Re: Support 64-bit default GCC on Solaris/x86

Rainer Orth wrote: In version 4.7, GCC will gain a 64-bit default Solaris/x86 configuration, similar to the existing sparcv9-sun-solaris2 configurations. In order for that to work with GNU ld (Sun ld works out of the box), I had to make the following minor patch to libtool.m4. This patch has be

[PATCH 3/3] maint: dynamically strip unused scripts from libltdl Makefile.

This is in preparation for punting maintenance of more of our code to gnulib... and rather than keep a hard-coded list of which files in $(aux_dir) should be installed by libtoolize for libltdl, this patch figures out which aux_files are not part of the libltdl build so that we can add gnulib modul

[PATCH 2/3] build: substitute paths into defs.m4sh instead of recalculating.

More DRY inspired cleanup. 72 hours... etc etc. * Makefile.am (configure_edit): Add substitutions for aux_dir, macro_dir and srcdir. * tests/defs.m4sh: Don't recalculate srcdir. (aux_dir, macro_dir, srcdir): Convert to absolute paths. Signed-off-by: Gary V. Vaughan --- Makefile.am |5 +

[PATCH 1/3] build: support AM_SILENT_RULES

Thanks to the flooding in Bangkok making my Internet connection unreliable enough that email has been a bit of a challenge even when the electricity is on... I've been tidying up Libtool some more. This patch makes errors in an end-user build extremely easy to spot. After a successful configure,

FYI: [PATCH 4/4] maint: add autobuild prerequisite only if autobuild.m4 is absent.

As discussed on another thread[1]. Applied as obvious, and also pending acceptance in upstream gnulib. [1]: http://lists.gnu.org/archive/html/libtool-patches/2011-10/msg00020.html * bootstrap (func_require_autobuild_buildreq): Skip if autobuild.m4 is already present. * bootstrap.conf (require_a

FYI: [PATCH 3/4] maint: ensure bootstrap runs from dist tarball.

This is a fix to the master-copy of bootstrap, pending acceptance into upstream gnulib. Applied as obvious. * bootstrap (func_gnulib_tool_copy_file): When gnulib-tool is not available, check that there is a copy of the file from a previous run available before bailing out with a diagnostic. (func

FYI: [PATCH 2/4] tests: add a keyword `expensive' to very long running tests

Applied as obvious. This shaves about 15 minutes of every casual make distcheck for me, while still leaving the default to run even the crazy slow test cases. :) * tests/cmdline_wrap.at, tests/stresstest.at (AT_KEYWORDS): Add `expensive'. * NEWS: Updated. Signed-off-by: Gary V. Vaughan --- NEW

FYI: [PATCH 1/4] libtoolize: fix some long-standing sed substitution bugs

Applied as obvious. * libtoolize.m4sh (func_fixup_Makefile): `\$' in a " quoted script that is evaluated twice before execution becomes a `\$', which sed interprets as a literal $. The script will then delete any lines with a first non-whitespace character of `$'. Instead use `$' in singl

Support 64-bit default GCC on Solaris/x86

In version 4.7, GCC will gain a 64-bit default Solaris/x86 configuration, similar to the existing sparcv9-sun-solaris2 configurations. In order for that to work with GNU ld (Sun ld works out of the box), I had to make the following minor patch to libtool.m4. This patch has been tested with a slig

autobuild bootstrap dependency [WAS Re: bug#9845: [PATCH 1/3] maint: use gnulib's maint.mk and support scripts release procedure.]

Hi Peter, On 24 Oct 2011, at 09:12, Gary V. Vaughan wrote: > On 24 Oct 2011, at 08:22, Peter O'Gorman wrote: >> On 10/23/2011 11:03 AM, Gary V. Vaughan wrote: >>> By the end of this series, making a release still involves an awful lot >>> of waiting, but after passing the bevy of make distcheck va