November 2007 - Pacman-dev - lists.archlinux.org

[pacman-dev] "local version is up to date. Upgrade anyway?" question
by Xavier 23 Nov '07

23 Nov '07

For testing the last patch from Nagy that makes the "dependency cycle" warning more verbose, I wanted to do : pacman -S base But then cancel the operation. The problem is that, for every package, I got the following question : local version is up to date. Upgrade anyway? I couldn't use --noconfirm , because I didn't want to actually do the operation, only preview it. Also, --noconfirm will answer other questions than this one. Actually, this is exactly the issue that was mentioned in this thread : http://bbs.archlinux.org/viewtopic.php?id=35901 The only way was to use the --ask option, which was too complicated and has been removed. We probably don't need the power of this function anyway. But maybe we could have a replacement, just for this "local version is up to date" question. Or, is this question really useful anyway? How often do we try installing an already installed package, without the intention to reinstall it? Or maybe pacman could just print a warning instead : # pacman -S foo warning : foo is already installed ...

4 8

[pacman-dev] pacman-git releases
by fancris3 23 Nov '07

23 Nov '07

I use pacman-git from 1 or 2 weeks (local build from PKGBUILD) , its fast and stable,good job.Thank you. Name : pacman-git Version : 20071121-1 ... Architecture : x86_64 ... [~]:: pacman -Q|wc -l 1251 :) no problems for pacman [~]:: pacman -Ss|wc -l 9521 this its a feature or a bug? i have made a build for bsc-beta,i have a line: make PREFIX=$startdir/pkg/opt/bsc install || return 1 variable its wrong,makepkg make empty pkg and skip "return 1" make: Nothing to be done for `install'. ==> Tidying install... -> Removing info/doc files... -> Compressing man pages... -> Stripping debugging symbols from binarie ==> Creating package... -> Generating .FILELIST file... -> Generating .PKGINFO file... -> Compressing package... ==> Leaving fakeroot environment. ==> Finished making: bsc-beta (Thu Nov 22 19: in the stable version this thing is working correct and return an error ==> Removing info/doc files... ==> Compressing man pages... ==> Stripping debugging symbols from libraries... ==> Stripping symbols from binaries... ==> Generating .FILELIST file... tar: *: Cannot stat: No such file or directory tar: Error exit delayed from previous errors ==> Generating .PKGINFO file... ==> Compressing package... tar: *: Cannot stat: No such file or directory tar: Error exit delayed from previous errors ==> ERROR: Failed to create package file. also makepkg doesn't work with directory with spaces for example try to build pacman-git in "~/new\ folder/pacman-git"

4 6

[pacman-dev] Usefulness of -F option (was: "local version is up to date. Upgrade anyway?" question)
by Dan McGee 22 Nov '07

22 Nov '07

On Nov 21, 2007 11:46 AM, Xavier <shiningxc(a)gmail.com> wrote: > Except that -F apparently doesn't want to install new packages, only upgrade > existing but outdated ones. > While the option I had in mind for -S would either install new packages, or > upgrading existing but outdated ones. > > But I never used -F once in all the time I used Arch. Who uses this option? > I actually was surprised when I first saw it some times ago. And I already > forgot it existed, since I never see it mentioned anywhere. Does anyone even use -F anymore? I could see its usefullness before sync DBs were introduced, but I do have a hard time seeing its use anymore. Kill it? Any objections? I dislike stale crud sitting around. -Dan

7 12

[pacman-dev] [PATCH] New alpm_list_join function
by Nagy Gabor 22 Nov '07

22 Nov '07

Patch attached Bye ---------------------------------------------------- SZTE Egyetemi Könyvtár - http://www.bibl.u-szeged.hu This mail sent through IMP: http://horde.org/imp/

5 11

[pacman-dev] [PATCH 1/5] Two memleak fixes in pacman.
by Nagy Gabor 21 Nov '07

21 Nov '07

Both memleak was an unfreed alpm_db_whatprovides list. Signed-off-by: Nagy Gabor <ngaba(a)bibl.u-szeged.hu> --- src/pacman/deptest.c | 1 + src/pacman/sync.c | 1 + 2 files changed, 2 insertions(+), 0 deletions(-) diff --git a/src/pacman/deptest.c b/src/pacman/deptest.c index 301558e..d17c821 100644 --- a/src/pacman/deptest.c +++ b/src/pacman/deptest.c @@ -70,6 +70,7 @@ int pacman_deptest(alpm_list_t *targets) break; } } + alpm_list_free(provides); } if(!found) { diff --git a/src/pacman/sync.c b/src/pacman/sync.c index 5f1072f..6f3c5d5 100644 --- a/src/pacman/sync.c +++ b/src/pacman/sync.c @@ -532,6 +532,7 @@ int sync_trans(alpm_list_t *targets, int sync_only) if(prov) { pmpkg_t *pkg = alpm_list_getdata(prov); pname = alpm_pkg_get_name(pkg); + alpm_list_free(prov); break; } } -- 1.5.3.5

6 6

[pacman-dev] FREELIST macros / little memleak fix
by Xavier 21 Nov '07

21 Nov '07

1) I remember having already asked that, but it was probably just on irc, and anyway I forgot what the answer was (if there was one ;)). It's about commit a57b2f233 : http://projects.archlinux.org/git/?p=pacman.git;a=commitdiff;h=a57b2f233f28… What was wrong / messy about these macros : -/* TODO we should do away with these... they're messy*/ -#define _FREELIST(p, f) do { alpm_list_free_inner(p, f); alpm_list_free(p); p = NULL; } while(0) -#define FREELIST(p) _FREELIST(p, free) -#define FREELISTPTR(p) do { alpm_list_free(p); p = NULL; } while(0) +#define FREELIST(p) do { alpm_list_free_inner(p, free); alpm_list_free(p); p = NULL; } while(0) Hm, actually while writing this mail, I realize this it not very problematic :P In the worst case, we just need two little additional calls. But I still don't find these very messy. 2) If I mentioning this, it's because I found a little memleak in _alpm_sync_free: a missing alpm_list_free call. So maybe the above _FREELIST(p, f) macro could have helped. Also, for some reasons, both _alpm_sync_free and _alpm_trans_free did the job of alpm_list_free_inner manually. So I changed that. 3) A little problem : all _alpm_*_free don't match the alpm_list_fn_free prototype except one : _alpm_graph_free . All others don't take a void * argument, so they need to be casted to alpm_list_fn_free for being used with alpm_list_free_inner . So should "static void _alpm_graph_free(void *data)" be changed to "static void _alpm_graph_free(pmgraph_t *graph)" ? >From fdf689932ca1fa919b95533a2469f797f22970f1 Mon Sep 17 00:00:00 2001 From: Chantry Xavier <shiningxc(a)gmail.com> Date: Wed, 21 Nov 2007 17:10:20 +0100 Subject: [PATCH] Fix a memleak in _alpm_sync_free. An alpm_list_free call was missing. Also make use of alpm_list_free_inner in both _alpm_sync_free and _alpm_trans_free. Signed-off-by: Chantry Xavier <shiningxc(a)gmail.com> --- lib/libalpm/sync.c | 7 ++----- lib/libalpm/trans.c | 10 ++-------- 2 files changed, 4 insertions(+), 13 deletions(-) diff --git a/lib/libalpm/sync.c b/lib/libalpm/sync.c index 16b1998..f03a78b 100644 --- a/lib/libalpm/sync.c +++ b/lib/libalpm/sync.c @@ -74,11 +74,8 @@ void _alpm_sync_free(pmsyncpkg_t *sync) /* TODO wow this is ugly */ if(sync->type == PM_SYNC_TYPE_REPLACE) { - alpm_list_t *tmp; - for(tmp = sync->data; tmp; tmp = alpm_list_next(tmp)) { - _alpm_pkg_free(tmp->data); - tmp->data = NULL; - } + alpm_list_free_inner(sync->data, (alpm_list_fn_free)_alpm_pkg_free); + alpm_list_free(sync->data); sync->data = NULL; } else { _alpm_pkg_free(sync->data); diff --git a/lib/libalpm/trans.c b/lib/libalpm/trans.c index 19ae2b6..d998826 100644 --- a/lib/libalpm/trans.c +++ b/lib/libalpm/trans.c @@ -246,8 +246,6 @@ pmtrans_t *_alpm_trans_new() void _alpm_trans_free(pmtrans_t *trans) { - alpm_list_t *i; - ALPM_LOG_FUNC; if(trans == NULL) { @@ -256,13 +254,9 @@ void _alpm_trans_free(pmtrans_t *trans) FREELIST(trans->targets); if(trans->type == PM_TRANS_TYPE_SYNC) { - for(i = trans->packages; i; i = alpm_list_next(i)) { - _alpm_sync_free(i->data); - } + alpm_list_free_inner(trans->packages, (alpm_list_fn_free)_alpm_sync_free); } else { - for(i = trans->packages; i; i = alpm_list_next(i)) { - _alpm_pkg_free(i->data); - } + alpm_list_free_inner(trans->packages, (alpm_list_fn_free)_alpm_pkg_free); } alpm_list_free(trans->packages); -- 1.5.3.6

2 1

[pacman-dev] pacman-git releases
by Dan McGee 21 Nov '07

21 Nov '07

I bumped the version of the i686 version again today, and I've also added an x86_64 build and DB at the same place. http://www.archlinux.org/~dan/pacman-git/ You should be able to figure out the DB name from there. For anyone using the 64-bit version, please let us know just so that we can feel like its actually been tested on that architecture. Aaron- the date on your machine is all messed up, thus the messed up version label on the 64-bit package. :) -Dan

2 1

[pacman-dev] [PATCH] Improved 'dependency cycle' warning
by Nagy Gabor 21 Nov '07

21 Nov '07

Hi! I did a -Su yesterday, and I got 4 dependency cycle warnings. I became curious and I made this little patch (English clarifiaction needed). So I found 4 dependency cycle in our repos; these are not necessarily bugs but imho they worth mentioning: glibc->tzdata->glibc readline->glibc->bash->readline coreutils->shadow->coreutils db->coreutils->pam->db Bye ---------------------------------------------------- SZTE Egyetemi Könyvtár - http://www.bibl.u-szeged.hu This mail sent through IMP: http://horde.org/imp/

3 2

[pacman-dev] Valgrind updated, blog article plug
by Dan McGee 21 Nov '07

21 Nov '07

I rebuilt valgrind tonight against glibc 2.7, so you may see improved performance using this tool with your pacman debugging and memleak-finding expeditions. I'm mainly letting you all know this because Xavier asked me to write up something on how I use valgrind with pacman. I threw up a short little article on my blog (http://toofishes.net/blog/) to help people get started, the actual article is here: http://toofishes.net/blog/using-valgrind-c-programming/ If you guys have any questions in the future about how I do stuff, whether its coding, algorithms, tools I use, etc. *please* feel free to ask and I will try to answer, either privately or publicly. -Dan P.S. I'm already number 5 on Google (OK, I had to cheat a bit with the search terms). :) http://www.google.com/search?source=ig&hl=en&rlz=&q=valgrind+c

1 0

[pacman-dev] [GIT] The official pacman repository branch, master, updated. v3.0.0-612-g83fa6aa
by Dan McGee 21 Nov '07

21 Nov '07

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 discards c419277dd2633b19a2a989f5b2a11cd9b4f4375b (commit) via 83fa6aa2896c0c940e8bf5b7499abc7ef955e658 (commit) This update added new revisions after undoing existing revisions. That is to say, the old revision is not a strict subset of the new revision. This situation occurs when you --force push a change and generate a repository containing something like this: * -- * -- B -- O -- O -- O (c419277dd2633b19a2a989f5b2a11cd9b4f4375b) \ N -- N -- N (83fa6aa2896c0c940e8bf5b7499abc7ef955e658) When this happens we assume that you've already had alert emails for all of the O revisions, and so we here report only the revisions in the N branch from the common base, B. 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 83fa6aa2896c0c940e8bf5b7499abc7ef955e658 Author: Chantry Xavier <shiningxc(a)gmail.com> Date: Sat Aug 25 00:10:40 2007 +0200 Remove duplicated get_upgrades function, use sysupgrade instead. The alpm_get_upgrades was exactly the same as find_replacements + _alpm_sync_sysupgrade, except that it automatically made the eventual replacements, without asking the user : Replace %s with %s/%s? [Y/n] The replace question, asked in find_replacements. can now be skipped by using a NULL trans argument, so that we get the same behavior as with alpm_get_upgrades. So alpm_db_get_upgrades() can now be replaced by alpm_sync_sysupgrade(db_local, syncdbs). Signed-off-by: Chantry Xavier <shiningxc(a)gmail.com> Signed-off-by: Dan McGee <dan(a)archlinux.org> ----------------------------------------------------------------------- Summary of changes: hooks/post-receive -- The official pacman repository

1 0