopen interface to open the .oct files that are
created with libtool. If so, then yes, I think this change should be
applied.
Thanks, and sorry for the confusion...
No problem, and yes, it was the idea to change the way *.oct files are loaded on
MacOSX and not how *.oct files are created. And here are the good news: This
change was one of the right things to do and it solved the problem for me. Here
is the same short Octave session output:
bash$ ./run-octave --quiet
octave:1> which repmat
`repmat' is a function from the file
/Users/Thomas/Development/octave.am2/scripts/general/repmat.m
octave:2> which reshape
`reshape' is a built-in function
octave:3> which cellfun
`cellfun' is a function from the file
/Users/Thomas/Development/octave.am2/src/DLD-FUNCTIONS/cellfun.oct
octave:4> cellfun (@disp, {1,2,3}, 'UniformOutput', false)
ans =
{
[1,1] = 1
[1,2] = 2
[1,3] = 3
}
Attached I'm sending a simple "hg diff configure.ac >configure.ac.patch", I
don't know if you prefer any other type of changeset in this stage, then let me
know and I will send something else tomorrow.
Best regards
Thomas
PS. I am very happy with the change to libtool because I still needed to add
more "-llibs" flags to src/Makefile.in with the old way. Now libtool does all
this work for me and these dependency problems all disappeared. This is great!
Thanks.
diff -r 52a70c7d4935 configure.ac
--- a/configure.ac Wed Oct 28 11:56:07 2009 -0400
+++ b/configure.ac Thu Oct 29 22:27:17 2009 +0100
@@ -1594,10 +1594,15 @@
## Check for dyld first since OS X can have a non-standard libdl
save_LIBS="$LIBS"
- AC_CHECK_HEADER(mach-o/dyld.h)
- if test "$ac_cv_header_mach_o_dyld_h" = yes; then
- dyld_api=true
- else
+ AC_ARG_WITH(mach-dyld,
+ [AS_HELP_STRING([--with-mach-dyld], [enable older Mach-O dyld interface])],
+ with_mach_dyld=yes, with_mach_dyld=$withval)
+ if test "$with_mach_dyld" = yes; then
+ AC_CHECK_HEADER(mach-o/dyld.h)
+ if test "$ac_cv_header_mach_o_dyld_h" = yes; then
+ dyld_api=true
+ fi
+ elif test "$dyld_api" = false; then
AC_CHECK_LIB(dld, shl_load, [DL_LIBS=-ldld; LIBS="$LIBS $DL_LIBS"])
AC_CHECK_FUNCS(shl_load shl_findsym)
if test "$ac_cv_func_shl_load" = yes \
- Re: gnulib and automake, (continued)
- Re: gnulib and automake, Thomas Treichl, 2009/10/28
- Re: gnulib and automake, John W. Eaton, 2009/10/28
- Re: gnulib and automake, Thomas Treichl, 2009/10/29
- Re: gnulib and automake, John W. Eaton, 2009/10/29
- Re: gnulib and automake, Thomas Treichl, 2009/10/29
- Re: gnulib and automake, John W. Eaton, 2009/10/29
- Re: gnulib and automake, John W. Eaton, 2009/10/29
- Re: gnulib and automake,
Thomas Treichl <=
- Re: gnulib and automake, John W. Eaton, 2009/10/29
- Re: gnulib and automake, Ben Abbott, 2009/10/29
- Re: gnulib and automake, John W. Eaton, 2009/10/28
- Re: gnulib and automake, Thomas Treichl, 2009/10/30
- Re: gnulib and automake, John W. Eaton, 2009/10/30
Re: gnulib and automake, Marco Atzeri, 2009/10/28
Re: gnulib and automake, Søren Hauberg, 2009/10/28