This commit introduces the preservation of SSH_AUTH_SOCK within the
chroot environment, to support SSH-based operations, such as cloning
repositories via SSH.
Introduce setarch-aliases.d/ which gets installed inside
/usr/share/devtools. This allows to assign aliases which map one CARCH
to a different name which gets provided as argument for setarch. This is
necessary on archlinuxarm ("armv6h" -> "armv6l", "armv7h" -> "armv7l")
and allows for more fine-grained architectures (e.g. archlinux32 has
"i686" and "pentium4", which differ in the required cpu capabilities).
Signed-off-by: Levente Polyak <anthraxx@archlinux.org>
The progress bar feature creates noisy log files, especially with
parallel downloads. Lets disable the progress bar and instead use
parallel downloads.
This effectively does nothing but keep the commented options in sync to
not deviate from the original file for no reason.
Signed-off-by: Levente Polyak <anthraxx@archlinux.org>
The only effective difference is that -Wp,-D_GLIBCXX_ASSERTIONS is now
only defined for CXXFLAGS as of pacman 6.0.0-1
Signed-off-by: Levente Polyak <anthraxx@archlinux.org>
Outdated sogrep cache has already bit us once in the past. Finding one
or more databases older than a day is a good indication that a refresh
is in order, so do that automatically.
Extracting these databases is painfully slow on HDDs (especially laptop
ones). There shouldn't be a drawback to keeping the tarballs around and
extracting them to a temporary directory (usually tmpfs) to parse them.
The implemented update logic tries to avoid redownloading unchanged dbs.
After further followups always moving all products isn't actually
desired as they can theoretically be broken in various ways if
arch-nspawn exists non successful.
However, as we would like to always preserve the produced log files we
instead split out the logfiles into an own function and call that for
unsuccessful buils.
Fixes 4f305aa3
Commit 09e169b741 changed the default
build host from dragon.archlinux.org to build.archlinux.org - match
in documentation.
Signed-off-by: Christian Hesse <mail@eworm.de>
We now accept:
1) # nothing
in which case we'll use the PKGBUILD to retrieve...
2) name, or repo/name
in which case we'll use pacman to cache the package and retrieve...
3) a filename
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Teach get_pkgfile to call itself in local-only mode and find a cached
file no matter what its extension is. Avoids repetitively trying to curl
random files, fail with 404 errors, and proceed to discover a cache hit
under a different file extension.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Since move_products() function is fairly robust we can make it run for
failed build also to expose logs for packages that fails in build(),
prepare() or package(). It also exposes partially packaged split
packages if they fail in latter package_xxx().
Building a package may change the PKGBUILD during update_pkgver. Let's
retrieve the PKGBUILD after building to ensure we have the very same
file as the one we used to build the package. Otherwise this may lead to
the inability to distribute the package during commitpkg in case the
expected and the actual hashsum mismatch.
First try a .zst location before falling back to legacy variants. This
should slightly speed up downloading of dependencies, especially over
time as .zst packages are or will be the dominant format.
Signed-off-by: Levente Polyak <anthraxx@archlinux.org>
This was incorrectly implemented in commit
0067176529, which added the host_mirrors
root directory as a cachedir, when we actually want to use the pool/*
subdirectories (the same ones installed on the build server's
/etc/pacman.conf).
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Levente Polyak <anthraxx@archlinux.org>
We don't want to check against the current version known to the host
system, because that will be incorrect in a wide variety of situations,
including:
- the build host hasn't done a full system upgrade yet
- we're building against staging, and want to see the delta between
different staging versions
- we're building against extra, but the host runs testing which carries
changes we don't want to visualize right now
- the chroot has a configured database not available to the host, and
the package is only available there
Essentially, it's rarely 100% correct to run checkpkg on the host, but
we already have a database we *know* is correct, and that is the one we
just built the package against. So let's use that.
This also fixes a bug in the current logic, where in order to try
downloading fresh databases, we work in a non-cached temporary working
database to download the package files, but then let checkpkg default to
comparing packages in the system database. Since we are explicitly
trying to compare against packages that differ from the host's pacman
database, we need to pass the package files as options to checkpkg,
using the additional modes added in commit c14338c0fe
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Levente Polyak <anthraxx@archlinux.org>
Thanks to anthraxx for the guidance.
Original-patch-by: Levente Polyak <anthraxx@archlinux.org>
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Levente Polyak <anthraxx@archlinux.org>
By specifying multiple package files, we assume they are all from the
same PKGBUILD, and try to check them all against the produced artifacts.
Since the buildinfo should be comparable for all of them, we simply use
the first one passed on the command line.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Levente Polyak <anthraxx@archlinux.org>
This ensures we take user specific config values for PKGDEST into
account when printing the package list. This is required as devtools
archbuild_cmd puts packages potentially into the user defined PKGDEST
which the package list would otherwise miss.
This fixes an issue with the usage of makepkg --packagelist to get the
produced artifacts filenames according to the PKGEXT used in devtools'
makepkg.conf instead of the one defined in pacman.
One goal we want to preserve is that devtools configuration should be
self contained and not require any editing of non owned files like
the host /etc/makepkg.conf to produce expected results.
Additionally modify the archbuild_cmd override for multilib builds to
use an independent variable and not fiddle with the actual arch
variable to select the appropriate cmd.
This ensures we use the same configuration for reproducing packages as
we use for building them via devtools.
One example of why we care about this are the COMPRESS* settings that
may differ from the guest's pacman shipped makepkg.conf that affect the
reproducibility of packages.
We don't want the default PKGEXT in the current version of devtools, we
want the PKGEXT we *know* the input file used.
Signed-off-by: Levente Polyak <anthraxx@archlinux.org>
We need to modify the matching of valid package files to support formats
like zstd. Let's try to use an eager approach instead of a simple
whitelist in order to be functional for arbitrary formats that may be
introduced in the future without the need to adjust any code.
Allow any single fragment word as compression type but filter out known
non-package content like signature files.
Signed-off-by: Levente Polyak <anthraxx@archlinux.org>
David Runge
Benjamin Denhartog
yoursweetie
Jelle van der Waa
tsuibin
Felix Yan
bartus
Levente Polyak
Erich Eckner
Chih-Hsuan Yen
Eli Schwartz
Evangelos Foutras
Christian Hesse
bartoszek
Konstantin Gizdov
Robin Broda