| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
This can be used to pick another tool to elevate privileges.
|
|
|
|
|
|
|
|
|
| |
This way Meson and CMake will use the C{,PP,XX}FLAGS we've set in
`/etc/abuild.conf`. Without this change CMake and meson will overwrite
our CFLAGS with custom flags. Most importantly, they prefer -O1/-O2 over
our -Os.
See also: https://lists.alpinelinux.org/~alpine/devel/%3C2896c13070c508a49cbaa72c8fb7f34ea947358b.camel%40cogitri.dev%3E
|
|
|
|
|
| |
- Call "$abuild_path" instead of plain "abuild"
- Pass $forceroot as done elsewhere
|
|
|
|
|
|
| |
CARGO_HOME tells cargo where to store installed dependencies, save it to
a directory in $SRCDEST so we don't need to download all dependencies
again when compiling a rust package.
|
|
|
|
|
| |
It is recommended by upstream CMake (it will warn when you do in-source
builds) to keep generated files away from the source.
|
|
|
|
|
|
|
| |
This makes it easier to figure out to which build the directory belongs
to. Occasionally, I have many failed abuild rootblds in my /var/tmp and
including the $pkgname in the directory would help me associating the
directories with failed builds I recently executed.
|
|
|
|
|
|
|
|
| |
Fixes the error:
df: .: can't find mount point
When running abuild inside a chroot when the root file system
mountpoint information is not necessarily directly available.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Originally "gcc -dumpmachine" was used to detect build gcc triplet.
However, abuild does not depend on gcc or build-base (but installs
it if needed to build) so gcc might not be there. Additionally
abuild-sign can be used standalone, and does not have gcc dependency.
Using ${CC:-gcc} is problematic in cross-compile, as CC might be
already set for the cross-compiler and would result giving the target
host triplet.
It was deemed simplest to use "apk --print-arch" exclusively to detect
the builder host type, or specify CBUILD manually. If there is need
to use abuild/abuild-sign on non-Alpine hosts withou apk, we can
later add fallback that uses "uname -m" to detect the architecture
and guess Alpine CBUILD from it.
Fixes #9974
Fixes: 5adf47c1 "functions.sh: use apk --print-arch for CARCH if gcc is missing"
Fixes: 95cd15c0 "functions.sh: dont die if gcc is missing"
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
Previously, a key size of 2048 bits was hardcoded. While this is still
the default, it can now be changed. Additionally, the default key size
might be changed to 4096 in the future.
|
| |
|
| |
|
|
|
|
| |
fixes #9981
|
|
|
|
|
|
| |
fix patches such as https://dev.alpinelinux.org/archive/domoticz/openzwave-1.6.patch
previously it would only work with the filenamename.patch::$url syntax
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Without this change abuild rootbld would fail with:
touch: invalid date '@'
Because SOURCE_DATE_EPOCH wasn't set when abuild rootbld was used. This
is a bug introduced in 71d9d5233b9db3be91510addcb28721545d93185. Instead
of reverting the aforementioned commit move the SOURCE_DATE_EPOCH
initialization to a custom function and also call it from the abuild
rootbld function.
Fixes #9978
|
|
|
|
|
|
|
| |
The introduction of the --pax-options seems to confuse apk and resulted
in `BAD archive' errors.
This reverts commit f04a2ee34b28a38c4349ef1f94686a07afce730f.
|
|
|
|
|
|
| |
abuild-sign does not use gcc.
fixes #9974
|
| |
|
|
|
|
| |
fixes !7
|
|
|
|
|
|
|
|
|
| |
Currently the permissions for abuild-sudo is set to 4111, this make it
impossible to move the sudo file after it has being created. Moving
the sudo file is needed by yocto when creating an apk-package of abuild.
The sudo binary in debian stretch since if debian does it, it can't be bad.
Signed-off-by: Fredrik Gustafsson <fredrigu@axis.com>
|
| |
|
|
|
|
|
| |
getting the commit date can be timeconsuming so only do it once we need
it. We also re-use the ABUILD_LAST_COMMIT to speed up the operation.
|
|
|
|
| |
rename the global variable to upppercase.
|
|
|
|
|
|
| |
Add -- to explicitly separate out the file path.
Let git_last_commit_epoch take an option with the hash.
|
|
|
|
|
| |
Flags to make the tarball reproducible is taken from here:
http://h2.jaguarpaw.co.uk/posts/reproducible-tar/
|
|
|
|
| |
This improves reproducibility of builds
|
|
|
|
|
| |
This is similar to what we do in abuild. It improves performance on
multicore machines.
|
|
|
|
|
| |
set datestamps to be used in the built packages to date of commit. This
makes it much easier to have reproducible builds.
|
| |
|
|
|
|
| |
Signed-off-by: Richard Mortier <mort@cantab.net>
|
|
|
|
|
| |
Aloow overrid sharedir with global ABUILD_SHAREDIR so we test the local
functions.sh instead of a system installed functions.sh
|
|
|
|
|
| |
abuild uses datadir as local variable in various functions. Rename the
global datadir to sharedir to avoid confusion.
|
| |
|
|
|
|
|
| |
GitHub allows to reference https://github.com/<user>/<proj>/archive/<ver>.tar.gz
also as https://github.com/<user>/<proj>/archive/<ver>/<anything>.tar.gz.
|
|
|
|
|
| |
This forces cargo to install the exact versions of the dependencies
specified in Cargo.lock. This is essential for reproducible builds!
|
|
|
|
|
| |
the function is not used since commit 3379e675512d (abuild: print
version of built package early)
|
| |
|
|
|
|
|
| |
we set sysconfdir in functions.sh so there is no need to set it in
abuild.
|
|
|
|
| |
The assignment of the oldpkg variable is missing the name of the package
|
| |
|
|
|
|
| |
Prereqs from the 'configure' phase also belong in makedepends
|
|
|
|
| |
The 'version' string from CPAN can start with 'v...' or similar.
|
| |
|
| |
|