- Pacman-dev - lists.archlinux.org

[pacman-dev] [PATCH] Turn off colours onces finished using them
by Allan McRae 04 Jun '16

04 Jun '16

The colour of the package version could leak onto the next line in both -Qo and -Fo. Signed-off-by: Allan McRae <allan(a)archlinux.org> --- src/pacman/files.c | 4 ++-- src/pacman/query.c | 5 +++-- 2 files changed, 5 insertions(+), 4 deletions(-) diff --git a/src/pacman/files.c b/src/pacman/files.c index 692fcd5..f5cba5b 100644 --- a/src/pacman/files.c +++ b/src/pacman/files.c @@ -78,10 +78,10 @@ static int files_fileowner(alpm_list_t *syncs, alpm_list_t *targets) { print_line_machinereadable(repo, pkg, filename); } else if(!config->quiet) { const colstr_t *colstr = &config->colstr; - printf(_("%s is owned by %s%s/%s%s %s%s\n"), filename, + printf(_("%s is owned by %s%s/%s%s %s%s%s\n"), filename, colstr->repo, alpm_db_get_name(repo), colstr->title, alpm_pkg_get_name(pkg), colstr->version, - alpm_pkg_get_version(pkg)); + alpm_pkg_get_version(pkg), colstr->nocolor); } else { printf("%s/%s\n", alpm_db_get_name(repo), alpm_pkg_get_name(pkg)); } diff --git a/src/pacman/query.c b/src/pacman/query.c index 522212d..a8f4e67 100644 --- a/src/pacman/query.c +++ b/src/pacman/query.c @@ -84,8 +84,9 @@ static void print_query_fileowner(const char *filename, alpm_pkg_t *info) { if(!config->quiet) { const colstr_t *colstr = &config->colstr; - printf(_("%s is owned by %s%s %s%s\n"), filename, colstr->title, - alpm_pkg_get_name(info), colstr->version, alpm_pkg_get_version(info)); + printf(_("%s is owned by %s%s %s%s%s\n"), filename, colstr->title, + alpm_pkg_get_name(info), colstr->version, alpm_pkg_get_version(info), + colstr->nocolor); } else { printf("%s\n", alpm_pkg_get_name(info)); } -- 2.8.3

1 0

[pacman-dev] A WIP fix for removing unneeded dependency cycles
by Ashley Whetter 21 May '16

21 May '16

This is a work in progress patchset. DO NOT MERGE! I've been looking at a fix for FS#41031 [1]. What I've attempted to do in this prototype is to fix the issue without needing to calculate the strongly connected components of the dependency graph to see if I could improve on run time performance or memory consumption. The way this works is I first perform depth-first cycle detection on the graph. When a cycle is detected the list of packages in the cycle is added to the removes list of each package in the cycle. This marks a package as "in a cycle". After this preprocessing step a package is determined as removable by pretending that every other package in the cycle is being removed, and then determining what or not the package itself is removable as normal. My thinking from this prototype though is that it isn't as great a solution as calculating the strongly connected components of the dependency graph. Doing cycle detection this way means that dependency graphs with packages belonging to more than one cycle aren't detected. Although this is a crazy edge case, it's theoretically a potential graph, and implementing Tarjan's strongly connected components algorithm instead of the cycle detection used here shouldn't be much more work. It can still use the removes array and the `_can_remove_cycle` method for determining whether a cycle is removable; only the preprocessing step would need to change. The only way that I think calculating the strongly connected components could be worse than this prototype is that it requires a list of every vertex in the dependency graph. This is easy enough to calculate with a breadth first search but it will use more memory than simple cycle detection. So I wanted to see what people's thoughts were on these two solutions and if there was a preference to one or the other. Ashley [1] https://bugs.archlinux.org/task/41031

3 4

[pacman-dev] [GIT] The official pacman repository branch, release/5.0.x, updated. v5.0.1-6-gc2f9758
by allan＠archlinux.org 18 May '16

18 May '16

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, release/5.0.x has been updated via c2f9758018479c15a34887d960f36ba9b532175f (commit) from 5469161dad387b71d605cfeb2c6c369df6f939e3 (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 c2f9758018479c15a34887d960f36ba9b532175f Author: Leonid Isaev <leonid.isaev(a)jila.colorado.edu> Date: Sat May 7 17:24:17 2016 -0600 Use a more generic regexp when parsing output of gpg(1) in signature verification. The current way of extracting key trust from output of gpg --verify is not very robust against changes in the format of said output. As a result, pacman-key can return an error even if the signature is actuall good. This change relaxes the regexp when parsing output of gpg. Signed-off-by: Leonid Isaev <leonid.isaev(a)jila.colorado.edu> Signed-off-by: Allan McRae <allan(a)archlinux.org> (cherry picked from commit 892a1076c00a2b0097145c35f5d8ef590216dac0) ----------------------------------------------------------------------- Summary of changes: scripts/pacman-key.sh.in | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) hooks/post-receive -- The official pacman repository

1 0

[pacman-dev] [GIT] The official pacman repository branch, master, updated. v5.0.1-35-g839417e
by allan＠archlinux.org 18 May '16

18 May '16

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 839417e8c65e73e632c77d3ecd68539494850de1 (commit) via f018317f488e35f3a9a3ce43c977baceec13d08b (commit) via 4f114f38178aeaffcc571aa42e2f930ca916f79d (commit) via 892a1076c00a2b0097145c35f5d8ef590216dac0 (commit) via 87082e3f44c4f92c63d881db38a0fff0565fc9bc (commit) via 7bebe4deff0afd4e220b5c4d6024bffb08bd794f (commit) via 02a0bf550a22e199f48537b7eee87361b112e8a0 (commit) from 3729ef7a9acf75080fb6f60d13ea80cfd36d855d (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 839417e8c65e73e632c77d3ecd68539494850de1 Author: Christian Hesse <mail(a)eworm.de> Date: Tue May 17 13:23:44 2016 +0200 pacman.8: add link to alpm-hooks(5) in see also Signed-off-by: Christian Hesse <mail(a)eworm.de> Signed-off-by: Allan McRae <allan(a)archlinux.org> commit f018317f488e35f3a9a3ce43c977baceec13d08b Author: Christian Hesse <mail(a)eworm.de> Date: Tue May 17 13:23:43 2016 +0200 libalpm.3: add link to alpm-hooks(5) in see also Signed-off-by: Christian Hesse <mail(a)eworm.de> Signed-off-by: Allan McRae <allan(a)archlinux.org> commit 4f114f38178aeaffcc571aa42e2f930ca916f79d Author: Ashley Whetter <ashley(a)awhetter.co.uk> Date: Sun May 8 17:28:28 2016 +0100 libmakepkg: extract functions for integrity checking Signed-off-by: Ashley Whetter <ashley(a)awhetter.co.uk> Signed-off-by: Allan McRae <allan(a)archlinux.org> commit 892a1076c00a2b0097145c35f5d8ef590216dac0 Author: Leonid Isaev <leonid.isaev(a)jila.colorado.edu> Date: Sat May 7 17:24:17 2016 -0600 Use a more generic regexp when parsing output of gpg(1) in signature verification. The current way of extracting key trust from output of gpg --verify is not very robust against changes in the format of said output. As a result, pacman-key can return an error even if the signature is actuall good. This change relaxes the regexp when parsing output of gpg. Signed-off-by: Leonid Isaev <leonid.isaev(a)jila.colorado.edu> Signed-off-by: Allan McRae <allan(a)archlinux.org> commit 87082e3f44c4f92c63d881db38a0fff0565fc9bc Author: Alastair Hughes <hobbitalastair(a)gmail.com> Date: Sun Apr 3 16:19:49 2016 +1200 makepkg: ignore the architecture for --printsrcinfo Signed-off-by: Alastair Hughes <hobbitalastair(a)gmail.com> Signed-off-by: Allan McRae <allan(a)archlinux.org> commit 7bebe4deff0afd4e220b5c4d6024bffb08bd794f Author: Allan McRae <allan(a)archlinux.org> Date: Tue Mar 29 22:00:02 2016 +1000 Report local file URL for -Sp operations if package is in cache When using "pacman -Sp" operation to get URLs of packages to download, it is useful to know which packages are already in the file cache and do not need downloaded. Print packages in the cache with a file:// prefix. e.g $ pacman -Sp glibc file:///var/cache/pacman/glibc-2.23-1-x86_64.pkg.tar.xz Also use package locations in case statements rather than opersations. This allows the ALPM_PKG_SYNCDB to fall thorough to just printing the package name for weird serverless repo setups. Fixes FS#15868 Signed-off-by: Allan McRae <allan(a)archlinux.org> commit 02a0bf550a22e199f48537b7eee87361b112e8a0 Author: Allan McRae <allan(a)archlinux.org> Date: Sat Apr 2 22:41:17 2016 +1000 makepkg: move build enviroment set-up to function This is a requirement to split the preparation of the build environment into libmakepkg, which will allow dropping in extensions (e.g. to allow PGO). After this patch, disabling buildflags or makeflags and enabling debug CFLAGS will only effect the build(), check() and package() functions. The relevant variables are no longer exported for the prepare() function. This should have zero impact for the prepare() function of a properly written PKGBUILD, as no building/linking is done there... Signed-off-by: Allan McRae <allan(a)archlinux.org> ----------------------------------------------------------------------- Summary of changes: doc/libalpm.3.txt | 3 +- doc/pacman.8.txt | 3 +- scripts/Makefile.am | 6 + scripts/libmakepkg/.gitignore | 2 + .../{lint_package.sh.in => integrity.sh.in} | 37 +- .../libmakepkg/integrity/generate_checksum.sh.in | 107 +++++ .../generate_signature.sh.in} | 44 +- scripts/libmakepkg/integrity/verify_checksum.sh.in | 130 ++++++ .../libmakepkg/integrity/verify_signature.sh.in | 216 +++++++++ scripts/libmakepkg/util/pkgbuild.sh.in | 20 +- scripts/makepkg.sh.in | 509 ++------------------- scripts/pacman-key.sh.in | 2 +- scripts/po/POTFILES.in | 5 + src/pacman/util.c | 27 +- 14 files changed, 600 insertions(+), 511 deletions(-) copy scripts/libmakepkg/{lint_package.sh.in => integrity.sh.in} (53%) create mode 100644 scripts/libmakepkg/integrity/generate_checksum.sh.in copy scripts/libmakepkg/{tidy/purge.sh.in => integrity/generate_signature.sh.in} (56%) create mode 100644 scripts/libmakepkg/integrity/verify_checksum.sh.in create mode 100644 scripts/libmakepkg/integrity/verify_signature.sh.in hooks/post-receive -- The official pacman repository

1 0

[pacman-dev] [PATCH 1/2] libalpm.3: add link to alpm-hooks(5) in see also
by Christian Hesse 18 May '16

18 May '16

From: Christian Hesse <mail(a)eworm.de> Signed-off-by: Christian Hesse <mail(a)eworm.de> --- doc/libalpm.3.txt | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/doc/libalpm.3.txt b/doc/libalpm.3.txt index e2ea32c..aa667a1 100644 --- a/doc/libalpm.3.txt +++ b/doc/libalpm.3.txt @@ -34,6 +34,7 @@ See linkman:pacman.conf[5] for more details on configuring libalpm using the See Also -------- -linkman:pacman[8], linkman:makepkg[8], linkman:pacman.conf[5] +linkman:alpm-hooks[5], linkman:makepkg[8], linkman:pacman[8], +linkman:pacman.conf[5] include::footer.txt[] -- 2.8.2

2 2

[pacman-dev] Proposal to split out makepkg integrity methods
by Ashley Whetter 18 May '16

18 May '16

Hello everyone, I'm looking to split some makepkg methods out into a of makepkg.sh.in and into a new libmakepkg module. My plan is to have an "integrity" folder in libmakepkg. In this folder there would be a "generate.sh" file with the `generate_checksums`, `generate_one_checksum`, and `create_signature` methods. Another file, "check.sh" would contain `check_source_integrity`and it's dependencies; namely `check_checksums`, `check_pgpsigs`, `parse_gpg_statusfile`, `verify_integrity_sums`, and `verify_integrity_one`. "integrity" would depend on "util/pkgbuild.sh" as `check_checksums` and `verify_integrity_sum` both call `array_build`, and `check_pgpsigs` calls `get_all_sources` and `get_all_sources_for_arch`. I also plan on moving `get_integlist` into "util/pkgbuild.sh", or maybe into an "integrity/util.sh" file (but "util/pkgbuild.sh" makes more sense to me). Another option would be to have an "integrity.sh" file under "integrity" (or maybe some other name) and also a "signature.sh" file, and to split the methods by whether they are related to strictly integrity checking or signature checking. Are there any thoughts or suggestions on this plan of attack? Ashley

2 5

[pacman-dev] [PATCH] makepkg: ignore the architecture for --printsrcinfo
by Alastair Hughes 05 May '16

05 May '16

Signed-off-by: Alastair Hughes <hobbitalastair(a)gmail.com> --- It appears that SRCINFO files should be independent of architecture, however makepkg will currently fail if the SRCINFO file is being generated on a machine with a different architecture. This is in the same vein as c72ca3d5; I believe that creating source tarballs would be the only other candidate, but IGNOREARCH gets set later in that case. scripts/makepkg.sh.in | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/scripts/makepkg.sh.in b/scripts/makepkg.sh.in index f80e37a..5487884 100644 --- a/scripts/makepkg.sh.in +++ b/scripts/makepkg.sh.in @@ -1707,7 +1707,8 @@ while true; do -o|--nobuild) NOBUILD=1 ;; -p) shift; BUILDFILE=$1 ;; --packagelist) PACKAGELIST=1 IGNOREARCH=1;; - --printsrcinfo) PRINTSRCINFO=1 ;; + # generating a srcinfo does not depend on architecture + --printsrcinfo) PRINTSRCINFO=1 IGNOREARCH=1;; -r|--rmdeps) RMDEPS=1 ;; -R|--repackage) REPKG=1 ;; --sign) SIGNPKG='y' ;; -- 2.8.0

2 1

[pacman-dev] [GIT] The official pacman repository branch, release/5.0.x, updated. v5.0.1-5-g5469161
by allan＠archlinux.org 05 May '16

05 May '16

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, release/5.0.x has been updated via 5469161dad387b71d605cfeb2c6c369df6f939e3 (commit) via 07ccbc26914bea24983931532b97661a75b8a53c (commit) via a66cba1b08cc7c9f55efcfda6f4ab3cafff6b874 (commit) from d4f499f5638a68a3b7e17c5171aabdf535b79bd4 (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 5469161dad387b71d605cfeb2c6c369df6f939e3 Author: Allan McRae <allan(a)archlinux.org> Date: Mon Mar 28 16:24:22 2016 +1000 PKGBUILD.5: document that the pkgver() function runs after prepare() Signed-off-by: Allan McRae <allan(a)archlinux.org> (cherry picked from commit 59010386106eded2ee67798fa80215f250a53761) commit 07ccbc26914bea24983931532b97661a75b8a53c Author: Allan McRae <allan(a)archlinux.org> Date: Sun Mar 20 21:23:32 2016 +1000 Do not add root prefix twice when checking database files When checking .INSTALL and .CHANGELOG files in the mtree file, we need to find the path they are stored in the local database. This was appending the root prefix twice as alpm_option_get_dbpath already returns the absolute path to the database. While fixing that issue I added checks that the paths for the database files were not longer than PATH_MAX. Fixes FS#48563. Signed-off-by: Allan McRae <allan(a)archlinux.org> (cherry picked from commit 2ee1706a72858ea365c0e7a1a72d92ef2dae2f75) commit a66cba1b08cc7c9f55efcfda6f4ab3cafff6b874 Author: Allan McRae <allan(a)archlinux.org> Date: Mon Mar 21 11:49:16 2016 +1000 libmakepkg: ensure emptydir find command acts on individual directories Using "-exec command {} +" systax exits on any error. Such errors occur when running rmdir on a non-empty directory. Switch to "{} ;" syntax instead which avoids exiting before the find command is completed. Fixes FS#48515. Note, we can not use "-empty" in the find command because it is not supported by Busybox find, and the "--ignore-fail-on-non-empty" flag for rmdir is not available on BSD rmdir variants. Signed-off-by: Allan McRae <allan(a)archlinux.org> (cherry picked from commit baf1ff64e6ea23e93f307480480d87159cd2e487) ----------------------------------------------------------------------- Summary of changes: doc/PKGBUILD.5.txt | 8 ++--- scripts/libmakepkg/tidy/emptydirs.sh.in | 3 +- src/pacman/check.c | 54 ++++++++++++++++++--------------- 3 files changed, 36 insertions(+), 29 deletions(-) hooks/post-receive -- The official pacman repository

1 0

[pacman-dev] [GIT] The official pacman repository branch, master, updated. v5.0.1-28-g3729ef7
by allan＠archlinux.org 05 May '16

05 May '16

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 3729ef7a9acf75080fb6f60d13ea80cfd36d855d (commit) via b012da645e14e5f79968d6670dcc5e51232c4adb (commit) via 6c96ad36e082755e8a008b34aedde55a9a9f3dcd (commit) via 5b9bc6024c246f232c64f6409088fedf558917af (commit) via 2694d17ad99bab1cea5aad0ea4781ea4d51e596b (commit) via 0088a7ab1cc2676a31debc05878a80d0bd9a6ba4 (commit) via 7e7c64f513460cfb391e2fc558e6fdbc83faa54d (commit) via 7d1ea0b814723d6a422c5569df19513c40bf99fe (commit) via 502e0d1db0f588d3067a2d464824e8f1cff20ca1 (commit) via a194fb434e030a1668f08283c8a4d71428be86b5 (commit) via ee960048b2394557f7155f053bf3ff5c8bb4b5e1 (commit) via 3ed57be450e35c3a3aaa99af722b615a989027c7 (commit) from 4960be7764678783c1ac81e884ad34a535769ad1 (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 3729ef7a9acf75080fb6f60d13ea80cfd36d855d Author: Allan McRae <allan(a)archlinux.org> Date: Wed May 4 17:39:34 2016 +1000 Avoid logical OR duplication warning from gcc-6 The value EAGAIN is allowed by POSIX to be the same as EWOULDBLOCK, but this is not guaranteed. Thus on some systems (e.g. glibc Linux), we get a warning that the logical OR is being performed on two expressions of the same type. We can not get rid of this test in case any system defines these as unique values. Suggested-by: Dave Reisner Signed-off-by: Allan McRae <allan(a)archlinux.org> commit b012da645e14e5f79968d6670dcc5e51232c4adb Author: Ashley Whetter <ashley(a)awhetter.co.uk> Date: Sun May 1 18:03:08 2016 +0100 Add -m/--nocolor options to updpkgsums Fixes FS#43346 Signed-off-by: Ashley Whetter <ashley(a)awhetter.co.uk> Signed-off-by: Allan McRae <allan(a)archlinux.org> commit 6c96ad36e082755e8a008b34aedde55a9a9f3dcd Author: Ashley Whetter <ashley(a)awhetter.co.uk> Date: Sun May 1 22:24:20 2016 +0100 pacsort help clearly states that files contain inputs to be sorted Fixes FS#44121 Signed-off-by: Ashley Whetter <ashley(a)awhetter.co.uk> Signed-off-by: Allan McRae <allan(a)archlinux.org> commit 5b9bc6024c246f232c64f6409088fedf558917af Author: Allan McRae <allan(a)archlinux.org> Date: Wed May 4 17:38:03 2016 +1000 Prevent wrapping of enum items GCC-6 points out that the value we use for the sentinal in enums is actually too large for the integer type. Reduce the bitshift by one to fix this. Signed-off-by: Allan McRae <allan(a)archlinux.org> commit 2694d17ad99bab1cea5aad0ea4781ea4d51e596b Author: Eric Engestrom <eric(a)engestrom.ch> Date: Fri Apr 29 22:32:13 2016 +0100 fix spelling mistakes Signed-off-by: Eric Engestrom <eric(a)engestrom.ch> Signed-off-by: Allan McRae <allan(a)archlinux.org> commit 0088a7ab1cc2676a31debc05878a80d0bd9a6ba4 Author: Xavion <Xavion (dot) 0 (at) Gmail (dot) com> Date: Mon Mar 28 11:54:24 2016 +1100 Add colour to the output of the "-{F, Q}o" operations. Matching output for -s operations, the repository is coloured 'magenta', the package name is 'bold', and the version is outputted in 'green'. Signed-off-by: Xavion <Xavion (dot) 0 (at) Gmail (dot) com> Signed-off-by: Allan McRae <allan(a)archlinux.org> commit 7e7c64f513460cfb391e2fc558e6fdbc83faa54d Author: Xavion <Xavion (dot) 0 (at) Gmail (dot) com> Date: Mon Mar 28 09:35:37 2016 +1100 Add colour to group selection dialog Colour the group name in 'blue' and the repository names in 'magenta'. Signed-off-by: Xavion <Xavion (dot) 0 (at) Gmail (dot) com> Signed-off-by: Allan McRae <allan(a)archlinux.org> commit 7d1ea0b814723d6a422c5569df19513c40bf99fe Author: Allan McRae <allan(a)archlinux.org> Date: Sat Apr 2 16:05:55 2016 +1000 Handle provides with -Q It is useful to be able to use "pacman -Qi" on any dependency, even if that dependency is a provide. For example, on Arch Linux systems, "sh" is provided by the "bash" package, and many packages depend on "sh". Querying the package that provides the "sh" dependency currently requires first searching for "sh". This patch allows the use of "pacman -Qi" on a provide. Fixes FS#20650. Signed-off-by: Allan McRae <allan(a)archlinux.org> commit 502e0d1db0f588d3067a2d464824e8f1cff20ca1 Author: Allan McRae <allan(a)archlinux.org> Date: Sat Apr 2 15:23:25 2016 +1000 pacman_query: move error messages into relevant if statements This ensures any additions to these test do not have to rely on the correct error condition being set by libalpm. Signed-off-by: Allan McRae <allan(a)archlinux.org> commit a194fb434e030a1668f08283c8a4d71428be86b5 Author: Christian Hesse <mail(a)eworm.de> Date: Wed Mar 30 22:24:41 2016 +0200 pacman.8: add actual operation to captions for easy searching I caught myself searching for operations when in need for an option. However this made me find transaction and update options only. So add the actual operations to captions for easy searching. Signed-off-by: Christian Hesse <mail(a)eworm.de> Signed-off-by: Allan McRae <allan(a)archlinux.org> commit ee960048b2394557f7155f053bf3ff5c8bb4b5e1 Author: Allan McRae <allan(a)archlinux.org> Date: Tue Mar 29 20:21:42 2016 +1000 Remove notification of system upgrade when only printing URLs Signed-off-by: Allan McRae <allan(a)archlinux.org> commit 3ed57be450e35c3a3aaa99af722b615a989027c7 Author: Allan McRae <allan(a)archlinux.org> Date: Tue Mar 29 16:36:59 2016 +1000 Print replacements when using -Sup When printing a list of URLs of packages to be updated, pacman was ignoring any replacements that would be made in the update process. Fixes FS#35812 Signed-off-by: Allan McRae <allan(a)archlinux.org> ----------------------------------------------------------------------- Summary of changes: contrib/updpkgsums.sh.in | 23 +++++++++---- contrib/zsh_completion.in | 6 ++-- doc/pacman.8.txt | 23 ++++++------- lib/libalpm/alpm.h | 4 +-- lib/libalpm/be_package.c | 2 +- lib/libalpm/db.h | 2 +- lib/libalpm/deps.c | 2 +- lib/libalpm/package.h | 2 +- lib/libalpm/util.c | 14 ++++++-- src/pacman/callback.c | 12 ++++--- src/pacman/files.c | 6 ++-- src/pacman/query.c | 40 ++++++++++++---------- src/pacman/sync.c | 15 ++++---- src/pacman/util.c | 3 +- src/util/pacsort.c | 5 ++- test/pacman/tests/TESTS | 1 + .../sync-sysupgrade-print-replaced-packages.py | 14 ++++++++ 17 files changed, 112 insertions(+), 62 deletions(-) create mode 100644 test/pacman/tests/sync-sysupgrade-print-replaced-packages.py hooks/post-receive -- The official pacman repository

1 0

[pacman-dev] [PATCH] Report local file URL for -Sp operations if package is in cache
by Allan McRae 05 May '16

05 May '16

When using "pacman -Sp" operation to get URLs of packages to download, it is useful to know which packages are already in the file cache and do not need downloaded. Print packages in the cache with a file:// prefix. e.g $ pacman -Sp glibc file:///var/cache/pacman/glibc-2.23-1-x86_64.pkg.tar.xz Also use package locations in case statements rather than opersations. This allows the ALPM_PKG_SYNCDB to fall thorough to just printing the package name for weird serverless repo setups. Fixes FS#15868 Signed-off-by: Allan McRae <allan(a)archlinux.org> --- v3: - use ALPM_PKG_FROM_* in the switch statement - get fall-through correct src/pacman/util.c | 27 ++++++++++++++++++++++++--- 1 file changed, 24 insertions(+), 3 deletions(-) diff --git a/src/pacman/util.c b/src/pacman/util.c index 0155493..81780f7 100644 --- a/src/pacman/util.c +++ b/src/pacman/util.c @@ -1025,16 +1025,37 @@ static char *pkg_get_location(alpm_pkg_t *pkg) { alpm_list_t *servers; char *string = NULL; - switch(config->op) { - case PM_OP_SYNC: + switch(alpm_pkg_get_origin(pkg)) { + case ALPM_PKG_FROM_SYNCDB: + if(alpm_pkg_download_size(pkg) == 0) { + /* file is already in the package cache */ + alpm_list_t *i; + const char *pkgfile = alpm_pkg_get_filename(pkg); + char path[PATH_MAX]; + struct stat buf; + + for(i = alpm_option_get_cachedirs(config->handle); i; i = i->next) { + snprintf(path, PATH_MAX, "%s%s", (char *)i->data, pkgfile); + if(stat(path, &buf) == 0 && S_ISREG(buf.st_mode)) { + pm_asprintf(&string, "file://%s", path); + return string; + } + } + } + servers = alpm_db_get_servers(alpm_pkg_get_db(pkg)); if(servers) { pm_asprintf(&string, "%s/%s", (char *)(servers->data), alpm_pkg_get_filename(pkg)); return string; } - case PM_OP_UPGRADE: + + /* fallthrough - for theoretical serverless repos */ + + case ALPM_PKG_FROM_FILE: return strdup(alpm_pkg_get_filename(pkg)); + + case ALPM_PKG_FROM_LOCALDB: default: pm_asprintf(&string, "%s-%s", alpm_pkg_get_name(pkg), alpm_pkg_get_version(pkg)); return string; -- 2.8.2

1 0