It's quite common to get the following message after aborting a pacman command, often after system restarts:
:: Synchronizing package databases...
error: failed to synchronize all databases (unable to lock database)
Right now, the locking mechanism is relatively rudimentary: if db.lck exists, then the database is considered locked. If something happens to the pacman process and it cannot remove the db.lck, it will be locked indefinitely, or until somebody Googles "pacman unable to lock database" and removes the lock.
This commit makes db.lck keep track of the process that locked it. If subsequent pacman processes cannot find the process that originally locked the database, the database will be considered unlocked.
Debug symbols should only be split from finally linked ELFs, not bare
object files. We're already excluding static libraries from splitting
for a similar reason.
The `.gnu_debuglink` sections are also mishandled by LLVM's LLD, which
copies them to its output. For example, this affects Arch Linux's
`/usr/lib/Scrt1.o`.
While we're here (and it changes the code less), also strip GNU LTO data
from bare objects, again for the same reason we're removing it from
static libraries, and apply static library stripping instead of shared
library stripping.
See: https://bugs.gentoo.org/787623
When objcopy encounters an already-present section, adding the new
debuglink will fail with a warning. Remove any existing `.gnu_debuglink`
section to work around this problem.
Arch Linux's `rust` package is affected by this. Apparently when LLVM's
LLD links in `/usr/lib/Scrt1.o` it will also copy the `.gnu_debuglink`
section.
See: https://bugs.gentoo.org/787623
Using objcopy can result in file permission changes. We work around this
by using "cat" to copy the temporary output file into the target. Extract
this code into a utility function.
Signed-off-by: Allan McRae <allan@archlinux.org>
For example, with landlock ABI < 3, LANDLOCK_ACCESS_FS_TRUNCATE is not
set in ruleset_attr.handled_access_fs, so it should not be set in
path_beneath.allowed_access either. Otherwise, landlock_add_rule fails
with -EINVAL, and pacman complains:
> error: restricting filesystem access failed because the landlock rule for the temporary download directory could not be added!
The change is tested on Debian Bookworm kernel
linux-image-6.1.0-25-cloud-amd64 6.1.106-3.
Some libaplm utilities sync databases as a non-root user for use in
actvities other than system updates. The ability to download as a
non-root user was broken as part of the download sandboxing.
Applying a minimial fix by preventing the chown of the downloaded file
if the user is non-root. A larger change increasing the robustness
and error checking of this path is warranted in the future.
Signed-off-by: Allan McRae <allan@archlinux.org>
Sorting modifies the list in place, causing any existing pointers to the
list to point to a random element.
Fixes#165
Signed-off-by: Andrew Gregory <andrew.gregory.8@gmail.com>
Git commands can fail in bare repositories when global git config contains safe.bareRepository=explicit.
Some users set this option for increased security.
To be compatible with this configuration, explicitly set safe.bareRepository=all when invoking git in a bare repository.
In f91fa546 (repo-add: do not recreate the database if nothing was changed),
repo-add was made to skip database write-out if there were no changes to
the database. However, this breaks the usage of repo-add to create a new
empty database: `repo-add /path/to/mydb/mydb.db.tar.xz`.
Bring back support for this use-case by always writing the database if
it is missing.
Original-patch-by: Ivan Shapovalov <intelfx@intelfx.name>
Signed-off-by: Allan McRae <allan@archlinux.org>
The Arch sharutils package was spewing messages about "Permission denied" when
copying source files into the debug package. This is due to the source files
having 444 permissions and being used in multiple binaries. Only copy each
source file into the debug package onces to avoid this error.
Signed-off-by: Allan McRae <allan@archlinux.org>
If a package was already downloaded but its signature file was not,
pacman would download the signature then error out despite all files
being present.
Also fixes a similar error when some, but not all, package databases
were updated with -Sy.
Fixes#156
Signed-off-by: Allan McRae <allan@archlinux.org>
The calculation used the size of the package rather than the amount
remaining to download for partially downloaded packages.
Signed-off-by: Allan McRae <allan@archlinux.org>
On Windows based systems (e.g. msys), running PKGBUILD linting is very
slow due to time taken spawning bash subshells. Additionally, some packages
have extreme amounts of (usually procedurally generated) splitting, which
also causes linting to be extremely slow. Provide an environment variable
to disable PKGBUILD linting.
Signed-off-by: Allan McRae <allan@archlinux.org>
Supporting git source fragments (branch, commit, tag) is difficult in
conjunction with GITFLAGS usage - particularly with the most common
use cases that reduce the amount of data cloned from the upstream repo.
Leaving GITFLAGS in place an documenting that various git source features
are not supported when GITFLAGS are in used is not an ideal 'solution'.
Instead, remove GITFLAGS support.
Signed-off-by: Allan McRae <allan@archlinux.org>
This causes issues when repeatedly building a package using the same
git checkout. There is also ambiguity of the default checkout when
trying to build from HEAD. See #142 and #143.
This reverts commit 85c421f1cb.
Add the -k parameter to the sudo call to prevent caching of credientials.
This would (potentailly) stop a rogue sudo use within a PKGBUILD.
Signed-off-by: Allan McRae <allan@archlinux.org>
Add a new error code to expose the 'not a clone of' error state of some source
providers (git and fossil). This allows other tools integrating further and
handle this specific error state.
One usecase evolves around frequently changing source locations in PKGBUILDs
of packages in the AUR.
It's extremely hard to figure out what's going from when bsdtar fails
here when we expect it to succeed. Stop tunneling stderr to /dev/null
to help users figure out what's going on when this fails.
