| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
Signed-off-by: Matt Turner <mattst88@gentoo.org>
|
|
|
|
| |
Signed-off-by: Matt Turner <mattst88@gentoo.org>
|
|
|
|
| |
Signed-off-by: Matt Turner <mattst88@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
https://github.com/gentoo/gentoo/pull/25896 tried to address cross-compiling in
the wrong way by setting the runtime sysroot and the libdir. The build uses the
latter to find dependent libraries by default, a common technique in GNU
software.
This behaviour can be avoided by passing the "without-prefix" configure option
for each library, allowing the runtime sysroot and the libdir to be set as they
should be. Contrary to what you might expect, the sysroot should actually not be
set at all except when building cross-*/gdb, otherwise the paths get double
prefixed at runtime.
gdb's configure script also has a --with-build-sysroot option that should be set
to assist with cross-compiling. This does not influence runtime behaviour.
The includedir for cross-*/gdb has been changed to prevent it from installing
under /usr/${CTARGET}, where installed headers would collide with those from a
cross-compiled sys-devel/gdb build. /usr/include/${CTARGET}/gdb is consistent
with the files we install under /usr/share, as well as the scheme we use for
multilib headers, e.g. /usr/include/i686-pc-linux-gnu.
Closes: https://bugs.gentoo.org/896008
Signed-off-by: James Le Cuirot <chewi@gentoo.org>
Closes: https://github.com/gentoo/gentoo/pull/29819
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When bootstrapping on a 3.10 kernel (e.g. CentOS 7.x), libgcrypt notes the following:
```
* Messages for package dev-libs/libgcrypt-1.10.1-r3:
* The getentropy function requires the getrandom syscall.
* This was introduced in Linux 3.17.
* Your system is currently running Linux 3.10.0-1160.45.1.el7.x86_64.
* Disable the 'getentropy' USE flag or upgrade your kernel.
* ERROR: dev-libs/libgcrypt-1.10.1-r3::gentoo failed (pretend phase):
* Kernel is too old for getentropy
```
Since without getrandom it will do a runtime check it's ok to mask.
Signed-off-by: Bart Oldeman <bart.oldeman@calculquebec.ca>
Closes: https://github.com/gentoo/gentoo/pull/29906
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Matt Turner <mattst88@gentoo.org>
|
|
|
|
|
|
| |
Signed-off-by: Guillermo Joandet <gjoandet@gmail.com>
Closes: https://github.com/gentoo/gentoo/pull/29914
Signed-off-by: Matt Turner <mattst88@gentoo.org>
|
|
|
|
|
|
| |
Signed-off-by: Guillermo Joandet <gjoandet@gmail.com>
Closes: https://github.com/gentoo/gentoo/pull/29915
Signed-off-by: Matt Turner <mattst88@gentoo.org>
|
|
|
|
|
| |
Closes: https://bugs.gentoo.org/893390
Signed-off-by: Matt Turner <mattst88@gentoo.org>
|
|
|
|
|
| |
Bug: https://bugs.gentoo.org/893390
Signed-off-by: Matt Turner <mattst88@gentoo.org>
|
|
|
|
|
|
|
| |
The assert-msg-test.py test doesn't work under sandbox.
Closes: https://bugs.gentoo.org/871330
Signed-off-by: Matt Turner <mattst88@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Mike Pagano <mpagano@gentoo.org>
|
|
|
|
| |
Signed-off-by: Mike Pagano <mpagano@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Bernard Cafarelli <voyageur@gentoo.org>
|
|
|
|
| |
Signed-off-by: Bernard Cafarelli <voyageur@gentoo.org>
|
|
|
|
| |
Signed-off-by: Bernard Cafarelli <voyageur@gentoo.org>
|
|
|
|
| |
Signed-off-by: Bernard Cafarelli <voyageur@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
| |
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
|
|
| |
Closes: https://bugs.gentoo.org/885485
Closes: https://bugs.gentoo.org/895496
Signed-off-by: Bernard Cafarelli <voyageur@gentoo.org>
|
|
|
|
| |
Signed-off-by: Bernard Cafarelli <voyageur@gentoo.org>
|
|
|
|
|
|
|
| |
Also remove now useless "-Dwerror=false"
Closes: https://bugs.gentoo.org/897714
Signed-off-by: Bernard Cafarelli <voyageur@gentoo.org>
|
|
|
|
|
| |
Closes: https://bugs.gentoo.org/894582
Signed-off-by: Maciej Barć <xgqt@gentoo.org>
|
|
|
|
|
|
|
| |
Builtins don't work correctly without this.
Closes: https://bugs.gentoo.org/892956
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Without this, statically linking even a basic example fails because
we're using llvm-libunwind rather than GCC's internals.
Quoting myself on the bug:
"""
FWIW the LLVM/Clang profiles are considered pretty experimental still.
For Clang, it's tricky because it depends on configuration and the (default) flags passed.
Ionen points out it should be harmless dep-wise (non-invasive to users, no need to then
enable static-libs for its deps) to default to +static-libs on llvm-libunwind,
which would cover most cases here. I think that generally makes sense.
I see three options:
1. Adding IUSE=llvm-libunwind to virtual/libc
2. Defaulting on (which doesn't feel like a full solution alone) +static-libs in llvm-libunwind
3. Making clang-common[default-compiler-rt] depend on libunwind[static-libs]
(or correspondingly llvm-libunwind[static-libs] if USE=llvm-libunwind)
I think 2+3 is probably a good fit?
"""
Bug: https://bugs.gentoo.org/892956
Signed-off-by: Sam James <sam@gentoo.org>
|