Those are a wrapper around xbps_{array,dictionary}_internalize_from_zfile()
that prints a debugging msg when the plist file cannot be internalized.
Update xbps to use these wrappers.
We use a simple file lock that is created with O_CREAT|O_EXCL.
This should fix the concurrency issues with multiple processes
running xbps-rindex -a/-c on the same repository/arch combo.
At the time we've been searching for the pkg in a repo, no repos
were registered resulting in EINVAL. While here, if there's no
declared repos return ENOENT instead.
There's no reason to make them absolute, simply store in the metadata
the target file as is. This vastly simplifies the code and makes all
test pass correctly.
If xbps-create(8) did not guess the target file of relative symlinks for
some reason, just compare the current symlink and what's stored as is,
without converting it to absolute.
This might happen with dangling relative symlinks or existing binary
packages that were not created with a newer xbps-create(8).
If xbps_repo_open() was called with the lock arg set, xbps_repo_close()
will now unlock the repo file lock, without the need to set it.
This avoids the need to always unlock the file lock even if it wasn't
locked previously. This also introduceds an ABI/API break, but this
way it's cleaner.
- Simplify xbps_repo_open::repo_get_dict().
- Use xbps_end() in the utils where necessary.
- Make xbps_end() call xbps_pkgdb_unlock() if necessary.
- Make xbps_end() release rpool resources.
- Make xbps_end() release resources from xbps_handle.
- Fixed 90% of reported leaks (still reachable at exit) from valgrind.
That was to silence valgrind's memcheck with --leak-check=full.
-i, --ignore can be specified multiple times and can be used to
ignore configuration of those packages while configuration of all
packages is being performed.
Close#67
This explicitly enables the in memory fetch/store of remote repository
data archives mode, ignoring existing on-disk repodata archives.
This changes the previous behaviour of falling back to this mode if no
on-disk repodata archives were found.
Thanks to @Gottox and @dominikh for comments.
These routines return a xbps_array_t with a full sorted dependency graph
for the target pkg, by querying pkgdb or rpool.
Update xbps-query(8) to use the new libxbps API.
This allows you to print to stdout any file stored in a binary package,
locally or remotely!
$ xbps-query -R --cat=/usr/bin/ls coreutils > ls
$ file ls
ls: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.32, BuildID[sha1]=7a195fc46d1d5cdca32bfccd3b30f81784e342ed, stripped
$
- This mode prints to stdout the matching FILE stored in a binary package.
- ABI break: renamed xbps_get_pkg_plist_from_binpkg() xbps_binpkg_get_plist().
- Added xbps_binpkg_get_file() as a generic way to get pkg file contents.
- Removed useless comments from xbps_api_impl.h.
The behaviour of this routine mimics the existing xbps_array_add() with
the difference that stored objects are moved to the right to insert
our object as the first element on the array.
Use this to add replaced packages in the transaction array at the head
rather than at the end, to preserve the proper sorting order.
Rather than using a random buffer from stack or heap, and decide
what size to use, create a private memory mapped object...
This simplifies the code in lib/verifysig.c and xbps-create.
This effectively checks that all pkgs in a transaction to be installed
or updated, have not unmatched shared libraries, which might happen with
pkgs that had major soname bumps.
This implementation slow downs xbps_transaction_prepare() 2x, due to
the use of xbps_pkgdb_get_pkg_metadata(). A new pkgdb format is required
that adds the pkg shlib-provides objects to pkgdb, avoiding the need
to use xbps_pkgdb_get_pkg_metadata().
- xbps_binpkg_pkgver() returns a heap allocated string with the pkgver component.
- xbps_binpkg_arch() returns a heap allocated string with the architecture component.
A basename, full path or relative path is supported, i.e:
/path/to/foo-1.0_1.x86_64.xbps
../../foo/blah-0.1_1.x86_64.xbps
baz-0.1_1.x86_64.xbps
- xbps_repo_open() accepts a third argument (bool) to acquire a POSIX file
lock on the repository archive.
- xbps_repo_close() accepts a second argument (bool) to release a POSIX file
lock on the repository archive.
This avoids the issue of multiple xbps-rindex(8) processes being blocked
even for different repositories on the same architecture, resulting in
unnecessary contention.
The reason is that even if the pkg was not configured, it should still be accepted
as installed. If installing packages via XBPS_TARGET_ARCH, pkgs are never configured,
so this must be taken into account.
Will be cherry-picked to 0.37 meanwhile.
If a package that is going to be installed or updated contains invalid
dependencies return ENXIO and XBPS_STATE_INVALID_DEP xbps state to clients.
This improves the error messages returned to the clients when such
condition happens.
The system virtualpkg directory set to <rootdir>/usr/share/xbps/virtualpkg.d contains
virtualpkg configuration files (.conf/.vpkg) that can be overrided by the admin
in <rootdir>/etc/xbps/virtualpkg.d bearing the same file name.
This obsoletes the "virtualpkgdir" keyword support from the xbps configuration file.
The system repository directory set to <prefix>/share/xbps/repo.d contains
system repository configuration files (.conf) that can be overrided by the admin
in <sysconfdir>/xbps/repo.d bearing the same file name.
- Rather than using a POSIX named semaphore use a POSIX lock (lockf(3))
for pkgdb for writers. Writers that cannot acquire the pkgdb lock will
get EAGAIN rather then being blocked.
- Due to using a file lock we cannot write the pkgdb every time a package
is being unpacked, configured or removed. Instead pkgdb is only written
at the end of a specific point in the transaction (unpack, configure, remove)
or via xbps_pkgdb_unlock().
This function is similiar to xbps_fetch_file(). In contrast to xbps_fetch_file()
xbps_fetch_file_dest has an extra paramenter which allow to define an output file
for the request.
The variables to set cachedir, rootdir and metadir have been
changed to "array of chars", this way there are no extra allocations.
Update clients accordingly and bump API version.
The previous internal "struct rpool" was an extra structure that
can be avoided by just using "struct xbps_repo" directly.
This makes rpool use (at least) 4KB less per repository and 1
extra allocation.
- Repository keys are now stored in a new directory on metadir (/var/db/xbps):
<metadir>/key>
- Repository keys are stored with the hex fingerprint of its RSA
public key in a plist dictionary:
<metadir>/keys/xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx.plist
- Drop xbps-rkeys(8) and merge its functionality into xbps-install(8) and
xbps-query(8).
- xbps-query(8) -vL now shows some more details of remote repositories:
3134 http://localhost:8000 (RSA signed, verified)
Signed-by: Void Linux
4096 60:ae:0c:d6:f0:95:17:80:bc:93:46:7a:89:af:a3:2d
16 http://localhost:8000/nonfree (RSA signed, verified)
Signed-by: Void Linux
4096 60:ae:0c:d6:f0:95:17:80:bc:93:46:7a:89:af:a3:2d
Bump XBPS_API_VERSION.
- Remove xbps_repo_get_plist() and try to internalize all members at
xbps_repo_open() time.
- Added xbps_repo_open_idxfiles() to also internalize the index-files
plist from repository, which is really huge and must only be internalized
when needed.
- Improve how signed and verified repositories are detected.
- Misc optimizations and small performance improvements.
Bump XBPS_API_VERSION.
In some tasks the single threaded implementation outperms the multithreaded
one. Use it where it really makes a difference. The _multi() routines do not
spawn any thread if _SC_NPROCESSORS_ONLN == 1.
Bump XBPS_API_VERSION.
