| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
| |
This ebuild was failing, due, apperently, to its poorly constructed
makefile. While I was at it, I noticed some QAwarns apperently due
to a missing prototype in upstream code.
Fixed both via patches.
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
|
|
| |
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
|
|
| |
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
|
|
|
|
| |
Also cleans up some documentation cruft that used to end up in libdirs
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
|
|
|
|
| |
multilib-utization)
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
|
|
|
|
|
| |
This fixes blt for parallel builds. It should be upstreamable to the
blt project.
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
|
|
| |
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
|
|
|
|
|
|
|
|
| |
Although recent automakes provide a handy-dandy AM_PROC_AR macro,
we are stuck with autoconf here, and can't use it. Instead
of injecting the value via sed-script, we add an argument and
provide it to the configure script. This will come in handy
later during multilib-utization.
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
|
|
|
|
|
|
| |
Although this is a bit of a dirty hack, it is less so than injecting the
libdir via sed-script. This will come in handy later for multilib
support.
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
|
|
|
|
|
|
|
|
|
| |
Stop using the deprecated AC_OUTPUT; use AC_CONFIG_FILES instead.
The variable substitution I've added is explained here:
http://www.gnu.org/software/automake/manual/html_node/Requirements.html
This can/should be upstreamed to blt.
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
|
|
|
|
|
|
| |
There is no need to inject ranlib into configure via sed.
Just provide the RANLIB environment variable at configure
time and the right thing will happen.
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
|
|
| |
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
|
|
|
|
|
|
| |
Nobody seems to care about bug 485880 -- anyhow no reason to subject
users of my overlay to it. I should figure out who to nudge about
getting this merged.
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
|
|
|
|
|
|
|
|
|
| |
Since some of our patches modify configure, but some do not, and we
exepect configure to be regenerated from configure.in, this makes the
ebuild quite sensitive to subtle patch ordering issues.
To forestall any possible confusion during future versions of the
ebuild, simply remove configure before running eautoreconf.
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
|
|
|
|
|
|
|
|
|
| |
The old solution to LDFLAGS required us to inject the LDFLAGS at
make-time. This improves things by picking them up at configure
time, like we're supposed to. It also makes sure the LDFLAGS apply
to all of the build directories, although this probably doesn't
matter. Finally, it implements an equivalent fix in the cygwin
makefile.
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|
|
Signed-off-by: Gregory M. Tuner <gmt@be-evil.net>
|