| Age | Commit message (Collapse) | Author |
|---|
|
maintain here.
* moved bash_completion from the "archlinux" script in the official bash_completion package
* moved zsh_completion from the AUR zsh-pacman package
|
|
MSG/ERR usage earlier, the trailing '\n's are no longer needed
* Oddly enough - *ADDED* some '\n's to the sync_info errors (last commit) as the
package info output does not use the pm_fprintf facility and thus does not
honor our newline/no-newline setup
|
|
* Also don't stop searching when one package is not found (output and continue)
|
|
language files with latest pot file.
|
|
* Updated pot translation templates
* Located culprit of progress bar moving when unicode characters are used,
added a TODO note about it
* Removed '(target)' string from the sync.c error message, just like we did
from add.c yesterday
* Updated my TODO
|
|
versions match. (reported by Travis Willard)
|
|
want the actual number of characters, not the number of bytes.
* Added a TODO to take care of later in scriptlet processing.
|
|
|
|
|
|
some needless complication in a few of these functions (checkdeps required
stderr output due to the way it was called/used) so that's all cleaned up now.
Also, makepkg no longer outputs a single dep when they are missing, it lists all
at once, including makedeps
|
|
alpm_pkg_get functions to do this.
|
|
|
|
translations, but may want to update at some point)
* More libalpm/conflict.c changes. Optimized more, hopefully won't need too
much more than this.
|
|
|
|
Douglas Soares de Andrade <douglas@archlinux-br.org>, 2007.
Hugo Doria <hugodoria@archlinux-br.org>, 2007.
Lincoln de Sousa <lincoln@archlinux-br.org>, 2007.
Leandro Inácio <leandro@archlinux-br.org>, 2007.
|
|
Updated missing newlines in German translation.
|
|
* German Translation from:
Pierre Schmitz <pierre@archlinux.de>
Niclas Pfeifer <macwolf@archlinux.de>
* Italian Translation from:
Giovanni 'voidnull' Scafora <linuxmania@gmail.com>
Alessio 'mOLOk' Bolognino <themolok@gmail.com>
Lorenzo '^zanDarK' Masini <lorenxo86@gmail.com>
|
|
|
|
* Wording error in makepkg.conf
|
|
unbelievable amount of strcmp() calls (25 million) due to the list searching.
This has been reimplemented with a set-intersection scheme, due to the fact
that file lists are always ordered. - NEEDS TESTING
* Minor clean up, "globalized" the str_cmp helper to match the alpm comparison
signature, so we can use it elsewhere.
|
|
--cachedir.
|
|
descriptions, nor do we actually build them with the newest makepkg. The actual
implementation in here seems like it could be done better, and it soaks a large
amount of CPU time according to profilers.
|
|
pacman didn't check targets and spit out an error message).
|
|
|
|
and not always MARCH). Also changes makepkg.conf.in to not have commented -mcpu
lines (due to the above), but to have a comment about the differences in -march
and -mcpu
|
|
|
|
|
|
now uses an strlen call.
* Updates of the *.pot files.
|
|
* converted "sudosync" to "usesudo" to match makepkg
* Removed nostrip, as it was removed from makepkg as well
|
|
- support for the 'BCemno' makepkg options
|
|
|
|
|
|
* Removed usage of 'realpath'. From the manpage:
"Avoid using this function. It is broken by design"
|
|
used instead
* Allow -T/--deptest to work without root privileges
|
|
issues - silly me.
Also removed config.rpath from CVS
|
|
|
|
|
|
|
|
* Removed the handle->needles param. It's not needed not that alpm_list_t is
public
|
|
testpkg <package file>, returns 0 on success, 1 on failure
failure will also output error messages to stdout
|
|
the array, e.g. arch=('i686').
* Updated my TODO.
* Changed the email address in rankmirrors script.
|
|
|
|
'--usesudo', and all install/dep/remove operations now honor it.
* Fixed my stupid typo. depricated -> deprecated :)
|
|
|
|
More or less a reverse 'repo-add'.
|
|
|
|
|
|
from the PKGINFO, and size is the stat() size of the archive
* Removed the useless 'output' param from package.c:parse_descfile
* Installation progress
- Call progress callback once at 0% for initialization
- 'needdisp' was useless
- alpm_list_count is called an excessive amount in these nested loops. Now we
only call it once per iteration
- Use the compressed sizes for PROGRESS calcs as uncompressed (isize) is not
exact (it is missing metadata sizes), and thus produces > 100% numbers
|
|
* downloaded packages now pass ONLY the package name to the callback
|
|
* Questions no longer start with "error:"
* downloaded size is output as a float now
|
