- Pacman-dev - lists.archlinux.org

[pacman-dev] [GIT] pacman branch, master now at v3.0.0-235-gcacf968
by Dan McGee 17 Jun '07

17 Jun '07

Hello, This is an automated email from the git hooks/update script, it was generated because a ref change was pushed to the repository. Updating branch, master, via cacf968970bbc63e8eb519e40194e6c083ba6a06 (commit) via 2ab320a3885226ce28da0f2f2bc84b5f2d0f0539 (commit) via 520841ee8a82460e0abb2a9a44c857556604cf56 (commit) via a486fec3c48d57d48c7d871161c1510be10e0924 (commit) from 7bd4486ebd15ba9cca0ee3e6a2c2ddc88ff104dc (commit) - Log ----------------------------------------------------------------- commit cacf968970bbc63e8eb519e40194e6c083ba6a06 Author: Dan McGee <dan(a)archlinux.org> Date: Thu Jun 14 11:12:21 2007 -0400 Update Doxyfile and add a few missing parts of documentation Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 2ab320a3885226ce28da0f2f2bc84b5f2d0f0539 Author: Dan McGee <dan(a)archlinux.org> Date: Wed Jun 13 11:33:45 2007 -0400 Remove some use of goto in the pacman frontend Refactor the goto cleanup; statements out of add.c and remove.c, and move what the cleanup: label did into a function. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 520841ee8a82460e0abb2a9a44c857556604cf56 Author: Dan McGee <dan(a)archlinux.org> Date: Wed Jun 13 10:36:33 2007 -0400 Quick updates to the bash completion script Signed-off-by: Dan McGee <dan(a)archlinux.org> commit a486fec3c48d57d48c7d871161c1510be10e0924 Author: Chantry Xavier <shiningxc(a)gmail.com> Date: Sun Jun 17 18:57:15 2007 +0200 FS7338: Fix conflicts typo totally this time. The last typo fix for conflicts wasn't complete, it only fixed one occurence of the conflicts variable, instead of the two : http://www.archlinux.org/pipermail/pacman-dev/2007-May/008222.html This caused all conflicts statement to be ignored, so all databases will need to be rebuilt with a fixed repo-add. Signed-off-by: Chantry Xavier <shiningxc(a)gmail.com> ----------------------------------------------------------------------- Diffstat: contrib/bash_completion | 25 +++++++++---------------- lib/libalpm/Doxyfile | 16 +++++++++++++++- lib/libalpm/db.c | 11 +++++++++++ scripts/repo-add.in | 2 +- src/pacman/add.c | 37 ++++++++++++++++++++++--------------- src/pacman/remove.c | 41 ++++++++++++++++++++++++++--------------- 6 files changed, 84 insertions(+), 48 deletions(-) hooks/update --- Git Source Code Management System hooks/update refs/heads/master \ 7bd4486ebd15ba9cca0ee3e6a2c2ddc88ff104dc \ cacf968970bbc63e8eb519e40194e6c083ba6a06

1 0

[pacman-dev] How does pacman interact with a package's INSTALL script
by lordbad 16 Jun '07

16 Jun '07

In the documentation I have read the following: There are also install files. This PKGBUILD specifies 'foo.install' as the package's install file. Here is an example install file: post_install() { /bin/true } post_upgrade() { /bin/true } pre_remove() { /bin/true } op=$1 shift $op "$@" Here are the function explainations : * post_install : this script is run right after files are installed, it takes one argument : o the package version * post_upgrade : this script is run after all files have been upgraded, it takes two arguments : o the new package version o the old package version * pre_remove : this script is run right before files are removed (stop a daemon for example) and takes one argument : o the package version The three lines at the bottom are needed in every install file so that they run properly. But I do not understand how does pacman handle these scripts. Does it source the scripts first and then call the shell functions defined in them with some C function like "system" for example? But if it does so why are the last 3 lines needed - I see little if any need for them. So if anyone can give me some hints on the subject and maybe even mention which of pacman's source files implements this functionality I'd be very grateful. Thanks in advance for your help. Best Regards, Bozhidar _____________________ http://www.e-card.bg/

1 0

Re: [pacman-dev] [arch] Plans for switching to new package naming format (name-ver-rel -> name-ver-rel-arch)
by Dan McGee 15 Jun '07

15 Jun '07

On 6/13/07, Aaron Griffin <aaronmgriffin(a)gmail.com> wrote: > On 6/13/07, Johan Grahn <johan(a)cgeek.net> wrote: > > On Wed, 2007-06-13 at 22:43 +0100, Andrew Fyfe wrote: > > > Are there any plans (if so when) to switch the official repositories > > > over to the new package naming format? > > > > > > Current/Old: name-ver-rel.pkg.tar.gz > > > New: name-ver-rel-arch.pkg.tar.gz > > > > Why do you wanna do this? I'm just curios... > > > > Is it not better to just have separate repos because they are compiled > > differently and better to maintain. > > > > I'm not saying that is it a bad idea, I just wanna know how you think > > No body "wants" to do it - it's already done, and part of pacman 3.0. > The rationalization is simple: the majority of the time you don't > touch real package files... it's pacman -S, done. There is really > nothing lost from adding the suffix - what is gained is clarity in > looking at the filename. > > To respond to Andrew: there actually is no real "plan" at the > moment... but we want to make sure pacman2 users have sufficient time > to make the switch before we kill their packages. This drifts into something that should be on the pacman ML, so I'm moving it there. Is there any reason for pacman to enforce a naming convention at all when it comes to the packages? We should do any information gathering from the .PKGINFO file in the package itself rather than the name, shouldn't we? I also might be way off here- maybe pacman doesn't care. If so, let me know- I haven't dug into the code yet on this issue. -Dan

5 4

[pacman-dev] [patch] patch IV rewritten for git
by Nagy Gabor 13 Jun '07

13 Jun '07

Hi! Some comments: 1. This is a simple topological sort algorithm, based on the "depth first search" algorithm, for more info plz visit: http://en.wikipedia.org/wiki/Topological_sorting (An alternative algorithm for...) Usually this is done by a recursive way but I not prefer that, because this version is more expressive to me (however this looks more difficult, because we need some extra variables to precisely store the current state). 2. As I noted in the patch (TODO), the most CPU-cost part of the algorithm is to build the graph, so this should somehow combined with depcheck (because that function does this too...) So, here is the patch: ---------------------- diff --git a/lib/libalpm/Makefile.am b/lib/libalpm/Makefile.am index 29a2f02..afbfed0 100644 --- a/lib/libalpm/Makefile.am +++ b/lib/libalpm/Makefile.am @@ -43,7 +43,7 @@ libalpm_la_SOURCES = \ versioncmp.h versioncmp.c libalpm_la_LDFLAGS = -no-undefined -version-info $(LIB_VERSION_INFO) -libalpm_la_LIBADD = -larchive -ldownload -lm +libalpm_la_LIBADD = -larchive -ldownload if HAS_DOXYGEN all: doxygen.in diff --git a/lib/libalpm/alpm.h b/lib/libalpm/alpm.h index d4c584f..e6c905f 100644 --- a/lib/libalpm/alpm.h +++ b/lib/libalpm/alpm.h @@ -50,6 +50,7 @@ typedef struct __pmsyncpkg_t pmsyncpkg_t; typedef struct __pmdepend_t pmdepend_t; typedef struct __pmdepmissing_t pmdepmissing_t; typedef struct __pmconflict_t pmconflict_t; +typedef struct __pmgraph_t pmgraph_t; /* * Library diff --git a/lib/libalpm/deps.c b/lib/libalpm/deps.c index b6f8cc8..ceee5c1 100644 --- a/lib/libalpm/deps.c +++ b/lib/libalpm/deps.c @@ -29,7 +29,6 @@ #ifdef __sun__ #include <strings.h> #endif -#include <math.h> /* libalpm */ #include "deps.h" @@ -46,6 +45,27 @@ extern pmhandle_t *handle; +pmgraph_t *_alpm_graph_new(void) +{ + pmgraph_t *graph = NULL; + + graph = (pmgraph_t *)malloc(sizeof(pmgraph_t)); + if(graph) { + graph->state = 0; + graph->data = NULL; + graph->father = NULL; + graph->sons = NULL; + graph->sonptr = NULL; + } + return(graph); +} + +void _alpm_graph_free(pmgraph_t *graph) +{ + alpm_list_free(graph->sons); + free(graph); +} + pmdepmissing_t *_alpm_depmiss_new(const char *target, pmdeptype_t type, pmdepmod_t depmod, const char *depname, const char *depversion) @@ -108,11 +128,10 @@ int _alpm_depmiss_isin(pmdepmissing_t *needle, alpm_list_t *haystack) alpm_list_t *_alpm_sortbydeps(alpm_list_t *targets, pmtranstype_t mode) { alpm_list_t *newtargs = NULL; - alpm_list_t *i, *j, *k, *l; - int change = 1; - int numscans = 0; - int numtargs = 0; - int maxscans; + alpm_list_t *i, *j, *k; + alpm_list_t *vertices = NULL; + alpm_list_t *vptr; + pmgraph_t *vertex; ALPM_LOG_FUNC; @@ -120,65 +139,64 @@ alpm_list_t *_alpm_sortbydeps(alpm_list_t *targets, pmtranstype_t mode) return(NULL); } + _alpm_log(PM_LOG_DEBUG, _("started sorting dependencies")); + + /* We create the vertices */ for(i = targets; i; i = i->next) { - newtargs = alpm_list_add(newtargs, i->data); - numtargs++; + pmgraph_t *vertex = _alpm_graph_new(); + vertex->data = (void *)i->data; + vertices = alpm_list_add(vertices, vertex); } - maxscans = (int)sqrt(numtargs); + /* We compute the edges */ + for(i = vertices; i; i = i->next) { + pmgraph_t *vertex_i = i->data; + pmpkg_t *p_i = vertex_i->data; + //TODO: this should be somehow combined with _alpm_checkdeps + for(j = vertices; j; j = j->next) { + pmgraph_t *vertex_j = j->data; + pmpkg_t *p_j = vertex_j->data; + int son=0; + for(k = alpm_pkg_get_depends(p_i); k && !son; k = k->next) { + pmdepend_t *depend = alpm_splitdep(k->data); + son = alpm_depcmp(p_j, depend); + free(depend); + } + if(son) vertex_i->sons=alpm_list_add(vertex_i->sons, vertex_j); + } + vertex_i->sonptr=vertex_i->sons; + } - _alpm_log(PM_LOG_DEBUG, _("started sorting dependencies")); - while(change) { - alpm_list_t *tmptargs = NULL; - change = 0; - if(numscans > maxscans) { - _alpm_log(PM_LOG_DEBUG, _("possible dependency cycle detected")); - continue; + vptr = vertices; + vertex = vertices->data; + while(vptr) { + vertex->state = -1; // we touched this vertex + int found = 0; + while(vertex->sonptr && !found) { + pmgraph_t *nextson = (vertex->sonptr)->data; + vertex->sonptr = (vertex->sonptr)->next; + if (nextson->state == 0) { + found = 1; + nextson->father = vertex; + vertex = nextson; + } + else if(nextson->state == -1) { + _alpm_log(PM_LOG_WARNING, _("dependency cycle detected\n")); + } } - numscans++; - /* run thru targets, moving up packages as necessary */ - for(i = newtargs; i; i = i->next) { - pmpkg_t *p = i->data; - _alpm_log(PM_LOG_DEBUG, " sorting %s", alpm_pkg_get_name(p)); - for(j = alpm_pkg_get_depends(p); j; j = j->next) { - pmdepend_t *depend = alpm_splitdep(j->data); - pmpkg_t *q = NULL; - if(depend == NULL) { - continue; - } - /* look for depend->name -- if it's farther down in the list, then - * move it up above p - */ - for(k = i->next; k; k = k->next) { - q = k->data; - const char *qname = alpm_pkg_get_name(q); - if(!strcmp(depend->name, qname)) { - if(!_alpm_pkg_find(qname, tmptargs)) { - change = 1; - tmptargs = alpm_list_add(tmptargs, q); - } - break; - } - for(l = alpm_pkg_get_provides(q); l; l = l->next) { - const char *provname = l->data; - if(!strcmp(depend->name, provname)) { - if(!_alpm_pkg_find(qname, tmptargs)) { - change = 1; - tmptargs = alpm_list_add(tmptargs, q); - } - break; - } - } + if(!found) { + newtargs = alpm_list_add(newtargs, vertex->data); + vertex->state = 1; //we leave this vertex + vertex = vertex->father; + if(!vertex) { + while(vptr = vptr->next) { + vertex = vptr->data; + if (vertex->state == 0) break; } - free(depend); - } - if(!_alpm_pkg_find(alpm_pkg_get_name(p), tmptargs)) { - tmptargs = alpm_list_add(tmptargs, p); } } - alpm_list_free(newtargs); - newtargs = tmptargs; } + _alpm_log(PM_LOG_DEBUG, _("sorting dependencies finished")); if(mode == PM_TRANS_TYPE_REMOVE) { @@ -189,6 +207,8 @@ alpm_list_t *_alpm_sortbydeps(alpm_list_t *targets, pmtranstype_t mode) newtargs = tmptargs; } + alpm_list_free_inner(vertices, _alpm_graph_free); + return(newtargs); } diff --git a/lib/libalpm/deps.h b/lib/libalpm/deps.h index 8f3a9b9..56028a4 100644 --- a/lib/libalpm/deps.h +++ b/lib/libalpm/deps.h @@ -42,6 +42,18 @@ struct __pmdepmissing_t { pmdepend_t depend; }; +/* Graphs */ +struct __pmgraph_t { + int state; //0: untouched, -1: entered, other: leaving time + void *data; + struct __pmgraph_t *father; //where did we come from? + alpm_list_t *sons; + alpm_list_t *sonptr; // points to a son in the sons list +}; + +pmgraph_t *_alpm_graph_new(void); +void _alpm_graph_free(pmgraph_t *graph); + pmdepmissing_t *_alpm_depmiss_new(const char *target, pmdeptype_t type, pmdepmod_t depmod, const char *depname, const char *depversion); diff --git a/src/pacman/Makefile.am b/src/pacman/Makefile.am index 6a7f4e4..1ab5916 100644 --- a/src/pacman/Makefile.am +++ b/src/pacman/Makefile.am @@ -26,11 +26,11 @@ pacman_SOURCES = \ util.h util.c pacman_LDADD = $(top_builddir)/lib/libalpm/.libs/libalpm.la \ - -ldownload + -ldownload -lm pacman_static_SOURCES = $(pacman_SOURCES) pacman_static_LDFLAGS = $(LDFLAGS) -all-static pacman_static_LDADD = $(top_builddir)/lib/libalpm/.libs/libalpm.la \ - -ldownload + -ldownload -lm # vim:set ts=2 sw=2 noet:

4 6

[pacman-dev] [GIT] pacman branch, master now at v3.0.0-231-g7bd4486
by Dan McGee 13 Jun '07

13 Jun '07

Hello, This is an automated email from the git hooks/update script, it was generated because a ref change was pushed to the repository. Updating branch, master, via 7bd4486ebd15ba9cca0ee3e6a2c2ddc88ff104dc (commit) from af813a8badd71c591a8c29d9bad1710ae3b2e945 (commit) - Log ----------------------------------------------------------------- commit 7bd4486ebd15ba9cca0ee3e6a2c2ddc88ff104dc Author: Andrew Fyfe <andrew(a)neptune-one.net> Date: Wed Jun 13 10:48:22 2007 +0100 scripts/makepkg.in: Forgot the fakeroot switch when changing to GNU getopt. Signed-off-by: Andrew Fyfe <andrew(a)neptune-one.net> ----------------------------------------------------------------------- Diffstat: scripts/makepkg.in | 5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-) hooks/update --- Git Source Code Management System hooks/update refs/heads/master \ af813a8badd71c591a8c29d9bad1710ae3b2e945 \ 7bd4486ebd15ba9cca0ee3e6a2c2ddc88ff104dc

1 0

[pacman-dev] Add fakeroot version to .PKGINFO header
by Andrew Fyfe 13 Jun '07

13 Jun '07

I've created a small patch (on my ready_to_pull branch) that adds the version of fakeroot to the header of .PKGINFO: # Generated by makepkg 3.1.0-dev # using fakeroot version 1.7.1 This is purely for debugging. I've currently got a couple of packages with wrong permissions which I think are a result of the coreutils/fakeroot bug from a couple of months ago (http://archlinux.org/news/310/) Andrew http://repo.or.cz/w/pacman.git?a=shortlog;h=ready_to_pull

2 2

[pacman-dev] Problems with alpm
by Georg Grabler 13 Jun '07

13 Jun '07

Hey guys, I'm once again having problems with the code in alpm. I can't seem to find out how to correctly register databases. I've the bindings basically ready, and need to do the implementation now (pyalpm). So, the following problem occured: I set the database path to /var/lib/pacman (or /var/lib/pacman/community, tried both). It seems as if alpm_option_set_dbpath can't cause errors (void return), so i ignored it. alpm_trans_init runs without returning -1, so i expect it to be succeeded. On alpm_trans_addtarget, i finally get the -1, with pm_errno 11, meaning PM_ERR_DB_NULL. I don't know (lacking docs) what this means, and why this can occur. Do you have any hints for me? Some codelines: cdef pyalpm_config c_config c_config = pyalpm_config() config = c_config alpm_initialize() def addPackage(in_pname) db = config.getDatabasePath() print db # output is OK pyalpm_db_register('community') if (alpm_trans_init(PM_TRANS_TYPE_ADD, <pmtransflag_t((<int>PM_TRANS_FLAG_CASCADE) | (<int>PM_TRANS_FLAG_RECURSE)), NULL, NULL, NULL) == -1): os.remove("/tmp/pacman.lck") print pm_errno return -1 if (alpm_trans_addtarget(in_name) == -1): os.remove("/tmp/pacman.lck") print pm_errno return -2 # here it exits ###test.py### import pyalpm pyalpm.config.setDatabasePath('/var/lib/pacman') print pyalpm.addPackage('yaz') ###output### 11 -2 Seems as if i'm missing something -- Yours, Georg (STiAT)

3 2

[pacman-dev] [GIT] pacman branch, master now at v3.0.0-230-gaf813a8
by Dan McGee 13 Jun '07

13 Jun '07

Hello, This is an automated email from the git hooks/update script, it was generated because a ref change was pushed to the repository. Updating branch, master, via af813a8badd71c591a8c29d9bad1710ae3b2e945 (commit) via 620a811f9136feb332a0f196a6f9a8818368c908 (commit) via 5e774d72b0ac7777aada8271e2162d764d4f0db4 (commit) via caf6c326ec9a376ffb794051cd09c40abc135fda (commit) via 83245aa58ef1eefcc3498a82923e55bc799ca361 (commit) via 54b71f0427e87e6d525423df06f8a06f2b71c518 (commit) from e2a75a8257e3402c32a208709f044373ef97caf7 (commit) - Log ----------------------------------------------------------------- commit af813a8badd71c591a8c29d9bad1710ae3b2e945 Author: Dan McGee <dan(a)archlinux.org> Date: Wed Jun 13 00:55:23 2007 -0400 Add ARM architecture to configure.ac Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 620a811f9136feb332a0f196a6f9a8818368c908 Author: Dan McGee <dan(a)archlinux.org> Date: Wed Jun 13 00:34:43 2007 -0400 Clean up pacman/sync.c and organize it a bit Re-add some checks for a valid target list. In addition, move variable declarations to the most local scope possible, which should help in finding errors that were previously uncaught. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 5e774d72b0ac7777aada8271e2162d764d4f0db4 Author: Dan McGee <dan(a)archlinux.org> Date: Wed Jun 13 00:31:01 2007 -0400 Remove some more diskspace checking holdover stuff Signed-off-by: Dan McGee <dan(a)archlinux.org> commit caf6c326ec9a376ffb794051cd09c40abc135fda Author: Dan McGee <dan(a)archlinux.org> Date: Tue Jun 12 23:55:14 2007 -0400 Functionalize query.c and pacman_query Move a lot of the operations that pacman query performs into their own functions for ease of following the flow and for possible later merges with their sync counterparts. We need to think less in terms of operations and more in term of overall concepts. Also (re)introduced checks for the target list on a -Q operation, ensuring that we only call certain query operations when it is sane to do so. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 83245aa58ef1eefcc3498a82923e55bc799ca361 Author: Andrew Fyfe <andrew(a)neptune-one.net> Date: Tue Jun 12 23:30:56 2007 +0100 scripts/makepkg.in: Added fakeroot version to .PKGINFO header. Signed-off-by: Andrew Fyfe <andrew(a)neptune-one.net> commit 54b71f0427e87e6d525423df06f8a06f2b71c518 Author: Andrew Fyfe <andrew(a)neptune-one.net> Date: Tue Jun 5 11:58:25 2007 +0100 scripts/makepkg.in: Switch to GNU getopt. Signed-off-by: Andrew Fyfe <andrew(a)neptune-one.net> ----------------------------------------------------------------------- Diffstat: configure.ac | 5 + lib/libalpm/alpm.h | 1 - lib/libalpm/error.c | 2 - scripts/makepkg.in | 129 +++++++++--------------- src/pacman/add.c | 14 --- src/pacman/query.c | 280 +++++++++++++++++++++++++++++++-------------------- src/pacman/sync.c | 61 ++++++++---- 7 files changed, 263 insertions(+), 229 deletions(-) hooks/update --- Git Source Code Management System hooks/update refs/heads/master \ e2a75a8257e3402c32a208709f044373ef97caf7 \ af813a8badd71c591a8c29d9bad1710ae3b2e945

1 0

[pacman-dev] [GIT] pacman branch, master now at v3.0.0-224-ge2a75a8
by Dan McGee 12 Jun '07

12 Jun '07

Hello, This is an automated email from the git hooks/update script, it was generated because a ref change was pushed to the repository. Updating branch, master, via e2a75a8257e3402c32a208709f044373ef97caf7 (commit) via 892e6a7fe926bccc8e3d18d4a03973a20d5208a4 (commit) from c1b6a6d616727280ab412ea06d15853b8c1f386a (commit) - Log ----------------------------------------------------------------- commit e2a75a8257e3402c32a208709f044373ef97caf7 Author: Nagy Gabor <ngaba(a)petra.hos.u-szeged.hu> Date: Tue Jun 12 01:09:37 2007 +0200 Add pactest requiredby004. Upgrade a package that no longer needs a depend. Reference: http://www.archlinux.org/pipermail/pacman-dev/2007-June/008503.html Signed-off-by: Chantry Xavier <shiningxc(a)gmail.com> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 892e6a7fe926bccc8e3d18d4a03973a20d5208a4 Author: Dan McGee <dan(a)archlinux.org> Date: Tue Jun 12 16:09:46 2007 -0400 pacman util.h: Add missing stdarg.h include Signed-off-by: Dan McGee <dan(a)archlinux.org> ----------------------------------------------------------------------- Diffstat: .../tests/{requiredby002.py => requiredby004.py} | 12 ++++++++---- src/pacman/util.h | 1 + 2 files changed, 9 insertions(+), 4 deletions(-) hooks/update --- Git Source Code Management System hooks/update refs/heads/master \ c1b6a6d616727280ab412ea06d15853b8c1f386a \ e2a75a8257e3402c32a208709f044373ef97caf7

1 0

[pacman-dev] UPGRADERM patch
by Xavier 12 Jun '07

12 Jun '07

About Nagy patch and comment here : http://www.archlinux.org/pipermail/pacman-dev/2007-April/008054.html http://www.archlinux.org/pipermail/pacman-dev/2007-April/008132.html I think adding a special UPGRADERM type for this pseudo remove transaction is better than the previous fix which just added this check for eliminating duplicated required by : if(!alpm_list_find_str(rqdby, pkgname)) { But this just makes the existence of this pseudo remove transaction more explicit. And as seen in add.c code, this should be fixed : 362 /* this is kinda odd. If the old package exists, at this point we make a 363 * NEW transaction, unrelated to handle->trans, and instantiate a "remove" 364 * with the type PM_TRANS_TYPE_UPGRADE. TODO: kill this weird behavior. */ I asked Dan about it on irc : 15:44 shining >> if there is a way to kill this weird behavior totally, then maybe there should be a bit more details about how to do, so we can reject Nagy patch ? 16:11 toofishes >> well i sure don't know it yet 16:12 toofishes >> i'm saying this- i hate the way alpm tries to have these "transactions" that aren't transactional in nature at all (meaning fully finish or never start) 16:12 toofishes >> alpm_trans_start() 16:12 toofishes >> alpm_install(pkg) 16:12 toofishes >> alpm_remove(pkg) 16:13 toofishes >> alpm_trans_end() (which checks to make sure transaction is possible) 16:13 toofishes >> alpm_trans_commit() (which carries out the install) So it looks like there is a lot of work to be done here, for fixing the problem correctly.

2 2