| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
Signed-off-by: Justin Lecher <jlec@gentoo.org>
|
|
|
|
| |
Signed-off-by: Justin Lecher <jlec@gentoo.org>
|
|
|
|
| |
Signed-off-by: Justin Lecher <jlec@gentoo.org>
|
|
|
|
| |
Signed-off-by: Justin Lecher <jlec@gentoo.org>
|
|
|
|
|
|
| |
Convert all URLs for sites supporting encrypted connections from http to https
Signed-off-by: Justin Lecher <jlec@gentoo.org>
|
|
|
|
|
|
| |
Convert all URLs for sites supporting encrypted connections from http to https
Signed-off-by: Justin Lecher <jlec@gentoo.org>
|
|
|
|
| |
Package-Manager: portage-2.2.20.1
|
|
|
|
|
|
|
|
|
| |
symlink #511474
The code path already does nothing whenever /etc/localtime is a symlink,
so deleting the different ways we message the user in favor a shortcut
at the start should be fine. It also means you don't get misleading
messages when running systemd.
|
| |
|
|
|
|
|
| |
Package-Manager: portage-2.2.20.1
Signed-off-by: Lars Wendler <polynomial-c@gentoo.org>
|
|
|
|
| |
Signed-off-by: Justin Lecher <jlec@gentoo.org>
|
|
|
|
|
|
| |
The only library pam uses pkg-config to look up is libtirpc, and we
use that only when USE=nis. Depend on pkg-config only when that is
enabled to avoid circular dependencies (especially when bootstrapping).
|
| |
|
| |
|
|
|
|
| |
Package-Manager: portage-2.2.20_p134
|
| |
|
| |
|
| |
|
|
|
|
| |
Package-Manager: portage-2.2.20.1
|
| |
|
| |
|
|
|
|
|
|
|
| |
There are more cases where a version of tic is needed that matches the
current version of source beyond cross-compiling. New installs, multilib,
and upgrades also run into this case. Drop the cross-compile test and run
the code whenever the host version isn't in sync.
|
| |
|
|
|
|
| |
Package-Manager: portage-2.2.20.1
|
|
|
|
|
| |
Package-Manager: portage-2.2.20.1
Signed-off-by: Justin Lecher <jlec@gentoo.org>
|
|
|
|
| |
Package-Manager: portage-2.2.20.1
|
|
|
|
| |
Package-Manager: portage-2.2.20.1
|
|
|
|
| |
Package-Manager: portage-2.2.20.1
|
|
|
|
|
| |
Package-Manager: portage-2.2.20.1
RepoMan-Options: --include-arches="x86"
|
|
|
|
|
|
| |
The progs change had a subtle effect that broke multilib builds.
A few ticlib headers aren't installed when progs are disabled.
Add a patch to always install them regardless of the progs state.
|
|
|
|
|
| |
Now that we support the new ABI 6, we can support installing the threaded
versions of ncurses alongside the non-threaded.
|
|
|
|
|
|
| |
Rather than maintain parallel dirs, create subdirs of the main BUILD_DIR.
This makes the overall flow a bit simpler as we can also stick the cross
logic in there too.
|
|
|
|
|
|
| |
This gives us a nice speed boost as the configure script is quite large.
Follow up changes will make this more apparent (when we run configure up
to 4 times instead of just 2).
|
|
|
|
|
|
| |
Since the configure script has an option to disable progs, leverage that
instead of whiting out a random make variable. It's the same result, but
feels cleaner and requires less code.
|
|
|
|
|
| |
If the user isn't going to run the testsuite, there's no point in building
all the test binaries, so put that logic behind the standard USE=test.
|
|
|
|
|
|
| |
Since upstream has a configure flag for this now, we should be able to
build+install the C++ layers as shared libs instead of only making the
static libs available.
|
|
|
|
|
|
|
| |
Block the emul package all the time now that it's no longer in the tree.
Unify the gpm deps since there's already a circular dep here and the
multilib code makes no real difference in that respect.
|
|
|
|
|
|
|
|
|
| |
The bare min changes are here to get it working as more updates will
follow. Focus is on making the base ABI match upstream defaults for
this release -- namely that we update to ABI 6 which includes colors
and other new extended features. #373767
Based on work by Lars Wendler.
|
|
|
|
|
|
|
|
| |
Since the gpm code relies on running ldd on the linked file in order to
calculate the SONAME, and we're passing this in as a configure flag some
of the time, just change the code to always pass it in. This stabilizes
behavior across cross-compiling/multilib/ldd output/etc... Relying on
ldd output in general is bad juju.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Package-Manager: portage-2.2.20
|
|
|
|
| |
Package-Manager: portage-2.2.20
|
|
|
|
| |
Package-Manager: portage-2.2.20
|
| |
|
|
|
|
| |
Package-Manager: portage-2.2.20
|
|
|
|
| |
Package-Manager: portage-2.2.20
|
|
|
|
| |
Package-Manager: portage-2.2.20
|