We were using the stat() system call in quite a few places when we didn't
actually need anything the stat struct returned- we were simply checking for
file existence. access() will be more efficient in those cases.
Before (strace pacman -Ss pacman):
% time seconds usecs/call calls errors syscall
------ ----------- ----------- --------- --------- ----------------
33.16 0.005987 0 19016 stat64
After:
% time seconds usecs/call calls errors syscall
------ ----------- ----------- --------- --------- ----------------
34.85 0.003863 0 12633 1 access
7.95 0.000881 0 6391 7 stat64
Signed-off-by: Dan McGee <dan@archlinux.org>
We used list_display() on -Sg output, which might have been slightly nicer
looking but made it much harder to parse in something like a shell script.
Reformat it in the 'grpname pkgname' format that -Qg is already using.
Signed-off-by: Dan McGee <dan@archlinux.org>
Some previous commits apparently broke the get_filename function for package
loaded with pkg_load (on a -Qip operation) because this field was no longer
filled. Now pkg_load fills it.
But the -Qip operation needs to be run like this : -Qip <filename>, so the
filename is already known. There is no need to display it again.
Besides, on a normal -Qi operation, the filename is not displayed either
because this information is not stored in the local database.
Signed-off-by: Xavier Chantry <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
By putting the search / group / info / list operations just after the -Sy
op, we can simplify several checks :
1) the check for "missing targets". Since we took care of the above
operations, we now have less cases to consider :
* -Syu or -Su : we can proceed
* -Sy : we can end now (this is actually a bugfix)
* -S : this op requires targets, so exit with an error
2) the check to see if a transaction is needed. If we arrive at the end of
the function, it is either because we have -Su or -S <targets> so we already
know a transaction is needed there.
Signed-off-by: Xavier Chantry <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
Add a new totaldlcb callback function to libalpm and make pacman utilize it
when the TotalDownload option is enabled. This callback function is pretty
simple- it is meant to be called once at the beginning of a "list download"
action, and once at the end (with value 0 to indicate the list has been
finished). The frontend is responsible for keeping track of adding
individual file download amounts to the total xfered amount in order to
display some sort of overall progress.
Signed-off-by: Dan McGee <dan@archlinux.org>
We have been using unsigned long as a file size type for a while, which
works but isn't quite correct and could easily break. Worse was probably our
use of int in the download callback functions, which could be restrictive
for packages > 2GB in size.
Switch all file size variables to use off_t, which is the preferred type for
file sizes. Note that at least on Linux, all applications compiled against
libalpm must now be sure to use large file support, where _FILE_OFFSET_BITS
is defined to be 64 or there will be some weird issues that crop up.
Signed-off-by: Dan McGee <dan@archlinux.org>
Add a new configure flag, --enable-git-version, that allows the output of
'git describe' to be used in the version string associated with this
package. This could aid in debugging for users that are using a development
version of pacman and we should be able to figure out which cut of code they
are using.
Sample output:
$ pacman --version
Pacman v3.1.4-190-g4cfa-dirty - libalpm v2.3.1
$ makepkg --version
makepkg (pacman) 3.1.4-190-g5861-dirty
Signed-off-by: Dan McGee <dan@archlinux.org>
If we don't set the pkgname var to NULL, we run into all sorts of beautiful
segfault behavior when a group spans multiple repositories and we try to
print out the location of the former list. Easy fix.
This regression was introduced in bf86700369.
Signed-off-by: Dan McGee <dan@archlinux.org>
This event was unused, was missing the equivalent EXTRACT_DONE event, and
was useless because we already have ADD / UPGRADE START and DONE events.
Signed-off-by: Xavier Chantry <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
This is one of those rare cases where we actually want to code in a
platform-specific #ifdef. Because you don't need to be the root user on a
Windows box, and fakeroot doesn't exist so we can do easy testing, lets
disable any checking of the UID.
Signed-off-by: Xavier Chantry <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
This patch offers a way to fix FS#9228.
By putting "SyncFirst = pacman" in pacman.conf, the version check will
happen before the transaction really starts, and before any replacements is
made.
Otherwise, no version check is done.
The sync301 pactest was updated to use this SyncFirst option.
Example session with SyncFirst = pacman, and a newer pacman version
available :
$ pacman -Su (or pacman -S <any targets>)
:: the following packages should be upgraded first :
pacman
:: Do you want to cancel the current operation
:: and upgrade these packages now? [Y/n]
resolving dependencies...
looking for inter-conflicts...
Targets: pacman-x.y.z-t
Total Download Size: x.xx MB
Total Installed Size: x.xx MB
Proceed with installation? [Y/n] n
As Nagy previously noted, doing this check on any -S operations might look
intrusive, but it can be required.
For example, the case where you want to install a package with versioned
provisions, using a pacman version which didn't support that feature yet
(and there is already a newer pacman in sync db supporting it).
Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
* remove obsolete and unused *_cmp helper functions like deppkg_cmp and
_alpm_grp_cmp
* new alpm_list_remove_str function, used 6 times in handle.c
* remove _alpm_prov_cmp / _alpm_db_whatprovides and replace them by
a more general alpm_find_pkg_satisfiers with a cleaner implementation.
before: alpm_db_whatprovides(db, targ)
after: alpm_find_pkg_satisfiers(alpm_db_getpkgcache(db), targ)
* remove satisfycmp and replace alpm_list_find + satisfycmp usage by
_alpm_find_dep_satisfiers.
before : alpm_list_find(_alpm_db_get_pkgcache(db), dep, satisfycmp)
after : _alpm_find_dep_satisfiers(_alpm_db_get_pkgcache(db), dep)
* remove _alpm_pkgname_pkg_cmp, which was used with alpm_list_remove, and
use _alpm_pkg_find + alpm_list_remove with _alpm_pkg_cmp instead.
This commit actually get rids of all complicated and asymmetric _cmp
functions. I first thought these functions were worth it, be caused it
allowed us to reuse list_find and list_remove. But this was at the detriment
of the clarity and also the ease of use of these functions, dangerous
because of their asymmetricity.
Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
Tested using many easily generated error conditions. Also added "malloc
failure" (conf.c) and "segmentation fault" (pacman.c) error messages for
translation.
Signed-off-by: Allan McRae <mcrae_allan@hotmail.com>
[Dan: fix trailing whitespace errors, other compilation issues]
Signed-off-by: Dan McGee <dan@archlinux.org>
The start of a few commits to remove some PATH_MAX usage from our code. Use
a dynamically allocated string instead.
Signed-off-by: Dan McGee <dan@archlinux.org>
Using the graph structures that Nagy set up for dependency sorting, we now
do a similar process for deltas. Load up all of the deltas into a graph
object on which we can then apply Dijkstra's algorithm, using the new weight
field of graph struct.
We initialize the nodes weight using the base files that we can use in our
filecache (both filename and md5sum must match). The algorithm then picks
the best path among those that can be resolved.
Note that this algorithm has a few advantages over the old one:
1. It is completely file agnostic. These delta chains do not have to consist
of package files- this could be adopted to do delta-fied DBs.
2. It does not use the local_db anymore, or even care if a package or file
is currently installed. Instead, it only looks in the filecache for files
and packages that match delta chain entries.
Original-work-by: Dan McGee <dan@archlinux.org>
Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
The calls to alpm_trans_init and alpm_trans_release (+ error checking) were
duplicated between remove.c, sync.c and upgrade.c
This patch introduces trans_init and trans_release functions in util.c to
have this code just once.
So instead of having to do the same change 3 times for fixing FS#10273, I
just had to do it once (so I did it too :))
Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
Darwin's binary format does support symbols with differing visibilities, but
it does not support the protected or internal visibilities- only hidden. For
Darwin only, we should fall back to this visibility to prevent warnings from
the compiler and because it is close enough for our library purposes.
See http://gcc.gnu.org/viewcvs/*checkout*/trunk/gcc/config/darwin.c, search
for the "darwin_assemble_visibility" function for more details.
Also add pacman.static.exe to gitignore.
Signed-off-by: Dan McGee <dan@archlinux.org>
Linux includes all the gettext stuff in glibc, so there is no need for the
libintl links which we failed to include in our linker variables. Update the
makefiles which should enable NLS support on all platforms, including OS X
and Cygwin.
Signed-off-by: Dan McGee <dan@archlinux.org>
Remove a few functions and things that were unnecessary, update the help
line calls to the current function name, and make the small change to
pacman.c for the signal handler return type that is defined in config.h.
Signed-off-by: Dan McGee <dan@archlinux.org>
We use this function once in our codebase, but fortunately the workaround is
relatively easy. swprintf() is not available on Cygwin so the compile failed
there, but we can do a series of mbstowcs() calls that produce the same end
result as the swprintf() call.
Signed-off-by: Dan McGee <dan@archlinux.org>
If we specify -q/--quiet on an --owns operation, only print a matching
package name rather than the verbose human-readable message.
Signed-off-by: Dan McGee <dan@archlinux.org>
Not too complicated of a fix, but just adds some code to loop over the
entire group list and space it out.
Signed-off-by: Dan McGee <dan@archlinux.org>
This segfault creeped in as a result of commit bf867003. We were incorrectly
assuming the group member of our package was a pmgrp_t list when in fact it
is just a string list, which caused a segfault on any -Qs operation.
Also slightly cleanup the -Ss code (which was originally correct unlike the
-Qs code).
Signed-off-by: Dan McGee <dan@archlinux.org>
As it was already mentioned several times, the new -Sc behavior in 3.1 is
great, but only when the package cache is not shared.
This option has two possible values : KeepInstalled and KeepCurrent
With KeepCurrent, -Sc will clean packages that are no longer available in
any sync db, rather than packages that are no longer in the local db. The
resulting behavior should be better for shared cache.
Ref :
http://www.archlinux.org/pipermail/pacman-dev/2008-February/011140.html
Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
These case insensitive comparisons didn't work in some locales, like tr_TR
where upper(i) != I. So a second case sensitive comparison had to be made
for each directive.
Only keeping case sensitive comparisons make the code cleaner and treat all
locales equally.
Ref: http://www.archlinux.org/pipermail/pacman-dev/2008-March/011445.html
Also fix pactests to use the correct case.
Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
* -Rss removes all dependencies (including explicitly installed ones).
* updated documentation
* two pactest files added to test the difference between -Rs and -Rss
Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu>
From signal man page :
"The behavior of signal() varies across Unix versions, and has also varied
historically across different versions of Linux. Avoid its use: use
sigaction(2) instead. See Portability below."
The code was taken from there :
http://www.gnu.org/software/libtool/manual/libc/Sigaction-Function-Example.html
Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
This function was used in two different ways :
- as a signal handler : the argument was the signal number
- called manually for freeing the resources : the argument was the return
value
So the first part is now handler(int), and the second cleanup(int).
Ref: http://www.archlinux.org/pipermail/pacman-dev/2008-March/011388.html
Remaining problems :
- the return values are messy. for example, 2 can mean both that it was
interrupted (SIGINT == 2), or that --help or -V was used (returned by
parseargs).
- apparently signal is not portable and sigaction should be used instead
Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
Change the pacman_upgrade stub function to do what pacman_add used to do so
we can eliminate pacman_add. Move the code to the more-descriptive name of
upgrade.c.
Note that we have made no changes to the backend libalpm, where an ADD type
transaction could still be supported.
Signed-off-by: Dan McGee <dan@archlinux.org>
Now pacman frontend uses this function instead of the compile-time libalpm
version number.
Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu>
[Dan: fix one more spot where LIB_VERSION was used]
Signed-off-by: Dan McGee <dan@archlinux.org>
(cherry picked from commit 49197b7492)
Now pacman frontend uses this function instead of the compile-time libalpm
version number.
Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu>
[Dan: fix one more spot where LIB_VERSION was used]
Signed-off-by: Dan McGee <dan@archlinux.org>
* mainly code cosmetics (indent fixes)
* remove debug message "spam"
* print also user friendly result
Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu>
[Dan: a few more whitespace/linebreak cleanups added]
Signed-off-by: Dan McGee <dan@archlinux.org>
test_delta_md5sum and test_pkg_md5sum were simple wrappers to test_md5sum,
and only used once, so not very useful. I removed them.
Also, test_md5sum and alpm_pkg_checkmd5sum functions were a bit duplicated,
so I refactored them with a new _alpm_test_md5sum function in libalpm/util.c
Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
I screwed up originally when I accepted the TotalDownload patch,
8ec27835f4. I didn't realize how deeply it
modified libalpm and I probably shouldn't have let it do what it did. This
commit reverts much of what that patch added in order to clean up our
internal function calls. We can find another way to do it right down the
road here but for now it has to go.
Signed-off-by: Dan McGee <dan@archlinux.org>
This change is similar to the one made in
3017b71cb5.
We had a "loading package data..." message, followed by either "failed" or
"done", but it didn't take into account that other warnings / questions
could be displayed between.
Ref: http://archlinux.org/pipermail/pacman-dev/2008-January/010971.html
Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
Add a preset paramater to yesno function saying which answer should be the
default. Ref:
http://www.archlinux.org/pipermail/pacman-dev/2007-June/008470.html
This allows us to answer no by default to some questions, like the -Scc one
mentioned in the above thread, and implemented by this patch.
Another advantage is that we don't have to repeat the [Y/n] in every
questions. It's only put once in yesno function. This highly reduces the
chances that YES and NO strings are translated, but not some questions,
which lead to obvious confusions.
Finally, the noconfirm variable only needs to be used in that yesno
function. So all other usages of it were removed.
Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
