- Pacman-dev - lists.archlinux.org

[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