- Pacman-dev - lists.archlinux.org

[pacman-dev] [PATCH 1/2] fix memory leak in _alpm_sync_commit
by Jonathan Conder 18 May '10

18 May '10

Signed-off-by: Jonathan Conder <j(a)skurvy.no-ip.org> --- lib/libalpm/sync.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/lib/libalpm/sync.c b/lib/libalpm/sync.c index 6b625ed..d9f4c28 100644 --- a/lib/libalpm/sync.c +++ b/lib/libalpm/sync.c @@ -955,7 +955,8 @@ int _alpm_sync_commit(pmtrans_t *trans, pmdb_t *db_local, alpm_list_t **data) } EVENT(trans, PM_TRANS_EVT_INTEGRITY_DONE, NULL, NULL); if(trans->flags & PM_TRANS_FLAG_DOWNLOADONLY) { - return(0); + ret = 0; + goto error; } trans->state = STATE_COMMITING; -- 1.7.1

2 1

[pacman-dev] [GIT] The official pacman repository branch, master, updated. v3.3.3-189-gc1fc005
by dan＠archlinux.org 18 May '10

18 May '10

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "The official pacman repository". The branch, master has been updated via c1fc00508e337f20eb94e95978e193d23cfeaf10 (commit) via 0e7ba6bddd4b6c67cb6344ee3af4bacb00cda76b (commit) via eada558e12da1fd605d8c7401ec3c17a37fed9a9 (commit) via c6f0fc27eda31943e10e2a89e221796efb87dce9 (commit) via d44e5099339616c29797b88ed19fe27ea282c9ae (commit) via d32f6daa66a454ea41574667d44a2f6c36d4f77e (commit) via 5752e276fb585d07bd0f8e8b2a1f2827afdf2799 (commit) from a6ace987a913852af30bb5e3ae6bf145dc9c6157 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit c1fc00508e337f20eb94e95978e193d23cfeaf10 Author: Allan McRae <allan(a)archlinux.org> Date: Sun Oct 11 21:16:09 2009 +1000 makepkg: allow skipping integrity checks when making source package Extends the use of the --skipinteg option to creating a source package. Fixes FS#15984. Signed-off-by: Allan McRae <allan(a)archlinux.org> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 0e7ba6bddd4b6c67cb6344ee3af4bacb00cda76b Author: Nagy Gabor <ngaba(a)bibl.u-szeged.hu> Date: Mon May 17 21:41:51 2010 +0200 New pactest: unresolvable001.py To test the regression of commit eada558e12. Signed-off-by: Nagy Gabor <ngaba(a)bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit eada558e12da1fd605d8c7401ec3c17a37fed9a9 Author: Nagy Gabor <ngaba(a)bibl.u-szeged.hu> Date: Mon May 17 02:33:28 2010 +0200 Partial fix for the phonon/qt issue This patch fixes the phonon/qt issue, if all to-be-upgraded packages are explicit targets (ie. only not-yet-installed packages are pulled by resolvedeps). This condition covers the most common situations, for example it should hold with every -Su operation. After this patch sync405.py passes, but sync406.py doesn't. The work is inspired by the patch of Henning Garus, thanks for his work: http://mailman.archlinux.org/pipermail/pacman-dev/2010-February/010429.html (I moved the alpm_list_diff computation to sync.c in order to compute it only once.) Signed-off-by: Nagy Gabor <ngaba(a)bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit c6f0fc27eda31943e10e2a89e221796efb87dce9 Author: Nezmer <git(a)nezmer.info> Date: Mon May 17 10:52:38 2010 -0400 Resolve dependencies whenever --syncdeps is passed and --nodeps is not With this patch, dependencies will be resolved and not silently ignored when running: makepkg --nobuild --syncdeps makepkg --repackage --syncdeps Also, a warning is displayed when repackaging and dependencies are not being resolved. Thank you Allan for the feedback. Signed-off-by: Nezmer <git(a)nezmer.info> [Allan: Only warn when repackaging with a package function] Signed-off-by: Allan McRae <allan(a)archlinux.org> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit d44e5099339616c29797b88ed19fe27ea282c9ae Author: Xavier Chantry <chantry.xavier(a)gmail.com> Date: Mon Mar 15 23:46:44 2010 +0100 configure.ac : disable doc by default and check for asciidoc This is a complaint that has been reported many many times. By default, docs are enabled and there is no check for asciidoc, so anyone building from git will see their build fail. We cannot do a strict check for asciidoc because released source tarballs have man pages already built, and it should be possible to install them without having asciidoc. This patch attempts to improve the situation in two ways : 1) disable doc by default 2) print a warning if docs are enabled but asciidoc is not installed Signed-off-by: Xavier Chantry <chantry.xavier(a)gmail.com> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit d32f6daa66a454ea41574667d44a2f6c36d4f77e Author: Dieter Plaetinck <dieter(a)plaetinck.be> Date: Sat May 15 21:06:42 2010 +0200 fix for incorrect checking of return code, which causes syntax errors Signed-off-by: Allan McRae <allan(a)archlinux.org> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 5752e276fb585d07bd0f8e8b2a1f2827afdf2799 Author: Marc-A. Dahlhaus <mad(a)wol.de> Date: Tue May 18 14:32:02 2010 +0200 Allow to include a path containing wildcards Dan: line wrapping and man page touchup. Signed-off-by: Marc-A. Dahlhaus <mad(a)wol.de> Signed-off-by: Dan McGee <dan(a)archlinux.org> ----------------------------------------------------------------------- Summary of changes: configure.ac | 18 ++++++++++++------ doc/pacman.conf.5.txt | 3 ++- lib/libalpm/deps.c | 10 +++------- lib/libalpm/deps.h | 2 +- lib/libalpm/sync.c | 6 +++++- pactest/tests/sync405.py | 2 -- pactest/tests/unresolvable001.py | 21 +++++++++++++++++++++ scripts/makepkg.sh.in | 20 +++++++++++++------- src/pacman/pacman.c | 31 +++++++++++++++++++++++++++++-- 9 files changed, 86 insertions(+), 27 deletions(-) create mode 100644 pactest/tests/unresolvable001.py hooks/post-receive -- The official pacman repository

1 0

[pacman-dev] Release plans
by Allan McRae 18 May '10

18 May '10

Hi, Back on the 15th of March, Dan mentioned the idea of making a release in a month (http://mailman.archlinux.org/pipermail/pacman-dev/2010-March/010514.html) So, given a month has passed, lets start making plans so we can do that next month! Things that need done before release: 1) Nagy's -D patch should be included. I am not sure what the final status of that was. 2) Revert/fix the cross-build library linking patch as this forces some linking that it should not. (perhaps consider making linking with --as-needed the default when supported too?) 3) Fix "makepkg --source" issue with split packages and install/changelog files (FS#18831) Related bugs FS#18394 and FS#16004 can be fixed at the same time. I have most of a patch prepared. 4) Pull in other stuff from working branches - Nagy has another couple of minor patches - Xavier has a couple (I am not sure about asciidoc one...) - I have few patches (all makepkg ones are well tested, pacman -Qo "functional which" could do with a final review) 5) Tidy up bash_completion stuff? (FS#16630) 6) Look at Flyspray roadmap: http://bugs.archlinux.org/index.php?do=index&tasks=&project=3&due=69 (I just removed package signing to 4.0...). I think the only thing not covered above is FS#15657 (presence of new db version is not checked when <dbfile>.part exists) and that is fairly minor... 7) Any other patches waiting on the mailing list? Any other points/concerns people want to mention? Allan

9 34

[pacman-dev] makepkg integrity check patches
by Allan McRae 18 May '10

18 May '10

Here is some justification for the patches on my working branch related to integrity checking. Put in order from least controversial to most... 1) cd3910b makepkg -g: use checksums defined in the pkgbuild This is Xavier's patch. I am not sure there was ever issues with this one, but I think it got caught up with the other patches. This patch just makes it easier to maintain checksums in a PKGBUILD other than the default in makepkg.conf. e.g. if I have sha1sums in the PKGBUILD, "makepkg -g" will generate sha1sums even if generating md5sums is the default. To generate md5sums only, you just need to remove the old checksums from the PKGBUILD or add an empty md5sum array to the PKGBUILD. 2) cd1378d makepkg: rework --skipinteg This is very, very, VERY useful. I did not have makepkg-git on my new computer earlier this week and the current makepkg behaviour annoyed me A LOT. without patch: > /usr/bin/makepkg --skipinteg ... ==> Validating source files with md5sums... abs-2.3.4.1.tar.gz ... FAILED ==> ERROR: One or more files did not pass the validity check! with patch: > makepkg --skipinteg ... ==> WARNING: Skipping integrity checks. ==> Extracting Sources... This is particularly useful when testing out a patch that you need to repeatedly modify. You only need to update your checksums once it is working. I use this very frequently, but then again I do more packaging than most. 3) 5d911ae makepkg: allow skipping integrity checks when making source package And here is the fun one... "makepkg --source" currently requires checking all checksums. Using "-source --skipinteg" does not skip this, which in itself makes little sense to me. The argument that this stops people distributing packages with bad checksums is flawed. There is nothing stopping them doing that now. They just have to not use makepkg when creating the tarball, which could lead to even worse PKGBUILDs being distributed as none of makepkg's other checks would be performed. Admittedly, this patch will see little use. I used to use it in combination with #2 when checking that a modified patch compiled on both my i686 and x86_64 machines. Now that I build both architectures on one machine, this is of less use to me. I have also used it when commenting on a PKGBUILD that was sent to me with a query about the best way to implement something. There I modified some of the build() function to be clearer and sent it to the author saying I would use this approach but it is not tested. Note I could have used tar manually (in fact the tar file name would have been tab autocompleted so the command would probably be simpler to type), but then I would not get the other checks makepkg provides. Note that this patch does not make --skipinteg the default. In fact, given how long the options is, it is very unlikely that a user could every accidentally type it! There is no difference in behaviour to the current "makepkg --source" unless the user specifically requests it. The patch just makes --skipinteg do what its name says; skip integrity checks. Allan

5 11

[pacman-dev] [PATCH] Resolve dependencies whenever --syncdeps is passed and --nodeps is not
by Nezmer 18 May '10

18 May '10

With this patch, dependencies will be resolved and not silently ignored when running: makepkg --nobuild --syncdeps makepkg --repackage --syncdeps Also, a warning is displayed when repackaging and dependencies are not being resolved. Thank you Allan for the feedback. Signed-off-by: Nezmer <git(a)nezmer.info> --- scripts/makepkg.sh.in | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/scripts/makepkg.sh.in b/scripts/makepkg.sh.in index a2db90b..f57e066 100644 --- a/scripts/makepkg.sh.in +++ b/scripts/makepkg.sh.in @@ -1867,9 +1867,9 @@ if (( SOURCEONLY )); then fi # fix flyspray bug #5973 -if (( NODEPS || NOBUILD || REPKG )); then +if (( NODEPS || ( (NOBUILD || REPKG) && !DEP_BIN ) )); then # no warning message needed for nobuild, repkg - if (( NODEPS )); then + if (( NODEPS || REPKG )); then warning "$(gettext "Skipping dependency checks.")" fi elif [ $(type -p "${PACMAN%% *}") ]; then -- 1.7.1

2 1

[pacman-dev] [GIT] The official pacman repository branch, master, updated. v3.3.3-182-ga6ace98
by dan＠archlinux.org 17 May '10

17 May '10

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "The official pacman repository". The branch, master has been updated via a6ace987a913852af30bb5e3ae6bf145dc9c6157 (commit) via 3a85f83840888cbab09f2413fa3eb7c3b71055e5 (commit) via df842e11cfa4a41f7a333270ab587fe6ee2e8a27 (commit) from 25cd6c2e8da4d0659d259a38f00d512f97dbc962 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit a6ace987a913852af30bb5e3ae6bf145dc9c6157 Author: Dan McGee <dan(a)archlinux.org> Date: Mon May 17 18:47:30 2010 -0500 Add some machinery to test the Include directive After the previous patch that re-enabled its use outside of sync repository sections which we had unintentionally disabled. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 3a85f83840888cbab09f2413fa3eb7c3b71055e5 Author: Xavier Chantry <chantry.xavier(a)gmail.com> Date: Mon May 17 19:10:50 2010 +0200 Allow Include directive in any sections Fix a regression of 51f9e5e40a7 that only allowed Include in repo sections. Thanks to Marc - A. Dahlhaus for reporting the issue. Signed-off-by: Xavier Chantry <chantry.xavier(a)gmail.com> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit df842e11cfa4a41f7a333270ab587fe6ee2e8a27 Author: Dan McGee <dan(a)archlinux.org> Date: Mon May 17 18:23:07 2010 -0500 pactest: remove cargo-cult option reset This doesn't need to be here. We don't even support non-CamelCase options anymore. Signed-off-by: Dan McGee <dan(a)archlinux.org> ----------------------------------------------------------------------- Summary of changes: pactest/pmdb.py | 1 + pactest/pmtest.py | 8 +------ pactest/tests/{query005.py => config001.py} | 10 ++------ pactest/tests/config002.py | 12 +++++++++++ pactest/util.py | 13 +++++++---- src/pacman/pacman.c | 28 +++++++++++++++----------- 6 files changed, 41 insertions(+), 31 deletions(-) copy pactest/tests/{query005.py => config001.py} (51%) create mode 100644 pactest/tests/config002.py hooks/post-receive -- The official pacman repository

1 0

[pacman-dev] [PATCH] New pactest: unresolvable001.py
by Nagy Gabor 17 May '10

17 May '10

To test the regression of commit a4084b91. Signed-off-by: Nagy Gabor <ngaba(a)bibl.u-szeged.hu> --- pactest/tests/unresolvable001.py | 21 +++++++++++++++++++++ 1 files changed, 21 insertions(+), 0 deletions(-) create mode 100644 pactest/tests/unresolvable001.py diff --git a/pactest/tests/unresolvable001.py b/pactest/tests/unresolvable001.py new file mode 100644 index 0000000..c98a4c0 --- /dev/null +++ b/pactest/tests/unresolvable001.py @@ -0,0 +1,21 @@ +self.description = "An unresolvable dependency" + +packageA1 = pmpkg("dep") +self.addpkg2db("local", packageA1) + +packageA1up = pmpkg("dep", "2.0-1") +packageA1up.depends = ["fake"]; +self.addpkg2db("sync", packageA1up) + +packageA2up = pmpkg("package") +packageA2up.depends = ["dep"]; +self.addpkg2db("sync", packageA2up) + +self.args = "-S package dep --ask=32" + +self.addrule("PACMAN_RETCODE=0") +self.addrule("PKG_EXIST=package") +self.addrule("PKG_EXIST=dep") +self.addrule("PKG_VERSION=dep|1.0-1") + +self.expectfailure = True -- 1.7.1

3 3

[pacman-dev] [PATCH] Partial fix for the phonon/qt issue
by Nagy Gabor 17 May '10

17 May '10

This patch fixes the phonon/qt issue, if all to-be-upgraded packages are explicit targets (ie. only not-yet-installed packages are pulled by resolvedeps). This condition covers the most common situations, for example it should hold with every -Su operation. After this patch sync405.py passes, but sync406.py doesn't. The work is inspired by the patch of Henning Garus, thanks for his work: http://mailman.archlinux.org/pipermail/pacman-dev/2010-February/010429.html (I moved the alpm_list_diff computation to sync.c in order to compute it only once.) Signed-off-by: Nagy Gabor <ngaba(a)bibl.u-szeged.hu> --- lib/libalpm/deps.c | 10 +++------- lib/libalpm/deps.h | 2 +- lib/libalpm/sync.c | 6 +++++- pactest/tests/sync405.py | 2 -- 4 files changed, 9 insertions(+), 11 deletions(-) diff --git a/lib/libalpm/deps.c b/lib/libalpm/deps.c index 7dc734d..26f9b16 100644 --- a/lib/libalpm/deps.c +++ b/lib/libalpm/deps.c @@ -578,7 +578,7 @@ pmpkg_t *_alpm_resolvedep(pmdepend_t *dep, alpm_list_t *dbs, /* Computes resolvable dependencies for a given package and adds that package * and those resolvable dependencies to a list. * - * @param local is the local database + * @param localpkgs is the list of local packages * @param dbs_sync are the sync databases * @param pkg is the package to resolve * @param packages is a pointer to a list of packages which will be @@ -594,7 +594,7 @@ pmpkg_t *_alpm_resolvedep(pmdepend_t *dep, alpm_list_t *dbs, * unresolvable dependency, in which case the [*packages] list will be * unmodified by this function */ -int _alpm_resolvedeps(pmdb_t *local, alpm_list_t *dbs_sync, pmpkg_t *pkg, +int _alpm_resolvedeps(alpm_list_t *localpkgs, alpm_list_t *dbs_sync, pmpkg_t *pkg, alpm_list_t *preferred, alpm_list_t **packages, alpm_list_t *remove, alpm_list_t **data) { @@ -605,10 +605,6 @@ int _alpm_resolvedeps(pmdb_t *local, alpm_list_t *dbs_sync, pmpkg_t *pkg, ALPM_LOG_FUNC; - if(local == NULL) { - return(-1); - } - if(_alpm_pkg_find(*packages, pkg->name) != NULL) { return(0); } @@ -624,7 +620,7 @@ int _alpm_resolvedeps(pmdb_t *local, alpm_list_t *dbs_sync, pmpkg_t *pkg, for(i = alpm_list_last(*packages); i; i = i->next) { pmpkg_t *tpkg = i->data; targ = alpm_list_add(NULL, tpkg); - deps = alpm_checkdeps(_alpm_db_get_pkgcache(local), 0, remove, targ); + deps = alpm_checkdeps(localpkgs, 0, remove, targ); alpm_list_free(targ); for(j = deps; j; j = j->next) { pmdepmissing_t *miss = j->data; diff --git a/lib/libalpm/deps.h b/lib/libalpm/deps.h index 6681c59..ffc3aee 100644 --- a/lib/libalpm/deps.h +++ b/lib/libalpm/deps.h @@ -49,7 +49,7 @@ void _alpm_depmiss_free(pmdepmissing_t *miss); alpm_list_t *_alpm_sortbydeps(alpm_list_t *targets, int reverse); void _alpm_recursedeps(pmdb_t *db, alpm_list_t *targs, int include_explicit); pmpkg_t *_alpm_resolvedep(pmdepend_t *dep, alpm_list_t *dbs, alpm_list_t *excluding, int prompt); -int _alpm_resolvedeps(pmdb_t *local, alpm_list_t *dbs_sync, pmpkg_t *pkg, +int _alpm_resolvedeps(alpm_list_t *localpkgs, alpm_list_t *dbs_sync, pmpkg_t *pkg, alpm_list_t *preferred, alpm_list_t **packages, alpm_list_t *remove, alpm_list_t **data); int _alpm_dep_edge(pmpkg_t *pkg1, pmpkg_t *pkg2); diff --git a/lib/libalpm/sync.c b/lib/libalpm/sync.c index 6b625ed..be36941 100644 --- a/lib/libalpm/sync.c +++ b/lib/libalpm/sync.c @@ -431,17 +431,21 @@ int _alpm_sync_prepare(pmtrans_t *trans, pmdb_t *db_local, alpm_list_t *dbs_sync } } + /* Compute the fake local database for resolvedeps (partial fix for the phonon/qt issue) */ + alpm_list_t *localpkgs = alpm_list_diff(_alpm_db_get_pkgcache(db_local), trans->add, _alpm_pkg_cmp); + /* Resolve packages in the transaction one at a time, in addtion building up a list of packages which could not be resolved. */ for(i = trans->add; i; i = i->next) { pmpkg_t *pkg = i->data; - if(_alpm_resolvedeps(db_local, dbs_sync, pkg, trans->add, + if(_alpm_resolvedeps(localpkgs, dbs_sync, pkg, trans->add, &resolved, remove, data) == -1) { unresolvable = alpm_list_add(unresolvable, pkg); } /* Else, [resolved] now additionally contains [pkg] and all of its dependencies not already on the list */ } + alpm_list_free(localpkgs); /* If there were unresolvable top-level packages, prompt the user to see if they'd like to ignore them rather than failing the sync */ diff --git a/pactest/tests/sync405.py b/pactest/tests/sync405.py index d170cec..941a1af 100644 --- a/pactest/tests/sync405.py +++ b/pactest/tests/sync405.py @@ -22,5 +22,3 @@ self.addrule("PKG_EXIST=qt") self.addrule("PKG_EXIST=phonon") self.addrule("PKG_VERSION=qt|4.6.1-1") - -self.expectfailure = True -- 1.7.1

3 5

[pacman-dev] [PATCH] makepkg: Always resolve dependencies When --syncdeps is explicitly passed
by Nezmer 17 May '10

17 May '10

This change makes sense for me in general. The use-case I ran into was needing the dependencies when repackaging. Signed-off-by: Nezmer <git(a)nezmer.info> --- scripts/makepkg.sh.in | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/scripts/makepkg.sh.in b/scripts/makepkg.sh.in index a2db90b..34261b5 100644 --- a/scripts/makepkg.sh.in +++ b/scripts/makepkg.sh.in @@ -1867,7 +1867,7 @@ if (( SOURCEONLY )); then fi # fix flyspray bug #5973 -if (( NODEPS || NOBUILD || REPKG )); then +if (( (NODEPS || NOBUILD || REPKG) && ! DEP_BIN )); then # no warning message needed for nobuild, repkg if (( NODEPS )); then warning "$(gettext "Skipping dependency checks.")" -- 1.7.0.4

3 3

[pacman-dev] [PATCH] redirect scriptlet stderr through alpm
by Jonathan Conder 17 May '10

17 May '10

Fixes FS#18770, and hopefully an occasional deadlock in my frontend as well. This introduces a new EVENT: PM_TRANS_EVT_SCRIPTLET_ERROR, and moves all scriptlet-related callbacks into the parent process. I modelled my work on the GLib function g_spawn_sync, but I am quite unfamiliar with POSIX so there could be some issues with the implementation. However, all the tests I have performed so far (against v3.3.3) seemed to work fine (even the kernel installed and gave me the right warnings). This patch might also make it easier to configure the shell used for install scriptlets. I'm not 100% happy with the code structure at the moment, so any suggestions would be welcome. Signed-off-by: Jonathan Conder <j(a)skurvy.no-ip.org> --- lib/libalpm/alpm.h | 4 + lib/libalpm/util.c | 162 ++++++++++++++++++++++++++++++++++++++----------- src/pacman/callback.c | 3 + 3 files changed, 133 insertions(+), 36 deletions(-) diff --git a/lib/libalpm/alpm.h b/lib/libalpm/alpm.h index 3329132..2692777 100644 --- a/lib/libalpm/alpm.h +++ b/lib/libalpm/alpm.h @@ -364,6 +364,10 @@ typedef enum _pmtransevt_t { * A line of text is passed to the callback. */ PM_TRANS_EVT_SCRIPTLET_INFO, + /** Scriptlet has printed information to stderr. + * A line of text is passed to the callback. + */ + PM_TRANS_EVT_SCRIPTLET_ERROR, /** Files will be downloaded from a repository. * The repository's tree name is passed to the callback. */ diff --git a/lib/libalpm/util.c b/lib/libalpm/util.c index 32eaa44..1f76372 100644 --- a/lib/libalpm/util.c +++ b/lib/libalpm/util.c @@ -444,10 +444,52 @@ int _alpm_logaction(int usesyslog, FILE *f, const char *fmt, va_list args) return(ret); } +static int _fddup2(int fds[2], int i, int fd) +{ + while(dup2(fds[i], fd) == -1) { + if(errno != EINTR) { + return(-1); + } + } + return(0); +} + +static int _fdselect(int fdout, int fderr, fd_set *fds) +{ + int nfds = -1; + FD_ZERO(fds); + + if(fdout >= 0) { + FD_SET(fdout, fds); + if(fdout > nfds) { + nfds = fdout; + } + } + if(fderr >= 0) { + FD_SET(fderr, fds); + if(fderr > nfds) { + nfds = fderr; + } + } + if(++nfds == 0) { + return(2); + } + + if(select(nfds, fds, NULL, NULL, NULL) == -1) { + if(errno == EINTR) { + FD_ZERO(fds); + } else { + return(1); + } + } + return(0); +} + int _alpm_run_chroot(const char *root, const char *cmd) { char cwd[PATH_MAX]; pid_t pid; + int fdout[2], fderr[2]; int restore_cwd = 0; int retval = 0; @@ -471,6 +513,13 @@ int _alpm_run_chroot(const char *root, const char *cmd) /* Flush open fds before fork() to avoid cloning buffers */ fflush(NULL); + /* Create pipes to redirect stderr and stdout */ + if(pipe(fdout) == -1 || pipe(fderr) == -1) { + _alpm_log(PM_LOG_ERROR, _("could not create a pipe (%s)\n"), strerror(errno)); + retval = 1; + goto cleanup; + } + /* fork- parent and child each have seperate code blocks below */ pid = fork(); if(pid == -1) { @@ -480,56 +529,97 @@ int _alpm_run_chroot(const char *root, const char *cmd) } if(pid == 0) { - FILE *pipe; /* this code runs for the child only (the actual chroot/exec) */ - _alpm_log(PM_LOG_DEBUG, "chrooting in %s\n", root); + close(fdout[0]); + close(fderr[0]); + + /* redirect standard output and error */ + _fddup2(fdout, 1, 1); + _fddup2(fderr, 1, 2); + close(fdout[1]); + close(fderr[1]); + if(chroot(root) != 0) { - _alpm_log(PM_LOG_ERROR, _("could not change the root directory (%s)\n"), - strerror(errno)); + fprintf(stderr, _("could not change the root directory (%s)\n"), strerror(errno)); exit(1); } + if(chdir("/") != 0) { - _alpm_log(PM_LOG_ERROR, _("could not change directory to / (%s)\n"), - strerror(errno)); + fprintf(stderr, _("could not change directory to / (%s)\n"), strerror(errno)); exit(1); } + + /* run the scriptlet and exit */ umask(0022); - pipe = popen(cmd, "r"); - if(!pipe) { - _alpm_log(PM_LOG_ERROR, _("call to popen failed (%s)\n"), - strerror(errno)); - exit(1); - } - while(!feof(pipe)) { - char line[PATH_MAX]; - if(fgets(line, PATH_MAX, pipe) == NULL) - break; - alpm_logaction("%s", line); - EVENT(handle->trans, PM_TRANS_EVT_SCRIPTLET_INFO, line, NULL); - } - retval = pclose(pipe); - exit(WEXITSTATUS(retval)); + execl("/bin/sh", "sh", "-c", cmd, (char *) NULL); + fprintf(stderr, _("call to execl failed (%s)\n"), strerror(errno)); + exit(1); } else { - /* this code runs for the parent only (wait on the child) */ - pid_t retpid; + /* this code runs for the parent only (process output from the child) */ + fd_set fds; + FILE *fout, *ferr; int status; - do { - retpid = waitpid(pid, &status, 0); - } while(retpid == -1 && errno == EINTR); - if(retpid == -1) { - _alpm_log(PM_LOG_ERROR, _("call to waitpid failed (%s)\n"), - strerror(errno)); + + close(fdout[1]); + close(fderr[1]); + + /* use buffered streams to read line by line */ + if((fout = fdopen(fdout[0], "r")) == NULL || (ferr = fdopen(fderr[0], "r")) == NULL) { retval = 1; - goto cleanup; + close(fdout[0]); + close(fderr[0]); } else { - /* check the return status, make sure it is 0 (success) */ - if(WIFEXITED(status)) { - _alpm_log(PM_LOG_DEBUG, "call to waitpid succeeded\n"); - if(WEXITSTATUS(status) != 0) { - _alpm_log(PM_LOG_ERROR, _("command failed to execute correctly\n")); - retval = 1; + while((retval = _fdselect(fdout[0], fderr[0], &fds)) == 0) { + char line[PATH_MAX]; + + if(FD_ISSET(fdout[0], &fds)) { + if(fgets(line, PATH_MAX, fout) != NULL) { + alpm_logaction("%s", line); + EVENT(handle->trans, PM_TRANS_EVT_SCRIPTLET_INFO, line, NULL); + } else { + fdout[0] = -1; + } + } + + if(FD_ISSET(fderr[0], &fds)) { + if(fgets(line, PATH_MAX, ferr) != NULL) { + alpm_logaction("%s", line); + EVENT(handle->trans, PM_TRANS_EVT_SCRIPTLET_ERROR, line, NULL); + } else { + fderr[0] = -1; + } } } + + fclose(fout); + fclose(ferr); + } + + while(waitpid(pid, &status, 0) == -1) { + if (errno != EINTR) { + _alpm_log(PM_LOG_ERROR, _("call to waitpid failed (%s)\n"), + strerror(errno)); + retval = 1; + goto cleanup; + } + } + + /* report error from above after the child has exited */ + if(retval == 1) { + _alpm_log(PM_LOG_ERROR, _("could not read from pipe (%s)\n"), + strerror(errno)); + goto cleanup; + } else { + retval = 0; + } + + /* check the return status, make sure it is 0 (success) */ + if(WIFEXITED(status)) { + _alpm_log(PM_LOG_DEBUG, "call to waitpid succeeded\n"); + if(WEXITSTATUS(status) != 0) { + _alpm_log(PM_LOG_ERROR, _("command failed to execute correctly\n")); + retval = 1; + } } } diff --git a/src/pacman/callback.c b/src/pacman/callback.c index f5bf17d..a33178e 100644 --- a/src/pacman/callback.c +++ b/src/pacman/callback.c @@ -226,6 +226,9 @@ void cb_trans_evt(pmtransevt_t event, void *data1, void *data2) case PM_TRANS_EVT_SCRIPTLET_INFO: printf("%s", (char*)data1); break; + case PM_TRANS_EVT_SCRIPTLET_ERROR: + fprintf(stderr, "%s", (char*)data1); + break; case PM_TRANS_EVT_RETRIEVE_START: printf(_(":: Retrieving packages from %s...\n"), (char*)data1); break; -- 1.7.1

2 4