In case we have a mirror failure, unlink_on_fail would remain set,
causing an interrupt in a successive download attempt to be wrongly
unlinked.
This also fixes a memory leak in the url member, as we would allocate
over the previous, unfreed URL.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
To avoid conflicts on reusing a payload after a failed download, ensure
that we reset the filename hints in the payload struct prior to the
download operation.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
Similar to how we manage receipt of SIGINT in pacman's internal
downloader, catch the signal and invoke our own trap handler before
unsetting it and calling the default. This requires a slight
modification to the arguments passed to trap_exit so we can pass the
raised signal to trap_exit (note that we substitue USR1 for ERR since
the latter is unique to bash).
Fixes FS#28491.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
Verifing the keyring at this point is useless as a malicious package is already
installed and as such has several options to bypass this check anyway.
Signed-off-by: Pierre Schmitz <pierre@archlinux.de>
Signed-off-by: Dan McGee <dan@archlinux.org>
* it updates to all translations
* minor fr, pt_BR, de, lt, sk and uk updates
* add new strings in pacman translation catalog
Signed-off-by: Dan McGee <dan@archlinux.org>
Passing the "-L" flag to stat means we get the size of the file
being pointed to for symlinks instead of the size of the symlink.
Keep "-L" usage in repo-add as we want the actual size of the
package/delta/signature there.
Signed-off-by: Allan McRae <allan@archlinux.org>
With multiple items in $libdepends this check only worked for the first
one, everything after this returned an error. This was probably an issue
with \s being treated wrong.
Fix-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Florian Pritz <bluewind@xinu.at>
libperl.so results in soname="libperl.so.so" which is wrong.
This returns the correct string: "libperl.so"
Fix-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Florian Pritz <bluewind@xinu.at>
If we begin to create a file list when loading a package, but abort
because of an error to one of our goto labels, the memory used to create
the file list will leak. This is because we use a set of local variables
to hold the data, and thus _alpm_pkg_free() cannot clean up for us.
Use the file list struct on the package object as much as possible to
keep state when building the file list, thus allowing _alpm_pkg_free()
to clean up any partially built data.
Signed-off-by: Dan McGee <dan@archlinux.org>
This is easily triggered via a `pacman -Sc` operation when it attempts
to open a delta file as a package- we end up leaking loads of memory
due to us never freeing the archive object. When you have upwards of
1200 delta files in your sync database directory, this results in a
memory leak of nearly 1.5 MiB.
Also fix another memory leak noticed at the same time- we need to call
the internal _alpm_pkg_free() function, as without the origin data being
set the public free function will do nothing.
Signed-off-by: Dan McGee <dan@archlinux.org>
Ensure we give database signatures special treatment like we already did
for package signatures. Attempt to parse the database name out of them
before taking the proper steps to handle their existence. This fixes
FS#28714.
We also add an unlink_verbose() helper method that displays any errors
that occur when unlinking, optionally opting to skip any ENOENT errors
from being fatal.
Finally, the one prompt per unknown database has been removed, this has
no real sound purpose and we don't do this for packages. Simply kill
databases we don't know about; other programs shouldn't have random data
in this directory anyway.
Signed-off-by: Dan McGee <dan@archlinux.org>
Mostly a waste of time. Sure, we no longer make sure your pacman
database partition has enough space, but if you are using this option
you better know what you are doing anyway.
Signed-off-by: Dan McGee <dan@archlinux.org>
(cherry picked from commit ee96900605)
This was noted when trying to perform an operation on a pacman database
on a read-only file system. Print the actual underlying errno string,
and only show the "you can remove" message if the lock file actually
exists.
Before:
$ pacman -Su
error: failed to init transaction (unable to lock database)
if you're sure a package manager is not already
running, you can remove /e/db.lck
After:
$ pacman -Su
error: failed to init transaction (unable to lock database)
error: could not lock database: Read-only file system
Signed-off-by: Dan McGee <dan@archlinux.org>
If one had a mountpoint at '/e' (don't ask), a file being installed to
'/etc' would map to it incorrectly. Ensure we do more than just prefix
matching on paths by doing some more sanity checks once the simple
strncmp() call succeeds.
Signed-off-by: Dan McGee <dan@archlinux.org>
Modify parse_options logic to fill an array instead of printing parsed
options. Avoid eval like the plague. Because it is the plague.
Fixes bugs such as FS#28445.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
Close FS#28434
Signed-off-by: Eric Bélanger <snowmaniscool@gmail.com>
Signed-off-by: Allan McRae <allan@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
This reverts commit f3fa77bcf1 along with
making other necessary changes to fully back this (mis)feature out until
we can do it correctly.
The quick summary here is this was not implemented correctly; provides
are not fully taken into account in this logic, and making that happen
exposes a lot of other flaws in this code that are covered up later on
in the dependency resolving process by several other pieces of
convoluted and conditional logic.
Tests have been adjusted accordingly. Some test EXISTS conditions have
been removed as we already know the package is installed locally, and we
also are checking the VERSION condition anyway.
With these two related revert commits, we do have some changes in test
pass/fail results:
* upgrade078.py: does not pass, this is due to --recursive getting
removed for -U/-S operations after this commit.
* sync302.py: the version checks have been disabled, so this test
continues to pass but has been scaled back in scope.
* sync303.py: now passes, was failing before.
* sync304.py: still failing, was failing before.
* sync305.py: now passes, was failing before.
* sync306.py: still passes, was passing before.
Signed-off-by: Dan McGee <dan@archlinux.org>
This reverts commit 0903452032.
Tests affected by this revert have been adjusted; additionally a few
EXIST tests have been removed where there is already a VERSION test
doing the job for us.
Signed-off-by: Dan McGee <dan@archlinux.org>
This fixes a problem that occurs if you tell makepkg to remove installed
dependencies (just to be sure) but there are none. As the $ was missing
in front of deplist, the check never happened and 'pacman -Rn' was
called which obviously failed.
Dan: later reported as FS#28448.
Signed-off-by: Christoph Vigano <mail@cvigano.de>
Signed-off-by: Dan McGee <dan@archlinux.org>
This wonderful/awful little bash shell variable wrecks havoc on `cd`
calls in shell scripts. Unset CDPATH in makepkg where we use `cd` quite
heavily. In pacman-optimize, we can move the change directory logic into
the bsdtar call so we are left with no usages of `cd` in the script.
Signed-off-by: Dan McGee <dan@archlinux.org>
The max filesize for a delta download must be the full size of the delta
file, not just what's remaining.
Fixes FS#28345
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
PKGBUILDs are advertised as being pure bash so it would be expected
that the default bash options are in effect when sourcing it.
This inadvertantly "fixes" FS#27780 where enabling extglob causes
the bash parser to error on non-valid bash function names like
package_libxml++(). Note that these function names are unsupported
in bash and could break again even with this "fix" in future bash
releases.
Signed-off-by: Allan McRae <allan@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
From mercurial-2.1, "hg pull" returns 1 when there are no updates.
Catch the return code and either pull the update or return an
error as appropriate.
Signed-off-by: Allan McRae <allan@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
This is after some manual massaging to fix issues with newlines in some
translations of the script catalogs.
Signed-off-by: Dan McGee <dan@archlinux.org>
The pacman-scripts catalog is omitted here due to various newline errors
I don't have the time to fix right now.
Signed-off-by: Dan McGee <dan@archlinux.org>
We hardly need the complexity (or slowness) provided by the libm power
function; add a super-cheap one that suits our needs and is specialized
for the values we plan on passing in.
Signed-off-by: Dan McGee <dan@archlinux.org>
Dan: don't compute lower bound unless needed, flip argument order so
out values are last, add param Doxygen documentation.
Signed-off-by: Dan McGee <dan@archlinux.org>
Provide a helpful error message for when creating a source tarball
and SRCPKGDEST is not writable.
Fixes FS#28197.
Signed-off-by: Allan McRae <allan@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
Very rarely a segfault would occur when removing a number of packages
due to a corrupted list for the local database (FS#27805, FS#28195).
This was caused by the alpm_list_msort function not correctly dealing
with the two new head node's prev values.
Signed-off-by: Allan McRae <allan@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
This addresses two problems:
1) echo's behavior is inconsistent when dealing with flags, and can
potentially be problematic.
$ echo -n
$ echo -- -n
-- -n
2) Always using the end of options markers prevents translated strings
from throwing errors, as shown in FS#28069.
The remaining "inconsistencies" are because printf is being used in a
guaranteed safe manner, e.g.
printf '%s\n' "$(gettext "--this can never break")"
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
This "fixes" two issues:
1) MAN_DIRS contains a glob by default so should not be quoted. It is
not quoted anywhere else so this should not cause breakage...
2) the find statement returns 1 when some of MAN_DIRS are missing. This
appears to only be exposed when running makepkg as root (which it appears
some wrappers do...).
Signed-off-by: Allan McRae <allan@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
This is particularly important in the case of FTP control connections,
which may be closed by rogue NAT/firewall devices detecting idle
connections on larger transfers which may take 5-10+ minutes.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
With pacman-4.0, using --noconfirm or --noprogressbar with -Q or -T
results in pacman reporting an "invalid option" error. Restrict the
passing of these options to pacman. Fixes FS#28012.
Signed-off-by: Allan McRae <allan@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
