March 2009 - Pacman-dev - lists.archlinux.org

[pacman-dev] Copyright notices in pacman code
by Allan McRae 12 Apr '09

12 Apr '09

Hi all, I was looking through makepkg for areas that need tidied before the upcoming(!) release when I noticed the copyright notice in the version output is is "2002-2007 Judd Vinet". (Note it is 2002-2008 in the header.) So I went to add Dan here for the newer years (being the current project lead - that way I can blame him when I break something...). Then I did a "grep -R Copyright *" and noticed there is a lot that has been altered that has not had copyrights updated. I don't know what the best practice with this should be. Should someone make a big patch bumps the copyright years and adds new lines as necessary? BTW, this is my favourite line: lib/libalpm/po/it.po:# Copyright (C) 2007 THE libalpm'S COPYRIGHT HOLDER Allan

4 5

[pacman-dev] Some comments on Bryan patches (bugs)
by Nagy Gabor 12 Apr '09

12 Apr '09

Hi! I know that I speak too late, but I haven't time/motivation to review this patch nowadays. 1. See alpm/sync.c/sync_prepare(). When user decides to remove unresolvable targets from the list, our code doesn't remove any targets from target->packages, just adds the pulled dependencies. The actual removal is done in the sortbydeps part (sortbydeps is called with list param!). This works, but it is hard to understand (imho), and it is a memleak: some pmsyncpkg_t structs magically disappeared. 2. Behaviour change (this pactest now fails) --- sync993.py --- self.description = "Choose a dependency satisfier (target-list vs. database)" sp1 = pmpkg("pkg1") sp1.depends = ["dep"] sp2 = pmpkg("pkg2") sp2.provides = ["dep"] sp3 = pmpkg("pkg3") sp3.provides = ["dep"] for p in sp1, sp2, sp3: self.addpkg2db("sync", p) self.args = "-S pkg1 pkg3" self.addrule("PACMAN_RETCODE=0") self.addrule("PKG_EXIST=pkg1") self.addrule("!PKG_EXIST=pkg2") self.addrule("PKG_EXIST=pkg3") ------------------ If you define conflict between pkg2 and pkg3, pacman will refuse to commit the transation. Before this patch, pacman always chose from target list. Bye

3 12

[pacman-dev] [PATCH] Remove pmsyncpkg_t
by Nagy Gabor 12 Apr '09

12 Apr '09

Hi! Since the patch is quite long and it depends on my working branch, I don't send it here directly. It can be found in my repo (it is a new branch): http://repo.or.cz/w/pacman-ng.git ---quote--- Remove pmsyncpkg_t pmsyncpkg_t data sructure was removed: 1. pmpkg_t.reason is used instead of pmsyncpkg_t.newreason. (The target packages come from sync repos, so we can use this field without any problems. Upgrade transaction also uses this field to store this info.) 2. pmsyncpkg_t.removes was moved to pmpkg_t.removes. This step requires careful programming, because we don't duplicate packages when we add them to trans->packages. So we modify sync pkgcache when we add this transaction-only info to our package. Hence it is important to free this list when we remove any package from the target list (remove_unresolvable, remove_conflicts, trans_free), otherwise this could confuse the new sync transactions (with non-pacman GUI). Overall, our code became ~100 line shorter, and we can call our helper functions directly on trans->packages in sync.c, we don't need to maintain parallel package lists. ---------- I simply merged pmsyncpkg_t to pmpkg_t, so the code became notable cleaner(?) and shorter, but I am not very satisfied, see the commit message. Bye

2 2

[pacman-dev] bash scripts errexit and nounset options
by Xavier 12 Apr '09

12 Apr '09

I found this page while implementing the locking system for repo-add, I stole some code from it : http://www.davidpashley.com/articles/writing-robust-shell-scripts.html It recommends using set -u and set -e, these both seem like a good idea to me. Should we try to be consistent using these options in all the bash scripts in pacman repo? Several already use set -e, but not all of them. And none seem to use set -u. While I find these two options useful, I wonder if it isn't a bad idea to introduce them now, as it could cause more harm than good. I guess that ideally, they would have been set from the beginning. Adding them now could add some non obvious problems to code that worked fine. Anyway, if we agree these flags are indeed useful, they could at least be considered when adding new scripts.

3 6

[pacman-dev] alpm_list regex matching
by leslie.polzer＠gmx.net 11 Apr '09

11 Apr '09

Hello, inlined diff (p0, hope that's not a problem) adds regex matching to alpm_list; useful for regular expressions in IgnorePkg etc. I'll leave the actual usage up to you and hope it helps somehow :) Leslie -- NEW homepage: https://viridian.dnsalias.net/~sky/homepage/ gpg --keyserver pgp.mit.edu --recv-keys DD4EBF83 Index: lib/libalpm/alpm_list.c =================================================================== RCS file: /home/cvs-pacman/pacman-lib/lib/libalpm/alpm_list.c,v retrieving revision 1.11 diff -a -p -u -r1.11 alpm_list.c --- lib/libalpm/alpm_list.c 11 Mar 2007 21:10:03 -0000 1.11 +++ lib/libalpm/alpm_list.c 14 Apr 2007 19:18:15 -0000 @@ -24,9 +24,11 @@ #include <stdlib.h> #include <string.h> #include <stdio.h> +#include <regex.h> /* libalpm */ #include "alpm_list.h" +#include "error.h" #include "util.h" /** \defgroup alpm_list functions */ @@ -458,6 +460,39 @@ int SYMEXPORT alpm_list_find_str(alpm_li return(0); } +/* Test for existence of a string in an alpm_list_t, treating the list + * elements as regular expressions +*/ +/** Is a _string_ in the list (optimization of alpm_list_find for strings) + * @param needle the string to compare + * @param haystack the list to search + * @return 1 if `needle` is found, 0 otherwise + */ + +int SYMEXPORT alpm_list_find_str_regex(alpm_list_t *haystack, const char *needle) +{ + alpm_list_t *lp = haystack; + + while(lp) { + if(lp->data) { + char* targ = lp->data; + regex_t reg; + + _alpm_log(PM_LOG_DEBUG, "matching needle '%s' against haystack item '%s'", + needle, lp->data); + + if(regcomp(&reg, targ, REG_EXTENDED | REG_NOSUB | REG_ICASE | REG_NEWLINE) != 0) { + RET_ERR(PM_ERR_INVALID_REGEX, NULL); + } + + if (regexec(&reg, needle, 0, 0, 0) == 0) { + return(1); + } + } + lp = lp->next; + } + return(0); +} /** * Calculate the items in list `lhs` that are not present in list `rhs` * @note Entries are not duplicated

3 2

[pacman-dev] help! - globbing not working in makepkg log cleanup
by Allan McRae 04 Apr '09

04 Apr '09

Hi, For some reason, this line in the clean_up() function is not working: rm -f "${pkgname}-${pkgver}-${pkgrel}-${CARCH}.log*" Printing out the value shows everything looks OK but the * does not appear to be wildcarding. To replicate, capy the abs files and "makepkg -Lf" a couple of times the "makepkg -cf". Note the pkg and src directories get cleaned up but none of the log files. Any ideas? Allan

3 3

[pacman-dev] [PATCH] makepkg: run tidy_install with no package() function
by Allan McRae 30 Mar '09

30 Mar '09

After the splitpkg implementation, the tidy_install function was not being called in PKGBUILDs with only the build() function. Signed-off-by: Allan McRae <allan(a)archlinux.org> --- scripts/makepkg.sh.in | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/scripts/makepkg.sh.in b/scripts/makepkg.sh.in index 716c5a9..349d0e0 100644 --- a/scripts/makepkg.sh.in +++ b/scripts/makepkg.sh.in @@ -1635,6 +1635,7 @@ if [ "$INFAKEROOT" -eq 1 ]; then if [ "$PKGFUNC" -eq 0 ]; then if [ "$REPKG" -eq 0 ]; then run_build + tidy_install fi else run_package -- 1.6.2.1

1 0

[pacman-dev] [PATCH] makepkg: do not bail on failure to install built package
by Allan McRae 29 Mar '09

29 Mar '09

Fixes FS#13417. Do no exit makepkg on a failure to install the built package(s). This allows clean-up to still occur. Signed-off-by: Allan McRae <allan(a)archlinux.org> --- scripts/makepkg.sh.in | 10 ++++++++-- 1 files changed, 8 insertions(+), 2 deletions(-) diff --git a/scripts/makepkg.sh.in b/scripts/makepkg.sh.in index ff2663b..716c5a9 100644 --- a/scripts/makepkg.sh.in +++ b/scripts/makepkg.sh.in @@ -1071,10 +1071,16 @@ install_package() { pkglist="${pkglist} $PKGDEST/${pkg}-${pkgver}-${pkgrel}-${CARCH}${PKGEXT}" done + local ret=0 if [ "$ASROOT" -eq 0 ]; then - sudo pacman $PACMAN_OPTS -U ${pkglist} || exit $? + sudo pacman $PACMAN_OPTS -U ${pkglist} || ret=$? else - pacman $PACMAN_OPTS -U ${pkglist} || exit $? + pacman $PACMAN_OPTS -U ${pkglist} || ret=$? + fi + + if [ $ret -ne 0 ]; then + warning "$(gettext "Failed to install built package(s).")" + return 0 fi } -- 1.6.2.1

1 0

[pacman-dev] makepkg repackage option text
by Allan McRae 26 Mar '09

26 Mar '09

Hi, Currently makepkg has a line like this: echo "$(gettext " -R, --repackage Repackage contents of pkg/ without building")" which is now true, but only to an extent... If a PKGBUILD has a package() function or split packages, then the packaging functions are run, not just a simple repacking of pkg/. So, alternative text is needed, and I need help generating it. How about: "Rerun the packaging stage only"? or does this need more description? For the man page: Rerun the packaging stage without rebuilding the package. This is useful if you forgot a depend or install file in your PKGBUILD and the build itself will not change. Comments? Allan

3 3

[pacman-dev] [PATCH] makepkg: adjust log clean-up for new filenames
by Allan McRae 25 Mar '09

25 Mar '09

The log files now have -build or -package at the end and there are separate log files for each *_package() function. Alter clean_up() to deal with this. Also, move glob outside quotes so this actually works. Signed-off-by: Allan McRae <allan(a)archlinux.org> --- scripts/makepkg.sh.in | 9 ++++++++- 1 files changed, 8 insertions(+), 1 deletions(-) diff --git a/scripts/makepkg.sh.in b/scripts/makepkg.sh.in index 95ee520..ff2663b 100644 --- a/scripts/makepkg.sh.in +++ b/scripts/makepkg.sh.in @@ -155,7 +155,14 @@ clean_up() { rm -rf "$pkgdir" "$srcdir" if [ -n "$pkgname" ]; then # Can't do this unless the BUILDSCRIPT has been sourced. - rm -f "${pkgname}-${pkgver}-${pkgrel}-${CARCH}.log*" + rm -f "${pkgbase}-${pkgver}-${pkgrel}-${CARCH}-build.log"* + if [ "$PKGFUNC" -eq 1 ]; then + rm -f "${pkgbase}-${pkgver}-${pkgrel}-${CARCH}-package.log"* + elif [ "$SPLITPKG" -eq 1 ]; then + for pkg in ${pkgname[@]}; do + rm -f "${pkg}-${pkgver}-${pkgrel}-${CARCH}-package.log"* + done + fi fi fi -- 1.6.2.1

1 0