- Pacman-dev - lists.archlinux.org

[pacman-dev] 3.4.1 release plans
by Dan McGee 03 Aug '10

03 Aug '10

Hey guys, We've had a handful of fixes go into maint so it might be time to think about a 3.4.1 release. Some other pending issues to think about before we release: * http://bugs.archlinux.org/task/19969, pacman download issues from googlecode.com domain. Looks like not our problem so will probably close this out. * http://bugs.archlinux.org/task/20022, Download progress and completion issues during repo sync. I saw this before the release in git, and it seems like someone else hit it too. It is always involving kernel.org HTTP that we've seen so far. * http://bugs.archlinux.org/task/19854, --ignore is ignored. This is the highest priority issue and something I definitely want to get in 3.4.1; I have started on a fix and it shouldn't be too bad but some code and checks need to move around. * http://bugs.archlinux.org/task/19853, No prompting about groups. Might even get fixed with the above, and with 24 votes, apparently people do care (but not enough to answer the patches welcome call). Thoughts/comments/concerns welcome as always. -Dan

2 3

[pacman-dev] [PATCH] makepkg: check tput support before using
by Allan McRae 03 Aug '10

03 Aug '10

Prevent makepkg aborting whe colors are enabled and the terminal does not support setting colors by tput. Signed-off-by: Allan McRae <allan(a)archlinux.org> --- scripts/makepkg.sh.in | 22 ++++++++++++++++------ 1 files changed, 16 insertions(+), 6 deletions(-) diff --git a/scripts/makepkg.sh.in b/scripts/makepkg.sh.in index 516e1d7..0527bd6 100644 --- a/scripts/makepkg.sh.in +++ b/scripts/makepkg.sh.in @@ -1629,12 +1629,22 @@ PACMAN=${PACMAN:-pacman} # check if messages are to be printed using color unset ALL_OFF BOLD BLUE GREEN RED YELLOW if [[ -t 2 && ! $USE_COLOR = "n" && $(check_buildenv color) = "y" ]]; then - ALL_OFF="$(tput sgr0)" - BOLD="$(tput bold)" - BLUE="${BOLD}$(tput setaf 4)" - GREEN="${BOLD}$(tput setaf 2)" - RED="${BOLD}$(tput setaf 1)" - YELLOW="${BOLD}$(tput setaf 3)" + # prefer terminal safe colored and bold text when tput is supported + if tput setaf 0 &>/dev/null; then + ALL_OFF="$(tput sgr0)" + BOLD="$(tput bold)" + BLUE="${BOLD}$(tput setaf 4)" + GREEN="${BOLD}$(tput setaf 2)" + RED="${BOLD}$(tput setaf 1)" + YELLOW="${BOLD}$(tput setaf 3)" + else + ALL_OFF="\033[1;0m" + BOLD="\033[1;1m" + BLUE="${BOLD}\033[1;34m" + GREEN="${BOLD}\033[1;32m" + RED="${BOLD}\033[1;31m" + YELLOW="${BOLD}\033[1;33m" + fi fi readonly ALL_OFF BOLD BLUE GREEN RED YELLOW -- 1.7.2.1

1 0

[pacman-dev] slovak translation for pacman
by Jozef Riha 27 Jul '10

27 Jul '10

attached. cheers, jose

2 3

[pacman-dev] [PATCH] Use the plural features of gettext
by Dan McGee 27 Jul '10

27 Jul '10

Gettext has this whole 'Plural-Form' thing that until now we haven't taken advantage of. Given that not all languages have the same plural form rules as English, take advantage of it by defining a new _n() macro which will normally define to ngettext(), and adjust a few messages as an example of how to use. There are surely other places where we do singular/plural logic without me having noticed, so further patches are welcome to fix those up too. Signed-off-by: Dan McGee <dan(a)archlinux.org> --- src/pacman/callback.c | 13 +++++++++---- src/pacman/query.c | 6 ++++-- src/pacman/util.h | 2 ++ 3 files changed, 15 insertions(+), 6 deletions(-) diff --git a/src/pacman/callback.c b/src/pacman/callback.c index 23804d7..3696d3c 100644 --- a/src/pacman/callback.c +++ b/src/pacman/callback.c @@ -283,11 +283,16 @@ void cb_trans_conv(pmtransconv_t event, void *data1, void *data2, namelist = alpm_list_add(namelist, (char *)alpm_pkg_get_name(i->data)); } - printf(_(":: the following package(s) cannot be upgraded due to " - "unresolvable dependencies:\n")); + printf(_n( + ":: The following package cannot be upgraded due to unresolvable dependencies:\n", + ":: The following packages cannot be upgraded due to unresolvable dependencies:\n", + alpm_list_count(namelist))); list_display(" ", namelist); - *response = noyes(_("\nDo you want to skip the above " - "package(s) for this upgrade?")); + printf("\n"); + *response = noyes(_n( + "Do you want to skip the above package for this upgrade?", + "Do you want to skip the above packages for this upgrade?", + alpm_list_count(namelist))); alpm_list_free(namelist); } break; diff --git a/src/pacman/query.c b/src/pacman/query.c index 7f064f2..5538e81 100644 --- a/src/pacman/query.c +++ b/src/pacman/query.c @@ -406,8 +406,10 @@ static int check(pmpkg_t *pkg) } if(!config->quiet) { - printf(_("%s: %d total files, %d missing file(s)\n"), - pkgname, allfiles, errors); + printf(_n("%s: %d total file, ", "%s: %d total files, ", allfiles), + pkgname, allfiles); + printf(_n("%d missing file\n", "%d missing files\n", errors), + errors); } return(errors != 0 ? 1 : 0); diff --git a/src/pacman/util.h b/src/pacman/util.h index bbf4382..d85ba7b 100644 --- a/src/pacman/util.h +++ b/src/pacman/util.h @@ -30,8 +30,10 @@ #include <libintl.h> /* here so it doesn't need to be included elsewhere */ /* define _() as shortcut for gettext() */ #define _(str) gettext(str) +#define _n(str1, str2, ct) ngettext(str1, str2, ct) #else #define _(str) str +#define _n(str1, str2, ct) (ct == 1 ? str1 : str2) #endif /* update speed for the fill_progress based functions */ -- 1.7.2

1 0

[pacman-dev] [GIT] The official pacman repository branch, maint, updated. v3.4.0-24-g0d6efb3
by dan＠archlinux.org 27 Jul '10

27 Jul '10

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "The official pacman repository". The branch, maint has been updated via 0d6efb35ce445747cdc37b4f58cdf63463de6e1a (commit) via 52118bf0f084e891b53bcf05f9e6021ebb26fb13 (commit) via ff689b6a383cbeb5ea699d587fe8975d71cae600 (commit) via fa4f25626c462e88ca0f6457bf9942980a5bd329 (commit) via f8d7cd6b2623a864aa85fbdcdd629f3ff92a631c (commit) via e702f56ea671c6cd1154a0ddb41fa63e97587c85 (commit) via 7f5c486666cc99fba4029b2c059ae062b7dd7933 (commit) from fcb4f0264f2b8e0a6ed1e7eebfe00f662ba94ef2 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- ----------------------------------------------------------------------- Summary of changes: contrib/bash_completion | 4 ++-- lib/libalpm/po/kk.po | 2 +- lib/libalpm/sync.c | 18 +++++++++++------- scripts/makepkg.sh.in | 6 +++--- src/pacman/util.c | 1 + test/pacman/tests/sync022.py | 25 +++++++++++++++++++++++++ test/pacman/tests/sync023.py | 29 +++++++++++++++++++++++++++++ 7 files changed, 72 insertions(+), 13 deletions(-) create mode 100644 test/pacman/tests/sync022.py create mode 100644 test/pacman/tests/sync023.py hooks/post-receive -- The official pacman repository

1 0

[pacman-dev] [GIT] The official pacman repository branch, master, updated. v3.4.0-56-gddc4130
by dan＠archlinux.org 27 Jul '10

27 Jul '10

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "The official pacman repository". The branch, master has been updated via ddc4130c978decc5b2498b24d3b20e0c4efbc3cf (commit) via 0d6efb35ce445747cdc37b4f58cdf63463de6e1a (commit) via 52118bf0f084e891b53bcf05f9e6021ebb26fb13 (commit) via ff689b6a383cbeb5ea699d587fe8975d71cae600 (commit) via fa4f25626c462e88ca0f6457bf9942980a5bd329 (commit) via f8d7cd6b2623a864aa85fbdcdd629f3ff92a631c (commit) via e702f56ea671c6cd1154a0ddb41fa63e97587c85 (commit) via 7f5c486666cc99fba4029b2c059ae062b7dd7933 (commit) from a83559949622992faafb628b742893891ead8390 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit ddc4130c978decc5b2498b24d3b20e0c4efbc3cf Merge: a835599 0d6efb3 Author: Dan McGee <dan(a)archlinux.org> Date: Tue Jul 27 10:18:35 2010 -0500 Merge branch 'maint' ----------------------------------------------------------------------- Summary of changes: contrib/bash_completion | 4 ++-- lib/libalpm/po/kk.po | 2 +- lib/libalpm/sync.c | 18 +++++++++++------- scripts/makepkg.sh.in | 6 +++--- src/pacman/util.c | 1 + test/pacman/tests/sync022.py | 25 +++++++++++++++++++++++++ test/pacman/tests/sync023.py | 29 +++++++++++++++++++++++++++++ 7 files changed, 72 insertions(+), 13 deletions(-) create mode 100644 test/pacman/tests/sync022.py create mode 100644 test/pacman/tests/sync023.py hooks/post-receive -- The official pacman repository

1 0

[pacman-dev] [PATCH 4/4] Add INTERRUPTED to transaction hooks support.
by Victor Lowther 27 Jul '10

27 Jul '10

If a transaction is interrupted, the hooks need to know so that they can roll back anything that needs to be rolled back. --- lib/libalpm/trans.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/lib/libalpm/trans.c b/lib/libalpm/trans.c index ed26aa6..d212f50 100644 --- a/lib/libalpm/trans.c +++ b/lib/libalpm/trans.c @@ -235,6 +235,7 @@ int SYMEXPORT alpm_trans_interrupt() RET_ERR(PM_ERR_TRANS_TYPE, -1)); trans->state = STATE_INTERRUPTED; + alpm_hook_transact("INTERRUPTED"); return(0); } -- 1.7.1.1

1 0

[pacman-dev] [PATCH 3/4] Add a PKGBUILD for easy testing.
by Victor Lowther 27 Jul '10

27 Jul '10

We don't need the whole usual pkgbuild, but having one in the git tree sure makes testing simpler. --- PKGBUILD | 60 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 files changed, 60 insertions(+), 0 deletions(-) diff --git a/PKGBUILD b/PKGBUILD new file mode 100644 index 0000000..16cf977 --- /dev/null +++ b/PKGBUILD @@ -0,0 +1,60 @@ +# $Id: PKGBUILD 83376 2010-06-21 12:52:09Z dan $ +# Maintainer: Aaron Griffin <aaron(a)archlinux.org> +# Maintainer: Dan McGee <dan(a)archlinux.org> + +pkgname=pacman-git +pkgver=$(date +%s) +pkgrel=$(git log --pretty=format:%h |head -n 1) +pkgdesc="A library-based package manager with dependency support" +arch=('i686' 'x86_64') +url="http://www.archlinux.org/pacman/" +license=('GPL') +groups=('base') +depends=('bash' 'libarchive>=2.7.1' 'libfetch>=2.25' 'pacman-mirrorlist') +makedepends=('asciidoc') +optdepends=('fakeroot: for makepkg usage as normal user') +provides=('pacman=9999') +backup=('etc/pacman.conf' 'etc/makepkg.conf') +conflicts=('pacman') +options=(!libtool) + +build() { + cd .. + ./autogen.sh + ./configure --prefix=/usr --sysconfdir=/etc --localstatedir=/var --enable-doc + make || return 1 +} + +package() { + cd $srcdir/.. + srcdir="$PWD" + make DESTDIR=$pkgdir install || return 1 + + # install Arch specific stuff + mkdir -p $pkgdir/etc + # set things correctly in the default conf file + case "$CARCH" in + i686) + mycarch="i686" + mychost="i686-pc-linux-gnu" + myflags="-march=i686 " + ;; + x86_64) + mycarch="x86_64" + mychost="x86_64-unknown-linux-gnu" + myflags="-march=x86-64 " + ;; + esac + sed -i $pkgdir/etc/makepkg.conf \ + -e "s|@CARCH[@]|$mycarch|g" \ + -e "s|@CHOST[@]|$mychost|g" \ + -e "s|@CARCHFLAGS[@]|$myflags|g" + + # install completion files + mkdir -p $pkgdir/etc/bash_completion.d/ + install -m644 contrib/bash_completion $pkgdir/etc/bash_completion.d/pacman + mkdir -p $pkgdir/usr/share/zsh/site-functions/ + install -m644 contrib/zsh_completion $pkgdir/usr/share/zsh/site-functions/_pacman +} + +# vim: set ts=2 sw=2 et: -- 1.7.1.1

1 0

[pacman-dev] [PATCH 2/4] Add very basic transaction hooks to alpm.
by Victor Lowther 27 Jul '10

27 Jul '10

libalpm compliant programs now support running pre-commit and post-commit hooks. Currently, the implementation is as simple as I can make it -- if /etc/pacman.d/hooks.d/transact.sh exists, it will be run with the 'pre-commit' parameter just before a transaction is committed, and it will be run with 'post-commit' after the transaction has been completed. The intent is for transact.sh to be a wrapper that calls scripts in /etc/pacman.d/hooks.d/transact.d with the arg it is passed, but since this is just for a demo I did not actually include any of that. I also didn't bother adhering to any of the coding standards, as that will be easy enough to fix up once we have something a bit more fleshed out. --- lib/libalpm/hooks.c | 28 ++++++++++++++++++++++++++++ lib/libalpm/hooks.h | 4 ++++ lib/libalpm/trans.c | 4 +++- 3 files changed, 35 insertions(+), 1 deletions(-) diff --git a/lib/libalpm/hooks.c b/lib/libalpm/hooks.c index 9630c70..5982ea5 100644 --- a/lib/libalpm/hooks.c +++ b/lib/libalpm/hooks.c @@ -30,6 +30,7 @@ #include <sys/types.h> #include <sys/stat.h> #include <sys/statvfs.h> +#include <sys/wait.h> #include <errno.h> /* libalpm */ @@ -46,3 +47,30 @@ #include "alpm.h" #include "deps.h" #include "cache.h" + +int alpm_hook_transact(char* phase) { + char filename[PATH_MAX]; + struct stat sbuf; + pid_t childpid, retpid; + int status; + /* ignore transaction hooks when targeting another root filesystem + if (!strcmp(handle->root,"/")) + return 1;*/ + snprintf(filename,PATH_MAX,"%s/transact.sh",ALPM_HOOKS_DIR); + /* if we cannot stat transact.sh, just return */ + if (stat(filename,&sbuf) != 0) + return 1; + fflush(NULL); + childpid=fork(); + if (childpid == -1) { + return 1; + } else if (childpid == 0) { + execl(filename,"transact.sh",phase,(char *)0); + } else { + /* return the exit status of the transact.sh program. */ + do { + retpid = waitpid(childpid,&status,0); + } while (!WIFEXITED(status) && !WIFSIGNALED(status)); + } + return !WEXITSTATUS(status); +} diff --git a/lib/libalpm/hooks.h b/lib/libalpm/hooks.h index d60b1f4..f79d2c4 100644 --- a/lib/libalpm/hooks.h +++ b/lib/libalpm/hooks.h @@ -25,4 +25,8 @@ #include "alpm.h" +#define ALPM_HOOKS_DIR "/etc/pacman.d/hooks.d" + +int alpm_hook_transact(char*); + #endif /*ALPM_HOOKS_H */ diff --git a/lib/libalpm/trans.c b/lib/libalpm/trans.c index 523f33b..ed26aa6 100644 --- a/lib/libalpm/trans.c +++ b/lib/libalpm/trans.c @@ -195,7 +195,8 @@ int SYMEXPORT alpm_trans_commit(alpm_list_t **data) if(trans->add == NULL && trans->remove == NULL) { return(0); } - + if (! alpm_hook_transact("pre-commit")) + return (-1); trans->state = STATE_COMMITING; if(trans->add == NULL) { @@ -211,6 +212,7 @@ int SYMEXPORT alpm_trans_commit(alpm_list_t **data) } trans->state = STATE_COMMITED; + alpm_hook_transact("post-commit"); return(0); } -- 1.7.1.1

1 0

[pacman-dev] [PATCH 1/4] Add skeleton for hooks support
by Victor Lowther 27 Jul '10

27 Jul '10

This begins work on adding some sort of hook infrastructure to pacman. There has been much discussion and little code that I have seen, so I figured I would go ahead an implement a very basic skeleton for transaction hooks support. --- lib/libalpm/Makefile.am | 1 + lib/libalpm/{sync.h => hooks.c} | 34 +++++++++++++++++++++++++--------- lib/libalpm/{add.h => hooks.h} | 19 ++++++++----------- 3 files changed, 34 insertions(+), 20 deletions(-) diff --git a/lib/libalpm/Makefile.am b/lib/libalpm/Makefile.am index 3473a73..b4fdadf 100644 --- a/lib/libalpm/Makefile.am +++ b/lib/libalpm/Makefile.am @@ -44,6 +44,7 @@ libalpm_la_SOURCES = \ sync.h sync.c \ trans.h trans.c \ util.h util.c \ + hooks.h hooks.c \ version.c libalpm_la_LDFLAGS = -no-undefined -version-info $(LIB_VERSION_INFO) diff --git a/lib/libalpm/sync.h b/lib/libalpm/hooks.c similarity index 65% copy from lib/libalpm/sync.h copy to lib/libalpm/hooks.c index 000a09c..9630c70 100644 --- a/lib/libalpm/sync.h +++ b/lib/libalpm/hooks.c @@ -1,9 +1,10 @@ /* - * sync.h + * hooks.c * * Copyright (c) 2006-2010 Pacman Development Team <pacman-dev(a)archlinux.org> * Copyright (c) 2002-2006 by Judd Vinet <jvinet(a)zeroflux.org> * Copyright (c) 2005 by Aurelien Foret <orelien(a)chez.com> + * Copyright (c) 2005 by Christian Hamar <krics(a)linuxforum.hu> * Copyright (c) 2005, 2006 by Miklos Vajna <vmiklos(a)frugalware.org> * * This program is free software; you can redistribute it and/or modify @@ -19,14 +20,29 @@ * You should have received a copy of the GNU General Public License * along with this program. If not, see <http://www.gnu.org/licenses/>. */ -#ifndef _ALPM_SYNC_H -#define _ALPM_SYNC_H -#include "alpm.h" - -int _alpm_sync_prepare(pmtrans_t *trans, pmdb_t *db_local, alpm_list_t *dbs_sync, alpm_list_t **data); -int _alpm_sync_commit(pmtrans_t *trans, pmdb_t *db_local, alpm_list_t **data); +#include "config.h" -#endif /* _ALPM_SYNC_H */ +#include <stdlib.h> +#include <stdio.h> +#include <string.h> +#include <unistd.h> +#include <sys/types.h> +#include <sys/stat.h> +#include <sys/statvfs.h> +#include <errno.h> -/* vim: set ts=2 sw=2 noet: */ +/* libalpm */ +#include "hooks.h" +#include "trans.h" +#include "alpm_list.h" +#include "package.h" +#include "util.h" +#include "log.h" +#include "handle.h" +#include "add.h" +#include "remove.h" +#include "sync.h" +#include "alpm.h" +#include "deps.h" +#include "cache.h" diff --git a/lib/libalpm/add.h b/lib/libalpm/hooks.h similarity index 73% copy from lib/libalpm/add.h copy to lib/libalpm/hooks.h index eb37dc7..d60b1f4 100644 --- a/lib/libalpm/add.h +++ b/lib/libalpm/hooks.h @@ -1,8 +1,11 @@ /* - * add.h + * hooks.h * * Copyright (c) 2006-2010 Pacman Development Team <pacman-dev(a)archlinux.org> * Copyright (c) 2002-2006 by Judd Vinet <jvinet(a)zeroflux.org> + * Copyright (c) 2005 by Aurelien Foret <orelien(a)chez.com> + * Copyright (c) 2005 by Christian Hamar <krics(a)linuxforum.hu> + * Copyright (c) 2006 by Miklos Vajna <vmiklos(a)frugalware.org> * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by @@ -17,15 +20,9 @@ * You should have received a copy of the GNU General Public License * along with this program. If not, see <http://www.gnu.org/licenses/>. */ -#ifndef _ALPM_ADD_H -#define _ALPM_ADD_H +#ifndef _ALPM_HOOKS_H +#define _ALPM_HOOKS_H -#include "db.h" -#include "alpm_list.h" -#include "trans.h" +#include "alpm.h" -int _alpm_upgrade_packages(pmtrans_t *trans, pmdb_t *db); - -#endif /* _ALPM_ADD_H */ - -/* vim: set ts=2 sw=2 noet: */ +#endif /*ALPM_HOOKS_H */ -- 1.7.1.1

1 0

Re: [pacman-dev] New patchset for package signing
by Denis A. Altoé Falqueto 26 Jul '10

26 Jul '10

On Thu, Jul 22, 2010 at 12:17 AM, Denis A. Altoé Falqueto <denisfalqueto(a)gmail.com> wrote: > Guys, > > The previous patches were not much commented. I believe the main problem was > the fact that I didn't use gpgme (and, instead, undone some work already done). > > So, I changed the code to use gpgme. Here are the patches for your evaluation. Hi, Now that ArchCon has ended, I believe we can continue with the discussion, right? The replies are rare, so I don't know if there is not enough interest or if my implementation is so bad that can't be fixed :) I am open to any constructive criticism and suggestions, as always. -- A: Because it obfuscates the reading. Q: Why is top posting so bad? ------------------------------------------- Denis A. Altoe Falqueto -------------------------------------------

2 2

[pacman-dev] [PATCH] bash_completion: negate expression inside brackets
by Dave Reisner 26 Jul '10

26 Jul '10

Avoids letting the shell evaluate ! as something else (e.g. an alias). Signed-off-by: Dave Reisner <d(a)falconindy.com> --- contrib/bash_completion | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/contrib/bash_completion b/contrib/bash_completion index 1ec2cd5..bdc4754 100644 --- a/contrib/bash_completion +++ b/contrib/bash_completion @@ -75,8 +75,8 @@ _pacman() { if [[ $? != 0 ]]; then _arch_ptr2comp core - elif ! [[ $prev =~ ^-\w*[Vbhr] || - $prev = --@(cachedir|config|dbpath|help|logfile|root|version) ]] + elif [[ ! $prev =~ ^-\w*[Vbhr] && + ! $prev = --@(cachedir|config|dbpath|help|logfile|root|version) ]] then [[ $cur = -* ]] && _arch_ptr2comp ${o#* } common || case ${o% *} in -- 1.7.2

1 0

[pacman-dev] [PATCH] bash_completion: negate expression inside brackets
by Dave Reisner 26 Jul '10

26 Jul '10

Avoids letting the shell evaluate the ! as something else (e.g. if ! is an alias). Signed-off-by: Dave Reisner <d(a)falconindy.com> --- contrib/bash_completion | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/contrib/bash_completion b/contrib/bash_completion index 1ec2cd5..486870f 100644 --- a/contrib/bash_completion +++ b/contrib/bash_completion @@ -75,7 +75,7 @@ _pacman() { if [[ $? != 0 ]]; then _arch_ptr2comp core - elif ! [[ $prev =~ ^-\w*[Vbhr] || + elif [[ ! $prev =~ ^-\w*[Vbhr] || $prev = --@(cachedir|config|dbpath|help|logfile|root|version) ]] then [[ $cur = -* ]] && _arch_ptr2comp ${o#* } common || -- 1.7.2

2 3

[pacman-dev] Unnecessary syscalls
by Tim Nieradzik 26 Jul '10

26 Jul '10

Hello, it's my first message on this list. I'd like to make a few contributions to improve Pacman's performance. Pacman is a great package manager which I'm now using for a few years without any problems. It also has a very clear and well-structured code but its disk stress concerns me a bit. Looking at strace, a few syscalls appear superfluous to me. Deleting them made pacman -Syu (right after a disk flush) about 10s faster (70% less syscalls). Here's the patch: http://pastebin.archlinux.fr/406557 I've already talked about it on the forum but they advised me to discuss it here instead. There are also some suggestions for further optimizations: http://bbs.archlinux.org/viewtopic.php?id=101702 What do you think? --Tim

2 1

[pacman-dev] [PATCH 1/2] Add two pactests for group and --needed interaction
by Dan McGee 26 Jul '10

26 Jul '10

The first step for resolving FS#20221. sync023 is the case from the bug report; sync022 is already working fine but we have no tests at all that test the --needed option in any form. Signed-off-by: Dan McGee <dan(a)archlinux.org> --- test/pacman/tests/sync022.py | 25 +++++++++++++++++++++++++ test/pacman/tests/sync023.py | 31 +++++++++++++++++++++++++++++++ 2 files changed, 56 insertions(+), 0 deletions(-) create mode 100644 test/pacman/tests/sync022.py create mode 100644 test/pacman/tests/sync023.py diff --git a/test/pacman/tests/sync022.py b/test/pacman/tests/sync022.py new file mode 100644 index 0000000..eebbe07 --- /dev/null +++ b/test/pacman/tests/sync022.py @@ -0,0 +1,25 @@ +self.description = "Install a group from a sync db using --needed" + +lp1 = pmpkg("pkg1") +lp2 = pmpkg("pkg2") +lp3 = pmpkg("pkg3") + +sp1 = pmpkg("pkg1", "1.1-1") +sp2 = pmpkg("pkg2") +sp3 = pmpkg("pkg3") + +for p in lp1, lp2, lp3, sp1, sp2, sp3: + setattr(p, "groups", ["grp"]) + +for p in lp1, lp2, lp3: + self.addpkg2db("local", p) + +for p in sp1, sp2, sp3: + self.addpkg2db("sync", p); + +self.args = "-S --needed grp" + +self.addrule("PACMAN_RETCODE=0") +for p in sp1, sp2, sp3: + self.addrule("PKG_EXIST=%s" % p.name) +self.addrule("PKG_VERSION=pkg1|1.1-1") diff --git a/test/pacman/tests/sync023.py b/test/pacman/tests/sync023.py new file mode 100644 index 0000000..9253497 --- /dev/null +++ b/test/pacman/tests/sync023.py @@ -0,0 +1,31 @@ +self.description = "Install a group from a sync db using --needed (testing repo)" + +lp1 = pmpkg("pkg1", "1.1-1") +lp2 = pmpkg("pkg2") +lp3 = pmpkg("pkg3") + +sp1 = pmpkg("pkg1") +sp2 = pmpkg("pkg2") +sp3 = pmpkg("pkg3") +newp1 = pmpkg("pkg1", "1.1-1") + +for p in lp1, lp2, lp3, sp1, sp2, sp3, newp1: + setattr(p, "groups", ["grp"]) + +for p in lp1, lp2, lp3: + self.addpkg2db("local", p) + +self.addpkg2db("testing", newp1); + +for p in sp1, sp2, sp3: + self.addpkg2db("sync", p); + +self.args = "-S --needed grp" + +self.addrule("PACMAN_RETCODE=0") +for p in sp1, sp2, sp3: + self.addrule("PKG_EXIST=%s" % p.name) +# The newer version should still be installed +self.addrule("PKG_VERSION=pkg1|1.1-1") + +self.expectfailure = True -- 1.7.2

1 1

[pacman-dev] [PATCH] Support reading package args from stdin
by Dave Reisner 25 Jul '10

25 Jul '10

Only occurs if no arguments were provided directly. Arguments can be separated by any amount of valid whitespace. This allows for piping into pacman from other programs or from itself, e.g.: pacman -Qdtq | pacman -Rs Signed-off-by: Dave Reisner <d(a)falconindy.com> --- src/pacman/pacman.c | 24 ++++++++++++++++++++++++ 1 files changed, 24 insertions(+), 0 deletions(-) diff --git a/src/pacman/pacman.c b/src/pacman/pacman.c index 049bc40..6d9d954 100644 --- a/src/pacman/pacman.c +++ b/src/pacman/pacman.c @@ -26,6 +26,7 @@ #define PACKAGE_VERSION GIT_VERSION #endif +#include <ctype.h> /* isspace */ #include <stdlib.h> /* atoi */ #include <stdio.h> #include <limits.h> @@ -1172,6 +1173,29 @@ int main(int argc, char *argv[]) cleanup(ret); } + /* read package arguments from stdin if we have none yet */ + if(!pm_targets && !isatty(0)) { + char line[BUFSIZ]; + int i = 0; + while((line[i] = fgetc(stdin)) != EOF) { + if(isspace(line[i])) { + line[i] = 0; + /* avoid adding zero length arg, if multiple spaces separate args */ + if(i > 0) { + pm_targets = alpm_list_add(pm_targets, strdup(line)); + i = 0; + } + continue; + } + ++i; + } + /* end of stream -- check for data still in line buffer */ + if(i > 0) { + pm_targets = alpm_list_add(pm_targets, strdup(line)); + } + freopen(ctermid(NULL), "r", stdin); + } + /* parse the config file */ ret = parseconfig(config->configfile); if(ret != 0) { -- 1.7.2

4 4

Re: [pacman-dev] Pacman.conf
by Dan McGee 21 Jul '10

21 Jul '10

if you think they aren't clear, by all means please file a bug report. And when you mention you volunteered to help, I guess I'm not sure who this is considering the email is from Tech Support Department, I'm not even quite sure who I'm talking to. -Dan On Wednesday, July 21, 2010, Tech Support Department <tech(a)frontrowcomputer.com> wrote: > > > > > > > Okay. It was just that since I had already offered > to help on the various man pages and was not taken up on the > offer, there was no way I could help fix the poor writing in the > man pages, and by extension, get the information I needed myself > unless I went directly to the source. I guess we users will have > to start a new document somewhere based upon trial and error. > > Thanks anyway. > > > On 07/21/2010 10:11 AM, Dan McGee wrote: > > On Wed, Jul 21, 2010 at 9:07 AM, Tech Support Department > <tech(a)frontrowcomputer.com> <tech(a)frontrowcomputer.com> wrote: > > > Could you please explain the following statements, which I found at > http://www.archlinux.org/pacman/pacman.conf.5.html: > > During parsing, pacman will define the $repo variable to the name of the > current section. This is often utilized in files specified using the Include > directive so all repositories can use the same mirrorfile. pacman also > defines the $arch variable to the value of Architecture, so the same > mirrorfile can even be used for different architectures. > > > > The Arch Linux forums, IRC channel, or mailing lists are a much better > venue for such questions. > > If you would still like my private assistance, I would be glad to > assist once we discuss payment options. My billing rate is $65/hour. > > -Dan > > > > > >

2 1

[pacman-dev] Explicit (opt)deps for makepkg.
by Xyne 19 Jul '10

19 Jul '10

Hi, I think the pacman package should explicitly list makepkg's optional dependencies, instead of simply assuming the presence of base and base-devel. Using a minimal 32-bit chroot, the only way to keep it minimal is to try to run makepkg and then track down each dependency-related error one by one until it completes. There is no reason to install a huge group when only a few of those packages are actually required (e.g. "file" from "base") and doing so simply leads to unnecessary downloads and chroot pollution. The shotgun approach to package installation for makepkg has always struck me as unminimalistic. I understand that the omission of explicit dependencies is seen as "simple" by some, but that's probably how the huge DE devs view their dep chains too. Maybe "base-devel" could encompass all hard dependencies of makepkg and nothing more, and then be listed as an optdep for makepkg. Packages that require anything beyond that should list the additional (build) deps explicitly. Thoughts? Regards, Xyne

5 6

[pacman-dev] pacman po files (translations)
by Dmitry Korzhevin 17 Jul '10

17 Jul '10

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Please tell me - where I can get pacman po files, in order to update the Russian and Ukrainian translation? - -- Best regards, Dmitry Korzhevin Tel: +38 (039) 295-0000 Office Phone: +38 (044) 383-14-12 E-mail: dkorzhevin(a)lsupport.net Jabber ID: dkorzhevin(a)lsupport.net Skype: dkorzhevin URL: http://lsupport.net Linux Support LLC -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQEcBAEBAgAGBQJMQQaeAAoJEGuRMpfwJUbRowgIAK2QXv/5uXJpqCqmWIntWaSG J/H2pb3OWH3aItkG2wEn5F4sfBsFpMFmxtVLbnAnAqoLPmbaUmwV/RV/ZoFBtFll 4mCyjANnMMPtk6Rlc7ff0HsTNe4ovnRRV9xWFRIM/sfhO3KYYmTNB+ESa59pcLvY nJtnC5i7lJZezc9bxB6GhoYk+hQ7MOOQY+t0YTEN9PyEJb5sI/l5r5h9VvMr5Lf3 XMIqGnqwWU+giRpMH7iQey6W6AvCL6P6h8ENI4i6aQemrQ3sLecQ72g4hS5JzJtL 150Wv9RhitYW0xZssd4H74Qrarpig9s/G2arsi7vbexuoNf/3TtwttqNAtoDC3A= =u4+o -----END PGP SIGNATURE-----

2 1

[pacman-dev] pacakge signing (I know again?)
by Chris Bannister 09 Jul '10

09 Jul '10

Hey guys. So before you ignore this as another "Where is package singing?" thread, I would like to help. Im a CS student just finished my first year with a decent understanding of C with a background of educated C++ (Uni forced me to learn this). I want to at least help get gpg signing into a usable testable state, but looking through the git project I cant really find where to start and what needs doing to get up to a working state. Sorry if this comes up a lot on the mailing list a lot, I just subscribed to it, been on the general/dev list for a while. Chris Bannister

3 2

[pacman-dev] makepkg aborts when using unknown terminal
by Eric Bélanger 09 Jul '10

09 Jul '10

Hi, This problem is somewhat related to the terminal color problem so I apologize if it's already taken care by these patches. I use rxvt-unicode as my terminal. Since pacman 3.4, when building packages with makechrootpkg, makepkg aborts with the following error: tput: unknown terminal "rxvt-unicode" This happens because I use a clean chroot in which rxvt-unicode isn't installed so the terminal definition file /usr/share/terminfo/r/rxvt-unicode isn't present. As a workaround, I added a 'TERM=xterm' in the wrapper script I use. It would be nice if, instead of aborting on unknown terminals, makepkg would fallback to the dumb terminal. The dumb terminal currently doesn's work as I believe it doesn't support colors. This change along with the color patch would make makpkg more robust when handling different and unusual terminals. Personnally, I will keep my workaround as I like color. ;) Eric

3 5

[pacman-dev] [PATCH 2/6] Remove use of gpgme
by Denis A. Altoé Falqueto 08 Jul '10

08 Jul '10

gpgme is very good when the keyring is the default, or if the requirements are simple. But for a specialized keyring, like pacman's, it is not flexible enough. This patch removes the checking of configure.ac. Signed-off-by: Denis A. Altoé Falqueto <denisfalqueto(a)gmail.com> --- configure.ac | 4 ---- lib/libalpm/alpm.h | 1 - lib/libalpm/error.c | 2 -- 3 files changed, 0 insertions(+), 7 deletions(-) diff --git a/configure.ac b/configure.ac index b6ce68b..008975b 100644 --- a/configure.ac +++ b/configure.ac @@ -154,10 +154,6 @@ else fi AM_CONDITIONAL(INTERNAL_DOWNLOAD, test "x$internaldownload" = "xyes") -# Check for gpgme -AC_CHECK_LIB([gpgme], [gpgme_check_version], , - AC_MSG_ERROR([gpgme is needed to compile pacman!])) - # Checks for header files. AC_CHECK_HEADERS([fcntl.h glob.h libintl.h limits.h locale.h string.h strings.h sys/ioctl.h sys/param.h sys/statvfs.h sys/syslimits.h sys/time.h syslog.h wchar.h]) diff --git a/lib/libalpm/alpm.h b/lib/libalpm/alpm.h index 2ce049d..d72cb8e 100644 --- a/lib/libalpm/alpm.h +++ b/lib/libalpm/alpm.h @@ -550,7 +550,6 @@ enum _pmerrno_t { /* External library errors */ PM_ERR_LIBARCHIVE, PM_ERR_LIBFETCH, - PM_ERR_GPGME, PM_ERR_EXTERNAL_DOWNLOAD }; diff --git a/lib/libalpm/error.c b/lib/libalpm/error.c index 907a9f2..1fccc46 100644 --- a/lib/libalpm/error.c +++ b/lib/libalpm/error.c @@ -158,8 +158,6 @@ const char SYMEXPORT *alpm_strerror(int err) /* obviously shouldn't get here... */ return _("download library error"); #endif - case PM_ERR_GPGME: - return _("gpgme error"); case PM_ERR_EXTERNAL_DOWNLOAD: return _("error invoking external downloader"); /* Unknown error! */ -- 1.7.1

2 3

[pacman-dev] pacman-translators list
by partizan 08 Jul '10

08 Jul '10

Hey guys, what you think about creating separate mailing list for pacman translators? I don't read anything from pacman-dev that come to mail, but must be subscribed to this list to send translations.

5 4

[pacman-dev] [PATCH] makepkg: check explicitly if the terminal supports colors
by Cedric Staniewski 08 Jul '10

08 Jul '10

Currently, makepkg aborts when colors are enabled (the default) and it runs in a terminal that does not support it, because tput returns a non-zero exit code. --- Another solution would be appending "|| true" to each tput line or disabling errexit in this section, so we get at least bold. Not sure if this is a real improvement, when the color option makes text bold but not colored. scripts/makepkg.sh.in | 5 ++++- 1 files changed, 4 insertions(+), 1 deletions(-) diff --git a/scripts/makepkg.sh.in b/scripts/makepkg.sh.in index dbc4047..1795b54 100644 --- a/scripts/makepkg.sh.in +++ b/scripts/makepkg.sh.in @@ -1620,8 +1620,11 @@ fi PACMAN=${PACMAN:-pacman} # check if messages are to be printed using color +# It is needed to explicitly check if the terminal supports colors, because +# otherwise, tput will return a non-zero exit code that stops makepkg due to +# bash's errexit option. unset ALL_OFF BOLD BLUE GREEN RED YELLOW -if [[ -t 2 && ! $USE_COLOR = "n" && $(check_buildenv color) = "y" ]]; then +if [[ -t 2 && ! $USE_COLOR = "n" && $(check_buildenv color) = "y" ]] && tput setaf 0 &>/dev/null; then ALL_OFF="$(tput sgr0)" BOLD="$(tput bold)" BLUE="${BOLD}$(tput setaf 4)" -- 1.7.1

4 4

[pacman-dev] Ignore the series of patches about package signing
by Denis A. Altoé Falqueto 08 Jul '10

08 Jul '10

Guys, Unfortunately, the patches were corrupted by GMail when I mailed them. I'll try tomorrow by attaching them instead of inserting them in the body of the message. I just don't feel comfortable with git send-mail. Sorry for the noise. -- A: Because it obfuscates the reading. Q: Why is top posting so bad? ------------------------------------------- Denis A. Altoe Falqueto -------------------------------------------

3 3