- Pacman-dev - lists.archlinux.org

[pacman-dev] Pacman's new feature
by Fabiano Furtado Pessoa Coelho 10 Jan '07

10 Jan '07

Hi people, I´v been thinking about new feature on pacman. Why not to show the each package size on the install package list, just befere the download/instalation? Like: <package name> (SIZE) | <package name> (SIZE) | <package name> (SIZE) Thanks in advance... Fusca PS: Sorry about my english

2 1

[pacman-dev] pacman packaging
by Dan McGee 10 Jan '07

10 Jan '07

This was just another thought of mine. Pacman is currently distributed as one package on Archlinux (and I would assume on any distribution that is using it). I was thinking it should maybe be two packages, pacman and pacman-utils, or something similar. The two things can be split fairly cleanly- the first is used for installing packages and general maintenence, while the other is used for building packages, looking at the ABS tree, building a custom repo, etc. This also seems more important with a few other scripts being added recently (re-pacman, repo-add) pacman: pacman (and conf/man files), pacman.static, vercmp, rankmirrors, pacman-optimize pacman-utils: abs (and conf files), makepkg (and conf file), makeworld (TODO still wondering if this is used), gensync, resync, repo-add (TODO should this be renamed to addsync?), re-pacman In addition, pacman-utils could then pull in scripts such as srcpac and namcap that are likely installed by most developers anyway. This isn't so much development related, but I thought this was the best list to bring this up. -Dan

4 7

[pacman-dev] Install files and messages
by Dan McGee 10 Jan '07

10 Jan '07

I don't know what other peoples thoughts are, but I think the current install file system and scriptlets...seems like a drop in at some point that wasn't fully done right. Most packages currently use the file for one of two purposes that I can see: 1. display informational messages 2. perform operations that can't be done in the build phase. The install scriptlets were likely originally meant more for number 2, but has become more and more about providing number 1. Here are some relevant bugs and postings: metadata messages in PKGBUILD <http://bugs.archlinux.org/task/1571> display all post-install messages at once <http://bugs.archlinux.org/task/6143> global on-install system <http://bbs.archlinux.org/viewtopic.php?p=219152#219152> Here are my thoughts on the situation. As every package builder seems to have their own presentation format for displaying messages (e.g. echo ">>>...", echo " **...**"), it would be nice to have a way to unify it. Most (all?) messages are in one of two places- post_install or post_upgrade. I propose these messages should be moved out of these scriptlets, and moved into a variable, likely still located in the install file as this is packaged up with the package (unlike the PKGBUILD). These variables could be called install_msg and upgrade_msg, respectively. The scriptlets would still remain, but any output would be 'nonessential'. By changing where the messages stem from, it would make it easier for bug 6143 to be fixed, and would also make for cleaner output to pacman.log, now that all messages are being logged there. Finally, _alpm_runscriptlet seems pretty complicated...I'm wondering if it couldn't be cleaned up a bit by someone that is better with forks and popens. There is also some duplicate code. -Dan

4 5

[pacman-dev] makepkg fix (FS #5923)
by James Rosten 06 Jan '07

06 Jan '07

This is a fix for Flyspray bug #5923 http://bugs.archlinux.org/task/5923. Patch is below. ~ Jamie / yankees26

3 2

[pacman-dev] Long functions and other minor fixes
by James Rosten 03 Jan '07

03 Jan '07

This is my massacre of _alpm_load_pkg in package.c (just joking about the massacre). I moved the extremely large for loop into its own function (and made it a while loop). Feel free to change the name of _alpm_pkg_checker (first thing that came to mind). The patches for package.c and package.h are below. ~ Jamie / yankees26

3 4

[pacman-dev] My $0.02 on the Frugalware v.s. Arch bit...
by Scott Horowitz 02 Jan '07

02 Jan '07

I hate to wade into this, but let me just say... there's no doubt in my mind that this will never be resolved while the (inflammatory) sentiment in the pacman-g2 README exists. For starters, the whole point being made about the API not being stable, which seems to be points 1 and 2 in the README, are plain silly. No project should take a project's cvs development code and expect the API to remain stable; Frugalware knew what it was getting into when it grabbed this code to put into their repositories. There's a reason that pacman3 hasn't been released yet on Archlinux, and that there haven't been Archlinux frontends developed. Pacman3 development shouldn't be hindered because someone decided to release the code into the wild, and it's clearly going under some large refactoring. As for point 3, Aaron's supposed disinterest in contributor's patches - I think it has been quite clear from the mailing list that this isn't the case. The real justification behind this sentiment, which was brought up in the original email and I agree needs improvement, is that there could use more openness (such as in response to a WONTFIX patch) between all the parties. Then there's the stuff about how development has fallen behind schedule (what project hasn't?) and Aaron asked Frugalware to fork things (I'm pretty sure this was more tongue-in-cheeck about how Frugalware is doing things than anything). These seem pretty petty to me. In my opinion, two things need to happen, one from each side. On the Archlinux side, I think Aaron should make a better attempt to be more transparent in his responses, which would hopefully encourage debate. On the Frugalware side, I think the whining about API stability needs to be dropped and, rather than take the "fork" talk as a challenge, they should be asking what needs to be done to make the two paths more compatible. Happy Holidays, Scott

7 11

[pacman-dev] pacman-optimize updates
by James Rosten 31 Dec '06

31 Dec '06

I came up with a patch for FS #5009 (the error when /var/lib/pacman is a loopback device or other mounted partition. I have not been able to test it yet, as I am on my mac (so sending this with Mail, so let us hope it doesn't mess this up :P), but in a few hours I'll be on my Arch computer and I'll test it then and obviously report back. Also, fixed the step number in the comments ('cause I am that nit pickish), because there were two step 1's. ~ Jamie / yankees26

1 0

[pacman-dev] Long functions and other minor fixes
by James Rosten 30 Dec '06

30 Dec '06

I've refactored _alpm_checkconflicts, but there appears to be something slightly off. With a normal cvs build of pacman3, trying to install sonata-svn with normal sonata produces this: loading package data... done.checking dependencies... :: sonata-svn conflicts with sonata. Remove sonata? [Y/n] y error: failed to prepare transaction (conflicting dependencies) :: sonata-svn: conflicts with sonata But with my patch: loading package data... done.checking dependencies... done. checking for file conflicts (1/1) [###############################################################] 100% error: failed to prepare transaction (conflicting files) sonata-svn: /usr/bin/sonata exists in filesystemsonata-svn: /usr/lib/python2.4/site-packages/mmkeys.so exists in filesystemsonata-svn: /usr/lib/python2.4/site-packages/mpdclient3.py exists in filesystemsonata-svn: /usr/lib/python2.4/site-packages/mpdclient3.pyc exists in filesystemsonata-svn: /usr/lib/python2.4/site-packages/sonata.py exists in filesystemsonata-svn: /usr/lib/python2.4/site-packages/sonata.pyc exists in filesystemsonata-svn: /usr/share/applications/sonata.desktop exists in filesystemsonata-svn: /usr/share/locale/de/LC_MESSAGES/sonata.mo exists in filesystemsonata-svn: /usr/share/locale/fr/LC_MESSAGES/sonata.mo exists in filesystemsonata-svn: /usr/share/locale/pl/LC_MESSAGES/sonata.mo exists in filesystemsonata-svn: /usr/share/locale/ru/LC_MESSAGES/sonata.mo exists in filesystemsonata-svn: /usr/share/pixmaps/sonata.png exists in filesystemsonata-svn: /usr/share/pixmaps/sonatacd.png exists in filesystemsonata-svn: /usr/share/pixmaps/sonatacd_large.png exists in filesystemsonata-svn: /usr/share/sonata/CHANGELOG exists in filesystemsonata-svn: /usr/share/sonata/README exists in filesystemsonata-svn: /usr/share/sonata/TODO exists in filesystemsonata-svn: /usr/share/sonata/TRANSLATORS exists in filesystem errors occurred, no packages were upgraded. As you can see it flipped out, but even with the normal pacman3 it still fails to install. Both were executed with: sudo ./pacman.static -A /var/abs/local/sonata-svn/sonata-svn-246-1.pkg.tar.gz Using -U instead of -A yields the same results. So below is the patch for you guys to look over and see what might be screwing this up (cause I have no idea). ~ Jamie / yankees26

1 0

[pacman-dev] Long functions and other minor fixes
by James Rosten 30 Dec '06

30 Dec '06

After trying to build pacman3 so I could test something for stonecrest, I found some errors in my patch and Dan's (very small, you forgot to put the label error before return(-1);) while compiling. So below is the better patch than the other patch (which should be ignored). It patches package.c, package.h, and sync.c of libalpm. ~ Jamie / yankees26

1 0

[pacman-dev] Long functions and other minor fixes
by Dan McGee 29 Dec '06

29 Dec '06

Some more functions that are worth taking a look at on the libalpm side: add.c: _alpm_add_commit be_files.c: _alpm_db_read _alpm_db_write conflict.c: _alpm_checkconflicts _alpm_db_find_conflicts package.c: _alpm_pkg_load (getting a bit long) remove.c: _alpm_remove_commit server.c: _alpm_downloadfiles_forreal sync.c: _alpm_sync_sysupgrade (getting there) _alpm_sync_prepare _alpm_sync_commit And some minor things I found in the code- cleaned up an extern enum, changed 'md5' variable to 'hash' as it can be either md5 or sha1, resorted db.h for clarity, and fixed some English grammar issues. Here is a patch: Index: lib/libalpm/alpm.h =================================================================== RCS file: /home/cvs-pacman/pacman-lib/lib/libalpm/alpm.h,v retrieving revision 1.64 diff -u -u -r1.64 alpm.h --- lib/libalpm/alpm.h 22 Nov 2006 09:03:41 -0000 1.64 +++ lib/libalpm/alpm.h 22 Dec 2006 18:34:01 -0000 @@ -399,7 +399,7 @@ /* * Errors */ -extern enum __pmerrno_t { +enum __pmerrno_t { PM_ERR_MEMORY = 1, PM_ERR_SYSTEM, PM_ERR_BADPERMS, @@ -469,7 +469,9 @@ /* Downloading */ PM_ERR_CONNECT_FAILED, PM_ERR_FORK_FAILED -} pm_errno; +}; + +extern enum __pmerrno_t pm_errno; char *alpm_strerror(int err); Index: lib/libalpm/backup.c =================================================================== RCS file: /home/cvs-pacman/pacman-lib/lib/libalpm/backup.c,v retrieving revision 1.7 diff -u -u -r1.7 backup.c --- lib/libalpm/backup.c 31 Oct 2006 06:39:59 -0000 1.7 +++ lib/libalpm/backup.c 22 Dec 2006 18:34:01 -0000 @@ -55,9 +55,9 @@ ptr++; /* now str points to the filename and ptr points to the md5 or sha1 hash */ if(!strcmp(file, str)) { - char *md5 = strdup(ptr); + char *hash = strdup(ptr); FREE(str); - return(md5); + return(hash); } FREE(str); } Index: lib/libalpm/db.h =================================================================== RCS file: /home/cvs-pacman/pacman-lib/lib/libalpm/db.h,v retrieving revision 1.20 diff -u -u -r1.20 db.h --- lib/libalpm/db.h 20 Nov 2006 09:10:24 -0000 1.20 +++ lib/libalpm/db.h 22 Dec 2006 18:34:01 -0000 @@ -44,10 +44,14 @@ pmlist_t *servers; }; +/* db.c, database general calls */ pmdb_t *_alpm_db_new(char *root, char *dbpath, char *treename); void _alpm_db_free(void *data); int _alpm_db_cmp(const void *db1, const void *db2); pmlist_t *_alpm_db_search(pmdb_t *db, pmlist_t *needles); +pmdb_t *_alpm_db_register(char *treename, alpm_cb_db_register callback); + +/* be.c, backend specific */ int _alpm_db_install(pmdb_t *db, const char *dbfile); int _alpm_db_open(pmdb_t *db); void _alpm_db_close(pmdb_t *db); @@ -58,7 +62,6 @@ int _alpm_db_remove(pmdb_t *db, pmpkg_t *info); int _alpm_db_getlastupdate(pmdb_t *db, char *ts); int _alpm_db_setlastupdate(pmdb_t *db, char *ts); -pmdb_t *_alpm_db_register(char *treename, alpm_cb_db_register callback); #endif /* _ALPM_DB_H */ Index: lib/libalpm/sync.c =================================================================== RCS file: /home/cvs-pacman/pacman-lib/lib/libalpm/sync.c,v retrieving revision 1.85 diff -u -u -r1.85 sync.c --- lib/libalpm/sync.c 1 Dec 2006 09:32:30 -0000 1.85 +++ lib/libalpm/sync.c 22 Dec 2006 18:34:03 -0000 @@ -382,7 +382,7 @@ { pmlist_t *deps = NULL; pmlist_t *list = NULL; /* list allowing checkdeps usage with data from trans->packages */ - pmlist_t *trail = NULL; /* breadcrum list to avoid running into circles */ + pmlist_t *trail = NULL; /* breadcrumb list to avoid running in circles */ pmlist_t *asked = NULL; pmlist_t *i, *j, *k, *l; int ret = 0;

2 4

[pacman-dev] pacman-optimize updates
by Dan McGee 28 Dec '06

28 Dec '06

pacman-optimize updates. Bug #5009 is out there, I'll think about it and try to make a patch hopefully without screwing up my own database. -Dan Index: pacman-optimize =================================================================== RCS file: /home/cvs-pacman/pacman-lib/scripts/pacman-optimize,v retrieving revision 1.3 diff -u -r1.3 pacman-optimize --- pacman-optimize 21 Dec 2006 01:53:41 -0000 1.3 +++ pacman-optimize 22 Dec 2006 19:37:41 -0000 @@ -20,7 +20,7 @@ # USA. # -myver='2.9.8' +myver='3.0.0' usage() { echo "pacman-optimize $myver" @@ -59,10 +59,6 @@ dbroot=$1 fi -if [ "`id -u`" != 0 ]; then - die "You must be root to optimize the database" -fi - # make sure pacman isn't running if [ -f /tmp/pacman.lck ]; then die "Pacman lockfile was found. Cannot run while pacman is running." @@ -72,6 +68,11 @@ die "$dbroot does not exist or is not a directory" fi +#if [ "$EUID" != 0 ]; then +if [ ! -w "$dbroot" ]; then + die "You must have correct permissions to optimize the database" +fi + # don't let pacman run while we do this touch /tmp/pacman.lck

2 2

[pacman-dev] New mirrors file
by Dan McGee 28 Dec '06

28 Dec '06

Saw this (http://bugs.archlinux.org/task/6083) bug today on the bugtracker and also knew this was a problem anyway, so here is my go at it. I tried to make a more up to date mirror list from both the Wiki and the download page. I ran the rankmirrors script on it a few times to hopefully weed out the old/bad mirrors. If someone that is skilled with automake/autoconf wants to finish the job, I'd appreciate it. Instead of having four more or less duplicate files in our repository, we could generate all of them from this one file. -Dan /etc/pacman.d/mirrorlist.in: # # @REPO@: Arch Linux @REPO@ repository # # United States Server = ftp://ftp.archlinux.org/@REPO@/os/@CARCH@ Server = ftp://ftp.nethat.com/pub/linux/archlinux/@REPO@/os/@CARCH@ Server = ftp://ftp-linux.cc.gatech.edu/pub/linux/distributions/archlinux/@REPO@/os/@… Server = ftp://mirror.cs.vt.edu/pub/ArchLinux/@REPO@/os/@CARCH@ Server = ftp://ftp.ibiblio.org/pub/linux/distributions/archlinux/@REPO@/os/@CARCH@ Server = http://www2.cddc.vt.edu/linux/distributions/archlinux/@REPO@/os/@CARCH@ # Europe # - Austria Server = ftp://gd.tuwien.ac.at/opsys/linux/archlinux/@REPO@/os/@CARCH@ # - Belgium Server = ftp://ftp.belnet.be/packages/archlinux/@REPO@/os/@CARCH@ # - Czech Republic Server = ftp://ftp.sh.cvut.cz/MIRRORS/arch/@REPO@/os/@CARCH@ # - Estonia Server = ftp://ftp.estpak.ee/pub/archlinux/@REPO@/os/@CARCH@ # - France Server = ftp://ftp.rez-gif.supelec.fr/pub/Linux/distrib/archlinux/@REPO@/os/@CARCH@ Server = ftp://mir1.archlinuxfr.org/archlinux/@REPO@/os/@CARCH@ Server = ftp://mir2.archlinuxfr.org/archlinux/@REPO@/os/@CARCH@ Server = ftp://distrib-coffee.ipsl.jussieu.fr/pub/linux/archlinux/@REPO@/os/@CARCH@ #Server = http://mirrors.jakimowicz.com/archlinux/@REPO@/os/@CARCH@ # - Germany #Server = ftp://ftp.archlinux.de/pub/archlinux/@REPO@/os/@CARCH@ Server = ftp://ftp.tu-chemnitz.de/pub/linux/sunsite.unc-mirror/distributions/archlinux/@REPO@/os/@CARCH@ Server = ftp://ftp.hosteurope.de/mirror/ftp.archlinux.org/@REPO@/os/@CARCH@ # - Great Britain Server = http://ftp.parrswood.manchester.sch.uk/Mirrors/ftp.archlinux.org/@REPO@/os/… # - Greece Server = ftp://ftp.ntua.gr/pub/linux/archlinux/@REPO@/os/@CARCH@ # - Ireland Server = ftp://ftp.heanet.ie/mirrors/ftp.archlinux.org/@REPO@/os/@CARCH@ # - Italy Server = ftp://mi.mirror.garr.it/mirrors/archlinux/@REPO@/os/@CARCH@ # - Lithuania Server = ftp://saule.mintis.lt/pub/linux/@REPO@/os/@CARCH@ # - Netherlands Server = ftp://ftp.nluug.nl/pub/metalab/distributions/archlinux/@REPO@/os/@CARCH@ Server = ftp://ftp.surfnet.nl/pub/os/Linux/distr/archlinux/@REPO@/os/@CARCH@ # - Norway Server = ftp://mirror.djzradio.net/archlinux/@REPO@/os/@CARCH@ # - Poland Server = ftp://sunsite.icm.edu.pl/site/linux-sunsite/distributions/archlinux/@REPO@/os/@CARCH@ Server = ftp://mirror.icis.pcz.pl/archlinux/@REPO@/os/@CARCH@ # - Portugal Server = ftp://gul.est.ips.pt/repos/archlinux/@REPO@/os/@CARCH@ Server = ftp://cesium.di.uminho.pt/pub/archlinux/@REPO@/os/@CARCH@ Server = ftp://darkstar.ist.utl.pt/pub/archlinux/@REPO@/os/@CARCH@ # - Romania Server = ftp://ftp.iasi.roedu.net/mirrors/archlinux.org/@REPO@/os/@CARCH@ # - Sweden Server = ftp://ftp.ds.hj.se/pub/os/linux/archlinux/@REPO@/os/@CARCH@ Server = ftp://ftp.gigabit.nu/@REPO@/os/@CARCH@ # - Switzerland Server = ftp://archlinux.puzzle.ch/@REPO@/os/@CARCH@ # Australia Server = ftp://mirror.pacific.net.au/linux/archlinux/@REPO@/os/@CARCH@ # Asia # - China Server = http://mirror.vxmatrix.net/ArchLinux/@REPO@/os/@CARCH@ Server = http://dxmirror.vxmatrix.net/ArchLinux/@REPO@/os/@CARCH@ Server = http://cncmirror.vxmatrix.net/ArchLinux/@REPO@/os/@CARCH@

4 4

[pacman-dev] [PATCH] conditional build of pacman.static
by Jürgen Hötzel 28 Dec '06

28 Dec '06

Hi, #1 from TODO.autoconf Jürgen

2 1

[pacman-dev] My $0.02 on the Frugalware v.s. Arch bit...
by Alex Smith 27 Dec '06

27 Dec '06

Hi all, I hope you all don't mind me writing this - I'm going to try to write it in a way that seems neutral (i.e. not on the Arch side or the Frugalware side), even though I'm from the Frugal side of things. And just in case you're wondering why I'm writing this here instead of telling you, VMiklos, I think this implies to both sides, therefore belongs here. Anyway, I'm going to start by saying bluntly what I think: This is just pathetic bickering. For goodness' sake, it's a package manager, not the world's most sacred thing. Why do I say this? Let's start with the MaxTries patch. Aaron said that it was "not valid". I've searched the archives of this list, and I can't find any reason for this. Yet, I've found reason for this here: http://bbs.archlinux.org/viewtopic.php?p=216399#216399 It would have been far more appropriate to say that here, then the problem with this patch could have been rectified. However, here you just said "not valid". Now, about our "patch queue". This wasn't a seperate codebase, it was simply a repo for us to develop our patches in and find/fix bugs before submitting to upstream. It was never a fork - we simply never had upstream commit access, so we had to do this. Personally, I don't think we wanted to fork, however it's hard to work with someone who says stuff is not valid without giving reason. But really, wouldn't it be so much easier if this was the case: 1. VMiklos creates the MaxTries patch, discusses it here. 2. Other people point out any problems/flaws with it, in a relaxed manner, rather than having a shouting match with 20 foot flames all around. 3. Problems get sorted after discussion, it gets committed. The main point to that is the relaxed manner. I'll repeat what I said at the beginning: this is a package manager, a small application, not something so important that the world will blow up if something goes wrong. I really think that we should merge back into 1 project, rather than leading 2 seperate paths with 2 different APIs, and co-operate better, have fun coding, instead of this bickering. Having 2 APIs would just suck - someone could come along and create a really great libalpm frontend that works on Frugalware, and then the Arch users will be disappointed when it won't even compile for them. We've cooperated up until now in relaxed way, why change it. But anyway, as I said in the topic, this is just my $0.02, so, choose to ignore it if you wish, but I seriously think that a relaxed attitude is the key here. Thanks, Alex ------ Alex Smith, Frugalware Linux developer - http://www.frugalware.org

2 1

[pacman-dev] IDEA: checkdb script
by Roman Kyrylych 24 Dec '06

24 Dec '06

Hi! After yet another local database corruption I've got an idea for a script that will do the following: go through files in /var/lib/pacman/local/ and check dependencies of them with those, listed in /var/lib/pacman/{current,extra,testing,unstable,community} and try to repair any found problems. For example both pacman and pacman3 write wrong dependency info into local database after some actions with some packages (there are few bugs about this already, but some may be unknown) - this script could fix this by rebuilding dependency tree from repos (it's highly recommended that all packages in current are up-to-date for this, but not strictly necessary, because depends are changed not so frequently). Of course there is the problem with this - what if this script will build wrong dependency tree or have other bugs. :p But I think there shouldn't be big problem to ensure it will work like swiss watch, and it can be used to find bugs in pacman when doing -S / -R intensively. What do you think about this? -- Roman Kyrylych (Роман Кирилич)

3 2

[pacman-dev] anoter "user vs. sudo" issue
by Roman Kyrylych 22 Dec '06

22 Dec '06

I'm thinking about why abs should be run as root currently? I don't see the point for this. :-/ Why not make it behave like makepkg - to be run as ordinary user and store everything in his home dir (or another dir, configurable)? What do you think? -- Roman Kyrylych (Роман Кирилич)

5 5

[pacman-dev] abs script changes
by Dan McGee 22 Dec '06

22 Dec '06

I don't think I can look at a whole lot without wanting to change/fix the code. Some changes on the abs script: 1. Remove old update function, and patch new one as needed. 2. Remove extra $ABSROOT directory check 3. Remove duplicate (and I believe unnecessary?) block of code to check for passive mode flag. This double code is actually in the current abs code, is there a reason? Also added a to-do note, let me know what you think. -Dan Index: abs =================================================================== RCS file: /home/cvs-pacman/pacman-lib/scripts/abs,v retrieving revision 1.2 diff -u -r1.2 abs --- abs 21 Dec 2006 01:53:41 -0000 1.2 +++ abs 21 Dec 2006 22:32:59 -0000 @@ -18,27 +18,17 @@ for sup in "${SUPFILES[@]}"; do if [ "$sup" != "testing" ]; then if [ "$sup" = "${sup#!}" ]; then - cvsup -L 1 -r 0 -g -b $ABSROOT -c .sup /etc/abs/supfile.$sup + $CVSUP -L 1 -r 0 -g -b $ABSROOT -P $CONNMODE -c .sup /etc/abs/supfile.$sup fi elif [ "$sup" = "testing" ]; then if [ ! -d /var/abs/testing ]; then mkdir /var/abs/testing; fi cd $ABSROOT/testing - cvsup -L 1 -r 0 -g -b $ABSROOT/testing -c .sup /etc/abs/supfile.testing + $CVSUP -L 1 -r 0 -g -b $ABSROOT/testing -P $CONNMODE -c .sup /etc/abs/supfile.testing cd $ABSROOT fi done -} - - -update() { - cd $ABSROOT - for sup in "${SUPFILES[@]}"; do - if [ "$sup" = "${sup#!}" ]; then - $CVSUP -L 1 -r 0 -g -b $ABSROOT -P $CONNMODE -c .sup /etc/abs/supfile.$sup - fi - done } if [ "$1" = "-h" ] || [ "$1" = "--help" ]; then @@ -46,11 +36,12 @@ exit 0 fi -if [ ! -d $ABSROOT ]; then +if [ ! -d "$ABSROOT" ]; then echo "abs: $ABSROOT does not exist (or is not a directory)" exit 1 fi -if [ ! -w $ABSROOT ]; then + +if [ ! -w "$ABSROOT" ]; then echo "abs: no write permissions in $ABSROOT" exit 1 fi @@ -64,24 +55,12 @@ exit 1 fi -if [ ! -d "$ABSROOT" ]; then - echo "abs: directory $ABSROOT does not exist" - exit 1 -fi - -if [ "$1" = "-p" ] || [ "$1" = "--passive" ]; then - CONNMODE="-" - shift -else - CONNMODE="m" - shift -fi - if [ "$1" = "-p" ] || [ "$1" = "--passive" ]; then CONNMODE="-" shift else CONNMODE="m" + #TODO: do we want to shift if this argument was not present? shift fi

3 7

[pacman-dev] [PATCH] configure.ac: remove unnecessary swig-checks + quote possible unset variables in test
by Jürgen Hötzel 22 Dec '06

22 Dec '06

Hi, configure checked 3 times for swig. Only one check is required for all swig-based bindings. path enclosed. Jürgen

3 2

[pacman-dev] Organization: pacman todo and timeline
by Aaron Griffin 22 Dec '06

22 Dec '06

Ok. So I think it's time i get this written up. Right now, I plan on getting pacman3 into [testing] ASAP. There were some NoUpgrade issues (previous thread) and they have been "handled" for the time being. This means that stability is the single most important thing. So, I need any and all available people to test this out: http://archlinux.org/~aaron/pacman/ I will periodically push new versions here based on changes which will most likely be discussed on this list. Now, beyond that, I have been keeping a "TODO.aaron" file separate from the other stuff. There is alot of reorganization I'd like to do before I get moving on any newer (complex) features. First and foremost, having functions which are over 100-200 lines tend to imply that something needs to be split out. Alot of these functions can be split out into logical units. For instance, _alpm_add_commit (lib/libalpm/add.c) contains a large chunk of lines for determining / checking md5sums. This would make alot of sense as a separate function which would, say, determine the ACTUAL output file. This is just an example. If anyone needs something to work on on the C level, I would HIGHLY appreciate some refactoring help here. Feel free to post any and all suggestions - nothing is stupid. So as for a timeless timeline, the plans are as follows: * release pacman3 to the masses * accept bug reports and fix what people complain about a lot * refactor the libalpm / pacman internals and remove some of the overlap * feature time (I have about 3 or 4 really good ideas) Unordered, but also important: * new backends (selectable at runtime?) * make the public libalpm interface easier to use

1 0

[pacman-dev] rankmirrors script
by James Rosten 22 Dec '06

22 Dec '06

Because rankmirrors just tests a whole file, I made a variation that takes a mirror's url as an argument and tests just that mirror. Sorry if the license information is wrong at the top (I got it from GNUs website), I noticed differences between it and the one that is in all the source files in pacman-lib's cvs. I think the only difference was only the address of FSF. http://jrosten.googlecode.com/svn/testmirror is where the python script can be found (pasting made it look funky, I am starting to not like copy and paste). Criticize away. ~ Jamie R.

1 0

[pacman-dev] [PATCH] fix in_array (usage) and code cleanup
by Jürgen Hötzel 21 Dec '06

21 Dec '06

Hi, the use of in_array in line 444 is invalid: bash-3.2$ makepkg-cvs /home/juergen/archlinux/pacman-lib/scripts/makepkg: line 444: [: too many arguments Helper function like in_array should not use global variables. I also did some code cleanup (the temp. file "filelist" was not even used). Less Code. Better Code. patch enclosed. Jürgen

2 1

[pacman-dev] My makepkg changes/updates
by Dan McGee 21 Dec '06

21 Dec '06

Last email of the day, hopefully. Here is my current diff on makepkg, not really meant to be applied in its current state but at least letting you know what I am working on and hopefully I can get some feedback out of it. I'll make a cleaner diff in a few days and hopefully this script will start to look a little better. 1. Adds --repackage option and $REPKG variable, should be noted by a comment in the code. The diff does some weird things where I had to indent a whole big block of code, so I'd recommend just trying to look at it as a whole at that part (basically the whole 'build' section). 2. Adds and removes blank lines to be more consistent, or maybe just more for my own readability. Also adds a description of makepkg. 3. Removes keepdocs and nostrip options (on the command line). 4. Removes hardcoded references to /var/cache/pacman/src. 5. Fixes a syntax warning in the architecture detection. 6. Adds a missing / in the cp statement for source file copying. 7. Lets you see my scattered/random TODOs throughout the code. -Dan --- pacman-lib.orig/scripts/makepkg 2006-12-20 21:30:06.000000000 -0500 +++ pacman-lib/scripts/makepkg 2006-12-20 22:19:46.000000000 -0500 @@ -1,6 +1,6 @@ #!/bin/bash # -# makepkg +# makepkg - make packages compatable for use with pacman # # Copyright (c) 2002-2006 by Judd Vinet <jvinet(a)zeroflux.org> # Copyright (c) 2005 by Aurelien Foret <orelien(a)chez.com> @@ -44,19 +44,21 @@ FORCE=0 GENINTEG=0 INSTALL=0 -DOWNLOAD="" NOBUILD=0 NODEPS=0 NOEXTRACT=0 RMDEPS=0 +REPKG=0 LOGGING=0 PACMAN_OPTS= -INFAKEROOT= +# determine if we are running with fakeroot if [ "$1" = "-F" ]; then INFAKEROOT=1 shift +else + INFAKEROOT=0 fi ### SUBROUTINES ### @@ -68,6 +70,7 @@ echo " $1" >&2 fi } + msg() { if [ "$USE_COLOR" = "Y" -o "$USE_COLOR" = "y" ]; then echo -e "\033[1;32m==>\033[1;0m \033[1;1m$1\033[1;0m" >&2 @@ -91,6 +94,7 @@ echo "==> WARNING: $1" >&2 fi } + error() { if [ "$USE_COLOR" = "Y" -o "$USE_COLOR" = "y" ]; then echo -e "\033[1;31m==> ERROR:\033[1;0m \033[1;1m$1\033[1;0m" >&2 @@ -239,6 +243,7 @@ fi missingdeps=1 fi + # rerun any additional sh scripts found in /etc/profile.d/ for i in /etc/profile.d/*.sh do @@ -246,8 +251,6 @@ . $i &>/dev/null fi done - # some applications (eg, blackbox) will not build with some languages - unset LC_ALL LANG return $missingdeps } @@ -269,18 +272,15 @@ echo " -h, --help This help" echo " -i, --install Install package after successful build" echo " -j <jobs> Set MAKEFLAGS to \"-j<jobs>\" before building" - echo " -k, --keepdocs Keep doc and info directories" echo " -L, --log Log package build process" echo " -m, --nocolor Disable colorized output messages" - echo " -n, --nostrip Do not strip symbols from binaries/libraries" echo " -o, --nobuild Download and extract files only" echo " -p <buildscript> Use an alternate build script (instead of '$BUILDSCRIPT')" echo " -r, --rmdeps Remove installed dependencies after a successful build" + # fix flyspray feature request #2978 + echo " -R, --repackage Repackage contents of $startdir/pkg without building" echo " -s, --syncdeps Install missing dependencies with pacman" echo " -S, --sudosync Install missing dependencies with pacman and sudo" - # fix flyspray feature request #5223 - Dan McGee <dpmcgee(a)gmail.com> - echo " -t <sourcedir> Cache source files in <sourcedir>" - echo " -w <destdir> Write package to <destdir> instead of the working dir" echo echo "These options can be passed to pacman:" echo @@ -294,15 +294,12 @@ ARGLIST=$@ if [ -f /etc/makepkg.conf ]; then - source /etc/makepkg.conf + source /etc/makepkg3.conf else - error "/etc/makepkg.conf not found. cannot continue" + error "/etc/makepkg3.conf not found. Cannot continue." exit 1 fi -#Let's be courteous and support frugalware's extensions -[ -e /usr/lib/frugalware/fwmakepkg ] && . /usr/lib/frugalware/fwmakepkg - while [ "$#" -ne "0" ]; do case $1 in # pacman @@ -319,12 +316,11 @@ --noextract) NOEXTRACT=1 ;; --install) INSTALL=1 ;; --force) FORCE=1 ;; - --keepdocs) KEEPDOCS=1 ;; - --nostrip) NOSTRIP=1 ;; --nobuild) NOBUILD=1 ;; --nocolor) USE_COLOR="n" ;; --geninteg) GENINTEG=1 ;; --rmdeps) RMDEPS=1 ;; + --repackage) REPKG=1 ;; --log) LOGGING=1 ;; --help) usage @@ -335,7 +331,7 @@ exit 1 ;; -*) - while getopts "bBcCdefghij:kLmnop:rsSt:w:-" opt; do + while getopts "bBcCdefghij:Lmop:rRsS-" opt; do case $opt in b) DEP_SRC=1 ;; B) USE_CCACHE=0 ;; @@ -346,30 +342,27 @@ f) FORCE=1 ;; g) GENINTEG=1 ;; h) - usage - exit 0 - ;; + usage + exit 0 + ;; i) INSTALL=1 ;; j) export MAKEFLAGS="-j$OPTARG" ;; - k) KEEPDOCS=1 ;; L) LOGGING=1 ;; m) USE_COLOR="n" ;; - n) NOSTRIP=1 ;; o) NOBUILD=1 ;; p) BUILDSCRIPT=$OPTARG ;; r) RMDEPS=1 ;; + R) REPKG=1 ;; s) DEP_BIN=1 ;; S) DEP_SUDO=1 ;; - t) SRCDEST=$OPTARG ;; - w) PKGDEST=$OPTARG ;; -) - OPTIND=0 - break - ;; + OPTIND=0 + break + ;; *) - usage - exit 1 - ;; + usage + exit 1 + ;; esac done ;; @@ -386,6 +379,7 @@ exit 1 fi +#TODO: is this necessary? # convert a (possibly) relative path to absolute cd $PKGDEST 2>/dev/null if [ $? -ne 0 ]; then @@ -396,20 +390,18 @@ cd $OLDPWD if [ "$CLEANCACHE" = "1" ]; then + # fix flyspray feature request #5223 if [ -n "$SRCDEST" ]; then msg "Cleaning up source files from the cache." - rm -rf /var/cache/pacman/src/* - rm -rf "${SRCDEST}"/* - exit 0 - else - if [ "$EUID" = "0" -a "$INFAKEROOT" != "1" ]; then - msg "Cleaning up source files from the cache." - rm -rf /var/cache/pacman/src/* - exit 0 - else - error "You must be root to clean the cache." + rm -rf "$SRCDEST"/* + if [ $? -ne 0 ]; then + error "Problem removing files; you may not have correct permissions in $SRCDEST" exit 1 + else + exit 0 fi + else + error "Source destination must be defined in makepkg.conf." fi fi @@ -452,7 +444,7 @@ error "pkgrel is not allowed to contain hyphens." exit 1 fi -if ! in_array $CARCH ${arch[@]}; then +if [ ! "in_array $CARCH ${arch[@]}" ]; then error "$pkgname is not available for the '$CARCH' architecture." plain "Note that many packages may need a line added to their $BUILDSCRIPT" plain "such as arch=('$CARCH')." @@ -504,7 +496,7 @@ msg "Making package: $pkgname $pkgver-$pkgrel ($(date))" unset deplist makedeplist -# fix flyspray bug #5973 - Dan McGee <dpmcgee(a)gmail.com> +# fix flyspray bug #5973 if [ $(type -p pacman) -a "$NODEPS" = "0" -a "$GENINTEG" = "0" -a "$NOBUILD" = "0" ]; then msg "Checking Runtime Dependencies..." deplist=$(checkdeps ${depends[@]}) @@ -524,7 +516,9 @@ if [ $? -gt 0 ]; then exit 1 fi -elif [ "$NODEPS" = "1" -o "$GENINTEG" = "1" -o "$NOBUILD" = "1" ]; then +elif [ "$GENINTEG" = "1" -o "$NOBUILD" = "1" ]; then + msg "skipping dependency checks." +elif [ "$NODEPS" = "1" ]; then warning "skipping dependency checks." else warning "pacman was not found in PATH. skipping dependency checks." @@ -537,16 +531,13 @@ mkdir -p src cd $startdir/src for netfile in ${source[@]}; do - file=$(strip_url $netfile) + file=$(strip_url "$netfile") if [ -f "../$file" ]; then - msg " Found $file in build dir" + msg2 "Found $file in build dir" cp "../$file" . elif [ -f "$SRCDEST/$file" ]; then - msg " Using cached copy of $file" - cp "$SRCDEST$file" . - elif [ -f "/var/cache/pacman/src/$file" ]; then - msg " Using cached copy of $file" - cp "/var/cache/pacman/src/$file" . + msg2 "Using cached copy of $file" + cp "$SRCDEST/$file" . else # check for a download utility if [ -z "$FTPAGENT" ]; then @@ -566,33 +557,28 @@ msg "Aborting..." exit 1 fi - msg " Downloading $file" + msg2 "Downloading $file" $FTPAGENT "$netfile" 2>&1 # fix flyspray bug #3289 ftpret=$? if [ $ftpret -gt 0 ]; then error "Failure while downloading $file" msg "Aborting..." - rm "$file" - exit 1 + #rm "$file" exit 1 fi - if [ "$EUID" = "0" -a "$INFAKEROOT" != "1" ]; then - mkdir -p /var/cache/pacman/src && cp $file /var/cache/pacman/src - if [ -n "$SRCDEST" ]; then - mkdir -p $SRCDEST && cp "$file" $SRCDEST - elif [ "$EUID" = "0" -a "$INFAKEROOT" != "1" ]; then - mkdir -p /var/cache/pacman/src && cp "$file" /var/cache/pacman/src - else + if [ -n "$SRCDEST" ]; then + mkdir -p $SRCDEST && cp "$file" $SRCDEST + if [ $? -ne 0 ]; then + warning "You do not have correct permissions to cache source in $SRCDEST" cp "$file" .. - fi + else + cp "$file" .. fi fi done - - -if [ "$NOEXTRACT" = "1" ]; then +if [ "$NOEXTRACT" = "1" -o "$REPKG" = "1" ]; then warning "Skipping source integrity checks -- using existing src/ tree" else for integ in ${INTEGRITY_CHECK[@]}; do @@ -635,7 +621,7 @@ fi done plain "" - #Validate integrity checks + #Validate integrity checks else integrity_sums=($(eval echo \${${integrity_name}s[@]})) @@ -645,6 +631,7 @@ idx=0 for netfile in "${source[@]}"; do file=$(strip_url "$netfile") + #TODO is there a reason this is echo and not msg/msg2? echo -n " $file ... " >&2 echo "${integrity_sums[$idx]} $file" | $integrity_name -c - >/dev/null 2>&1 if [ $? -ne 0 ]; then @@ -671,8 +658,8 @@ fi #Extract sources -if [ "$NOEXTRACT" = "1" ]; then - warning "Skipping source extraction -- using existing src/ tree" +if [ "$NOEXTRACT" = "1" -o "$REPKG" = "1" ]; then + warning "Skipping source extraction -- using existing src/ tree" else msg "Extracting Sources..." for netfile in "${source[@]}"; do @@ -699,7 +686,7 @@ cmd="bunzip2 -f $file" ;; esac if [ "$cmd" != "" ]; then - msg " $cmd" + msg2 "$cmd" $cmd if [ $? -ne 0 ]; then # unzip will return a 1 as a warning, it is not an error @@ -713,71 +700,73 @@ done fi -if [ "$EUID" = "0" ]; then - # chown all source files to root.root - chown -R root.root $startdir/src -fi - -# check for existing pkg directory -if [ -d $startdir/pkg ]; then - msg "Removing existing pkg/ directory..." - rm -rf $startdir/pkg -fi -mkdir -p $startdir/pkg - -if [ "$NOBUILD" = "1" ]; then - msg "Sources are ready." - exit 0 -fi - -# use distcc if requested -if [ "$USE_DISTCC" = "y" ]; then - [ -d /usr/lib/distcc/bin ] && export PATH=/usr/lib/distcc/bin:$PATH -fi - -# use ccache if it's available -if [ "$USE_CCACHE" = "1" ]; then - [ -d /usr/lib/ccache/bin ] && export PATH=/usr/lib/ccache/bin:$PATH -fi - -# build -msg "Starting build()..." +if [ "$REPKG" = "0" ]; then + if [ "$EUID" = "0" ]; then + # chown all source files to root.root + chown -R root.root $startdir/src + fi -#check for "exit on syntax error" shell option -echo $SHELLOPTS | grep errexit 2>&1 >/dev/null -set_e=$? + # check for existing pkg directory + if [ -d $startdir/pkg ]; then + msg "Removing existing pkg/ directory..." + rm -rf $startdir/pkg + fi + mkdir -p $startdir/pkg -if [ "$LOGGING" = "1" ]; then - BUILDLOG="${startdir}/${pkgname}-${pkgver}-${pkgrel}-${CARCH}.log" - if [ -f "$BUILDLOG" ]; then - i=1 - while true; do - if [ -f "$BUILDLOG.$i" ]; then - i=$(($i +1)) - else - break - fi - done - mv "$BUILDLOG" "$BUILDLOG.$i" + if [ "$NOBUILD" = "1" ]; then + msg "Sources are ready." + exit 0 fi - #use 'errexit' to bail on syntax error - [ $set_e -eq 1 ] && set -e - build 2>&1 | tee "$BUILDLOG" - [ $set_e -eq 1 ] && set +e + # use distcc if requested + if [ "$USE_DISTCC" = "y" ]; then + [ -d /usr/lib/distcc/bin ] && export PATH=/usr/lib/distcc/bin:$PATH + fi + + # use ccache if it's available + if [ "x$USE_CCACHE" = "1" ]; then + [ -d /usr/lib/ccache/bin ] && export PATH=/usr/lib/ccache/bin:$PATH + fi + + # build + msg "Starting build()..." + + #check for "exit on syntax error" shell option + echo $SHELLOPTS | grep errexit 2>&1 >/dev/null + set_e=$? + + if [ "x$LOGGING" = "1" ]; then + BUILDLOG="${startdir}/${pkgname}-${pkgver}-${pkgrel}-${CARCH}.log" + if [ -f "$BUILDLOG" ]; then + i=1 + while true; do + if [ -f "$BUILDLOG.$i" ]; then + i=$( ($i +1)) + else + break + fi + done + mv "$BUILDLOG" "$BUILDLOG.$i" + fi - if [ ${PIPESTATUS[0]} -gt 0 ]; then - error "Build Failed. Aborting..." - exit 2 - fi -else - #use 'errexit' to bail on syntax error - [ $set_e -eq 1 ] && set -e - build 2>&1 - [ $set_e -eq 1 ] && set +e - if [ $? -gt 0 ]; then - error "Build Failed. Aborting..." - exit 2 + #use 'errexit' to bail on syntax error + [ $set_e -eq 1 ] && set -e + build 2>&1 | tee "$BUILDLOG" + [ $set_e -eq 1 ] && set +e + + if [ ${PIPESTATUS[0]} -gt 0 ]; then + error "Build Failed. Aborting..." + exit 2 + fi + else + #use 'errexit' to bail on syntax error + [ $set_e -eq 1 ] && set -e + build 2>&1 + [ $set_e -eq 1 ] && set +e + if [ $? -gt 0 ]; then + error "Build Failed. Aborting..." + exit 2 + fi fi fi @@ -788,6 +777,8 @@ rm -rf pkg/usr/info pkg/usr/share/info rm -rf pkg/usr/doc pkg/usr/share/doc rm -rf pkg/{usr,opt/gnome}/share/gtk-doc + # fix flyspray bug #5021 + #rm -rf ${DOC_DIRS[@]} fi # move /usr/share/man files to /usr/man

4 11

[pacman-dev] NoUpgrade issues
by Aaron Griffin 21 Dec '06

21 Dec '06

First off, thanks to Thomas and Travis for helping me out with this one. There is an issue with pacman and the NoUpgrade removal. Let me try to breeze over it. Files marked in the backup=() array have saved MD5 sums in pacman's db. Theses are used for comparison when upgrading. It compares the 'local' md5 sum (on the file system), the 'new' md5 sum, from the package, and the 'original' md5 sum from the database. Here is the problem: in BOTH pacman 2 and 3, when a package is marked NoUpgrade, pacman stores the 'local' md5 sum in the database instead of the 'original' md5 sums. I can see this _possibly_ being intentional, but here's where the problem occurs. When you remove NoUpgrade, the 'original' and 'local' md5 sums are the same. The comparission says "oh this file hasn't been touched since it was installed, we can overwrite it" and kills your configs. So, the proposed solution, which is the easiest: Fix pacman3 to store the correct md5 sum even if it is a NoUpgrade file. Then, we leave the NoUpgrade lines in pacman.conf and wait one release (or more) to remove them. Removing them will happen after each NoUpgrade package has been upgraded (or artificially version bumped). Those packages: initscripts module-init-tools pacman grub lilo filesystem ...and one or two others I forget... Does anyone have an issue with this solution?

3 3

[pacman-dev] [PATCH] fix automake conditional for java bindings
by Jürgen Hötzel 21 Dec '06

21 Dec '06

Hi, I work on Lua bindings and found this problem. Patch enclosed: Don't build java bindings when swig is NOT installed. Jürgen

2 1

[pacman-dev] My thoughts for development
by Dan McGee 21 Dec '06

21 Dec '06

Since I've been spending quite a bit of time looking at the code lately, I've been making a TODO list of things worth looking into. I thought I'd put it up on the list for input. makepkg issues: 1. We went through and added some of the undocumented options to the makepkg.conf file, but I haven't done a whole lot of testing on this to see if it works correctly. I also don't know if all the checking is correct, and I see Aaron put some x's in front of these checks. Has anyone else? They also aren't documented in the man page if anyone wants to step up to the plate... 2. Remove use of multiple `` or $() calls to the same function. I have a local update of calling $(id -u) once and storing it to a variable for later use, I can publish a patch sometime. 3. I posted a patch for #5223 (source cache) and it has been patched. However, it is now a triple conditional with three different possible locations for package source (current dir, $SRCDEST, and /var/cache/pacman/src). Is the last one necessary since it is highly recommended to not build as root? Should we deprecate this path? Looking even further ahead, should we make fakeroot a requirement for running makepkg? 4. Aaron, you brought up that `` is the older standard and $() should be used. It is a pretty simple find that will allow you to change these for consistency, I've done it on my working copy. 5. Already mentioned, but a modeline in makepkg. 6. Overabundance of options in makepkg- I was trying to find a new letter for a '--repackage' option and we are pretty much out of lowercase letters. At the moment we have two basic types of options: things that can be done in the pkgbuild options line, and things that directly affect the way makepkg works (geninteg, skip integs, location of PKGBUILD). Are the first type truly necessary? It would make makepkg a bit more streamlined if the only options were those that can't be done somewhere else. So i propose the '--nostrip', '--keepdocs', and any other option like this be removed. 7. Integrity checking- currently generating integrity checks will only generate those from the first one in the INTEGRITY_CHECK array. This seems to make sense as long as the array in makepkg.conf is ordered the way we want it to be. When checking, we do get warning messages for other types that may be in the array but are not in the PKGBUILD. I just wanted to confirm this is the behavior you are looking for. Other issues: 1. Move the PKGBUILD man page to section 5 2. How are you guys doing side-by-side builds with the old pacman? At least for me, the current build system seems horribly complicated with shell scripts, configure, automaking and such. Is there a good reason for this? 3. Everyone talks about pacman slowness (even though it isn't really that bad), and looking though the mailing list archives I believe Aaron implemented a form of lazy caching (not loading the descriptions when unnecessary). Anything else on this front? I thought of something and I may be way off track, but what if when generating a repository we make (in addition to the flat file format) a single file with lines consisting of pkgname-pkgver:description? This file could be loaded and searched a lot quicker than flat files, and I think the operation people complain about being slow the most is a search. Wow, long list, but maybe it will provoke other great ideas. Feel free to respond to any or all of it. -Dan

4 10