- Pacman-dev - lists.archlinux.org

[pacman-dev] [PATCH] Allow -Qo to perform a functional 'which'
by Allan McRae 10 May '10

10 May '10

When pacman queries the ownership of an object that is not a path, it will check in the users PATH for a match. Implements FS#8798. Original-patch-by: Shankar <jatheendra(a)gmail.com> Signed-off-by: Allan McRae <allan(a)archlinux.org> --- This should address all comments I had when this was submitted a few months back. src/pacman/query.c | 57 +++++++++++++++++++++++++++++++++++++++++++++++---- 1 files changed, 52 insertions(+), 5 deletions(-) diff --git a/src/pacman/query.c b/src/pacman/query.c index 6b6a25d..038072f 100644 --- a/src/pacman/query.c +++ b/src/pacman/query.c @@ -56,10 +56,44 @@ static char *resolve_path(const char* file) return(str); } +/* check if filename exists in PATH */ +static int search_path(char *filename, size_t size, struct stat * bufptr) +{ + char *envpath, *envpathsplit, *path, fullname[PATH_MAX+1]; + int len; + + if ((envpath = getenv("PATH")) == NULL) { + return(-1); + } + if ((envpath = envpathsplit = strdup(envpath)) == NULL) { + return(-1); + } + + while ((path = strsep(&envpathsplit, ":")) != NULL) { + len = strlen(path); + + /* strip the trailing slash if one exists */ + while(path[len - 1] == '/') { + path[--len] = '\0'; + } + + snprintf(fullname, sizeof(fullname), "%s/%s", path, filename); + + if(stat(fullname, bufptr) == 0) { + strncpy(filename, fullname, size); + free(envpath); + return(0); + } + } + free(envpath); + return(-1); +} + static int query_fileowner(alpm_list_t *targets) { int ret = 0; alpm_list_t *t; + size_t len=PATH_MAX+1; /* This code is here for safety only */ if(targets == NULL) { @@ -67,19 +101,31 @@ static int query_fileowner(alpm_list_t *targets) return(1); } + char *filename = calloc(len, sizeof(char)); + for(t = targets; t; t = alpm_list_next(t)) { int found = 0; - char *filename = alpm_list_getdata(t); + strncpy(filename, alpm_list_getdata(t), len); char *bname, *dname, *rpath; const char *root; struct stat buf; alpm_list_t *i, *j; if(lstat(filename, &buf) == -1) { - pm_fprintf(stderr, PM_LOG_ERROR, _("failed to read file '%s': %s\n"), - filename, strerror(errno)); - ret++; - continue; + /* if it is not a path but a program name, then check in PATH */ + if(strchr(filename, '/') == NULL) { + if(search_path(filename, len, &buf) == -1) { + pm_fprintf(stderr, PM_LOG_ERROR, _("failed to read file '%s': %s\n"), + filename, strerror(errno)); + ret++; + continue; + } + } else { + pm_fprintf(stderr, PM_LOG_ERROR, _("failed to read file '%s': %s\n"), + filename, strerror(errno)); + ret++; + continue; + } } if(S_ISDIR(buf.st_mode)) { @@ -140,6 +186,7 @@ static int query_fileowner(alpm_list_t *targets) free(rpath); } + free(filename); return ret; } -- 1.7.0.1

4 6

Re: [pacman-dev] [PATCH] [bash_completion] Undeclared local vars/filenames with spaces/other
by Xavier Chantry 09 May '10

09 May '10

On Sat, May 8, 2010 at 10:31 AM, Xavier Chantry <shiningxc(a)gmail.com> wrote: > > > Attached list of changes between maint and master, + the new patch > rebased on top of master. > > Of course if you want to keep working on it, you're very welcome to do so :) > What mostly annoyed me with pacman bash completion is the lag/delay on completing packages in pacman -S. That clearly feels better after your patch, but I've noticed a funny problem and different behavior. I am sitting in my pacman git tree, and wanted to reinstall pacman : $ pacman -S pac<tab> after your patch : $ sudo pacman -S pac pacbuild pacman-contrib pacman-mirrorlist paco pacpan pacupdate before : $ sudo pacman -S pac pacbuild pacman/ pacman-contrib pacman-mirrorlist paco pacpan pacupdate So actually I couldnt complete pacman with your patch, it proposed me pacman- directly. The old behavior didnt propose me pacman either, but since it hesitated between pacman/ and pacman- , the completion stopped on pacman which is the good package name :) And this weird behavior is because I had a pacman dir in my PWD : $ ls -ld pacman/ drwxr-xr-x 2 xavier users 4096 Sep 23 2009 pacman/

2 1

[pacman-dev] Provides/replaces fields and backup files
by Nezmer 09 May '10

09 May '10

Hi, Suppose package B replaces package A. Both packages share the same content with minor differences. They both have many backup files with the same initial content. When package B replaces package A. Modified backup files will move to <backupfile.pacsave>. What If pacman could check for initial content of the backup files for both packages and If the content is the same, treat those backup files as If the process is a package upgrade not a package replacement. This is a real-life example. Packages A and B contain files like "/etc/passwd" and "/etc/group". So not moving the files to ".pacsave" adds more than a convenience for the lazy. My C skills are next-to-zero so unfortunately I can't help coding this. Do you consider this a valid minor bug? Should I report this to flyspray?

3 2

[pacman-dev] [PATCH] [bash_completion] Undeclared local vars/filenames with spaces/other
by Andres Perera 08 May '10

08 May '10

Fix: 3 Undeclared local vars with common enough names to warrant breakage Performance issues with _pacman trying to replicate /usr/bin/pacman with find and other slow tools. Performance issues with expanding an array (with sometimes hundreds of items) over three times. Expanding said array to remove already completed entries had the side effect of braking filenames with spaces and or \n. The full description of fixes are already posted at: http://bugs.archlinux.org/task/16630#comment55779 along with time diffs. Signed-off-by: Andres P <andres87p(a)gmail.com> --- contrib/bash_completion | 535 ++++++++++++++++++----------------------------- 1 files changed, 202 insertions(+), 333 deletions(-) diff --git a/contrib/bash_completion b/contrib/bash_completion index 62e5bc9..65051f5 100644 --- a/contrib/bash_completion +++ b/contrib/bash_completion @@ -1,365 +1,234 @@ -# vim: set ft=sh ts=2 sw=2 et: -# file: /etc/bash_completion.d/pacman - -# Bash completion for pacman -# Original: Manolis Tzanidakis <mtzanidakis(a)freemail.gr> +# pacman/makepkg completion by Andres Perera <andres87p gmail> # -# Distributed under the terms of the GNU General Public License, v2 or later. +# Distributed under the terms of the GNU General Public License v3 or +# later. # - -## initial functions - -rem_selected () -{ - # (Adapted from bash_completion by Ian Macdonald <ian(a)caliban.org>) - # This removes any options from the list of completions that have - # already been specified on the command line. - COMPREPLY=($(/bin/echo "${COMP_WORDS[@]}" | \ - (while read -d ' ' i; do - [ "${i}" == "" ] && continue - # flatten array with spaces on either side, - # otherwise we cannot grep on word boundaries of - # first and last word - COMPREPLY=" ${COMPREPLY[@]} " - # remove word from list of completions - COMPREPLY=(${COMPREPLY/ ${i%% *} / }) +# Local variables: common core cur glob list long m o prev query r +# remove s short sync upgrade w +# +# vim: ft=sh sts=2 sw=2 et: + +# Removes packages/files already present in the line +_arch_rem_selected() { + local w r + + for (( w=0; w<${#COMP_WORDS[@]}-1; w++)); do + # Don't touch operands + [[ ${COMP_WORDS[w]} == -* ]] && continue + for r in "${!COMPREPLY[@]}"; do + # Only deal with full matches + if [[ ${COMP_WORDS[w]} == ${COMPREPLY[r]} ]]; then + unset 'COMPREPLY[r]'; break + fi done - /bin/echo ${COMPREPLY[@]}))) - return 0 -} - -_available_repos () -{ - COMPREPLY=( $( compgen -W "$(/bin/grep '\[' /etc/pacman.conf | /bin/grep -v -e 'options' -e '^#' | tr -d '[]' )" -- $cur ) ) -} - -_installed_pkgs () -{ - local installed_pkgs - installed_pkgs=$( /bin/ls /var/lib/pacman/local/ ) - COMPREPLY=( $( compgen -W "$( for i in $installed_pkgs; do /bin/echo ${i%-*-*}; done )" -- $cur ) ) + done } -_available_pkgs () -{ - #find balks easilly on a find /foo/*/* type dir, especially one like - # /var/lib/pacman/*/* - # This little change-up removes the find *and* only uses enabled repos - local available_pkgs - local enabled_repos - enabled_repos=$( /bin/grep '\[' /etc/pacman.conf | /bin/grep -v -e 'options' -e '^#' | tr -d '[]' ) - available_pkgs=$( for r in $enabled_repos; do /bin/echo /var/lib/pacman/sync/$r/*; done ) - COMPREPLY=( $( compgen -W "$( for i in $available_pkgs; do j=${i##*/}; echo ${j%-*-*}; done )" -- $cur ) ) -} +# makepkg purge functions +_makepkg_deselect() { + local o -_installed_groups () -{ - local installed_groups - installed_groups=$( /bin/find /var/lib/pacman/local -name desc -exec /bin/sed -ne '/%GROUPS%/,/^$/{//d; p}' {} \; | /bin/sort -u ) - COMPREPLY=( $( compgen -W "$( for i in $installed_groups; do /bin/echo ${i%-*-*}; done )" -- $cur ) ) + if [[ $1 == l ]]; then + for o in "${!long[@]}"; do + if [[ ${COMP_WORDS[w]} == ${long[o]} ]]; then + unset 'long[o]'; break + fi + done + else + for o in "${!short[@]}"; do + if [[ ${COMP_WORDS[w]} == -*([^- ])${short[o]}* ]]; then + unset 'short[o]' + fi + done + fi } -_available_groups () -{ - #find balks easilly on a find /foo/*/* type dir, especially one like - # /var/lib/pacman/*/* - # This little change-up removes the find *and* only uses enabled repos - local available_groups - local enabled_repos - enabled_repos=$( /bin/grep '\[' /etc/pacman.conf | /bin/grep -v -e 'options' -e '^#' | tr -d '[]' ) - available_groups=$( for r in $enabled_repos; do /bin/sed '/%GROUPS%/,/^$/{//d; p}' /var/lib/pacman/sync/$r/*/desc | /bin/sort -u; done ) - COMPREPLY=( $( compgen -W "$( for i in $available_groups; do /bin/echo ${i%-*-*}; done )" -- $cur ) ) +_makepkg_count_words() { + local w + + for (( w=0; w<${#COMP_WORDS[@]}; w++)); do + [[ ${COMP_WORDS[w]} == @(-*([^- ])[hC]*|--@(help|cleancache)) ]] && return 0 + _makepkg_deselect "$1" + done } -## makepkg completion - -_makepkg () -{ - local cur prev +# makepkg main +_makepkg() { COMPREPLY=() - cur=${COMP_WORDS[COMP_CWORD]} - prev=${COMP_WORDS[COMP_CWORD-1]} - case "$prev" in - -p) - _filedir - return 0 - ;; - --help|--cleancache) - COMPREPLY='' - return 0 - ;; - esac - - if [[ "$cur" == -* ]]; then - COMPREPLY=( $( compgen -W '\ - -A --ignorearch \ - -b --builddeps \ - -c --clean \ - -C --cleancache \ - -d --nodeps \ - -e --noextract \ - -f --force \ - -g --geninteg \ - -h --help \ - -i --install \ - -L --log \ - -m --nocolor \ - -o --nobuild \ - -p \ - -r --rmdeps \ - -s --syncdeps \ - --asroot \ - --source \ - --noconfirm \ - --noprogressbar' -- $cur ) ) + local prev cur short long parse glob + prev=${COMP_WORDS[COMP_CWORD-1]} + cur=`_get_cword` + + short=( A C L R c d e f g h i m o p r s ) + + long=( --allsource --asroot --clean --cleancache --config --force + --geninteg --help --holdver --ignorearch --install --log + --nobuild --nocolor --noconfirm --nodeps --noextract + --noprogressbar --repackage --rmdeps --skipinteg --source + --syncdeps ) + + if (( COMP_CWORD == 1 )) && [[ $cur != -[^-]* ]]; then + # First word check + COMPREPLY=($(compgen -W "$(echo "${short[@]/#/-}" "${long[@]}")" -- "$cur")) + elif [[ $prev == @(--config|-*([^- ])p*) ]]; then + # Complete filenames for selected options + _arch_rem_selected + return 0 + else + # Generate the list depending on current word + case $cur in + -+($(IFS='|'; echo "${short[*]}"))) + _makepkg_count_words s + COMPREPLY=($(compgen -W "$(echo "${short[@]/#/$cur}")" -- "$cur")) + ;; + -*) + # If $cur isn't a combination option, fall back to displaying + # long options only. These are the prime target for completion + _makepkg_count_words l + COMPREPLY=($(compgen -W "$(echo "${long[@]}")" -- "$cur")) + ;; + esac fi - - rem_selected } -complete -o default -F _makepkg makepkg -## pacman completion - -_instring () -{ - str="${1}" - shift 1 - for c in "${@}"; do - if [ $(/bin/expr index "${str}" "${c}") -gt 0 ]; then - return 0 - fi - done - return 1 +# pacman packages +_pacman_pkg() { + COMPREPLY=($(compgen -W "$(if [[ $2 ]]; then + command pacman -"$1" |\ + command sed 's, .*,,' |\ + command sort -u + else + command pacman -"$1" + fi)" -- "$cur" )) } -_pacman () -{ - local a arg toparse op mod cur +# pacman main +_pacman() { COMPREPLY=() - # This argument parsing is done so we can check for flag existance later - # right now it's a tad crappy, but does the job - for (( i=1; i < ${#COMP_WORDS[@]}-1; i++ )); do - a=${COMP_WORDS[i]} - arg="${a:0:2}" - toparse="${a:2}" + local prev cur query remove sync upgrade core common list w m o s=() + prev=${COMP_WORDS[COMP_CWORD-1]} + cur=`_get_cword` - case "${arg}" in - -@(U|R|S|Q|h|V)) - op="${arg/-}" - mod="${mod}${a:2}" - ;; - --) - arg="${a:2}" - case "${arg}" in - remove) op="R" ;; - upgrade) op="U" ;; - query) op="Q" ;; - sync) op="S" ;; - help) op="h" ;; - version) op="V" ;; - verbose) mod="${mod}v" ;; - root) mod="${mod}r" ;; - dbpath) mod="${mod}b" ;; - nodeps) mod="${mod}d" ;; - force) mod="${mod}f" ;; - groups) mod="${mod}g" ;; - info) mod="${mod}i" ;; - list) mod="${mod}l" ;; - print-uris) mod="${mod}p" ;; - search) mod="${mod}s" ;; - sysupgrade) mod="${mod}u" ;; - upgrades) mod="${mod}u" ;; - downloadonly) mod="${mod}w" ;; - refresh) mod="${mod}y" ;; - changelog) mod="${mod}c" ;; - deps) mod="${mod}d" ;; - explicit) mod="${mod}e" ;; - unrequired) mod="${mod}t" ;; - foreign) mod="${mod}m" ;; - owns) mod="${mod}o" ;; - file) mod="${mod}p" ;; - search) mod="${mod}s" ;; - upgrades) mod="${mod}u" ;; - cascade) mod="${mod}c" ;; - check) mod="${mod}k" ;; - dbonly) mod="${mod}k" ;; - nosave) mod="${mod}n" ;; - recursive) mod="${mod}s" ;; - unneeded) mod="${mod}u" ;; - esac ;; - *) toparse="${a}" ;; - esac + # These are split in a subshell with IFS=' '. No point in listing + # them by \n since compgen -W can't handle spaces or new lines in + # operands + query=( '--changelog --check --deps --explicit --file --foreign --groups --info + --list --owns --quiet --search --unrequired --upgrades' 'c e g i k l m o p q s t u' ) - arglen=$(( ${#toparse}-1 )) - for c in $(/bin/seq 0 "${arglen}"); do - arg=${toparse:$c:1} - [ "${arg}" != "-" ] && mod="${mod}${arg}" - done - done + remove=( '--cascade --dbonly --nodeps --nosave --recursive --unneeded' 'c k n s u' ) - cur=${COMP_WORDS[COMP_CWORD]} + sync=( '--asdeps --asexplicit --clean --downloadonly --force --groups --ignore --ignoregroup + --info --list --needed --nodeps --print-uris --quiet --refresh --search --sysupgrade' + 'c f g i l p q s u w y' ) - if [ $COMP_CWORD -eq 1 ] && [[ "$cur" == -* ]]; then - COMPREPLY=( $( compgen -W '\ - -h --help \ - -Q --query \ - -R --remove \ - -S --sync \ - -U --upgrade \ - -V --version \ - ' -- $cur ) ) - rem_selected - return 0 - fi + upgrade=( '--asdeps --asexplicit --force --nodeps' f ) - if [[ "$cur" == -* ]]; then - case "${op}" in - U) - COMPREPLY=( $( compgen -W '\ - --asdeps \ - --asexplicit \ - -d --nodeps \ - -f --force \ - -h --help \ - --config \ - --logfile \ - --noconfirm \ - --noprogressbar \ - --noscriptlet \ - -v --verbose \ - -r --root \ - -b --dbpath \ - --cachedir \ - ' -- $cur ) ) - return 0 - ;; - R) - COMPREPLY=( $( compgen -W '\ - -c --cascade \ - -d --nodeps \ - -h --help \ - -k --dbonly \ - -n --nosave \ - -s --recursive \ - -u --unneeded \ - --config \ - --logfile \ - --noconfirm \ - --noprogressbar \ - --noscriptlet \ - -v --verbose \ - -r --root \ - -b --dbpath \ - --cachedir \ - ' -- $cur ) ) - return 0 - ;; - S) - COMPREPLY=( $( compgen -W '\ - --asdeps \ - --asexplicit \ - -c --clean \ - -d --nodeps \ - -f --force \ - -g --groups \ - -h --help \ - -i --info \ - -l --list \ - -p --print-uris \ - -s --search \ - -u --sysupgrade \ - -w --downloadonly \ - -y --refresh \ - --needed \ - --ignore \ - --ignoregroup \ - --config \ - --logfile \ - --noconfirm \ - --noprogressbar \ - --noscriptlet \ - -v --verbose \ - -r --root \ - -b --dbpath \ - --cachedir \ - ' -- $cur ) ) - return 0 - ;; - Q) - COMPREPLY=( $( compgen -W '\ - -c --changelog \ - -d --deps \ - -e --explicit \ - -g --groups \ - -h --help \ - -i --info \ - -k --check \ - -l --list \ - -m --foreign \ - -o --owns \ - -p --file \ - -s --search \ - -t --unrequired \ - -u --upgrades \ - --config \ - --logfile \ - --noconfirm \ - --noprogressbar \ - --noscriptlet \ - -v --verbose \ - -r --root \ - -b --dbpath \ - --cachedir \ - ' -- $cur ) ) - return 0 - ;; - esac - rem_selected + common=( '--cachedir --config --dbpath --debug --help --logfile --noconfirm + --noprogressbar --noscriptlet --root --verbose' 'b d h r v' ) + + core=( --query --remove --sync --upgrade --version -Q -R -S -U -V ) + + if (( COMP_CWORD == 1 )) && [[ $cur != -[^-]* ]]; then + COMPREPLY=($(compgen -W "$(echo "${core[@]}") '--help' '-h'" -- "$cur")) + return 0 + elif [[ $prev == @(-*([^- ])[br]*|--@(config|logfile|root|dbpath|cachedir)) ]]; then + _arch_rem_selected + return 0 else - case "${op}" in - U) - COMPREPLY=( $( compgen -d -- "$cur" ) \ - $( compgen -f -X '!*.pkg.tar.gz' -- "$cur" ) ) + # Main option check + for m in 'Q query' 'R remove' 'S sync' 'U upgrade' 'null'; do + if [[ $m == null ]]; then + # Go back to core if null + COMPREPLY=($(compgen -W "$(echo "${core[@]}")" -- "$cur")) return 0 - ;; - h|V) - COMPREPLY='' - return 0 - ;; - Q) - if _instring $mod g; then - _installed_groups - elif _instring $mod o; then - COMPREPLY=( $( compgen -d -- "$cur" ) \ - $( compgen -f -- "$cur" ) ) - elif _instring $mod p; then - COMPREPLY=( $( compgen -d -- "$cur" ) \ - $( compgen -f -X '!*.pkg.tar.gz' -- "$cur" ) ) - elif _instring $mod u; then - COMPREPLY='' + fi + # Break on match + [[ $COMP_LINE == @(* -*([^- ])${m% *}*|* --${m#* } *) ]] && break + done + fi + + # See if $cur is an operand before calling _pacman_pkg + case $cur in + + # Check if it's a combination option first + -"${m% *}"*) + # Generate the list from the option arrays. eval is being using on + # possible matches between query, remove, sync, upgrade and common + # arrays; completely sanitized input + eval "list=($(command sed 's,.*,\${&[1]},' <<<"${m#* }"; IFS=' '; echo ${common[1]}))" + # Parse previous arguments + for (( w=0; w<${#COMP_WORDS[@]}; w++)); do + if [[ ${COMP_WORDS[w]} == -*([^-])[hV]* ]]; then return 0 - else - _installed_pkgs fi - return 0 + for o in "${!list[@]}"; do + if [[ ${COMP_WORDS[w]} == -*([^-])*${list[o]}* ]]; then + unset 'list[o]' + fi + done + done + # Allow special double letters + case ${m% *} in + Q) [[ $cur == *([^i])i*([^i]) ]] && s+=(i) ;; + R) [[ $cur == *([^s])s*([^s]) ]] && s+=(s) ;; + S) [[ $cur == *([^c])c*([^c]) ]] && s+=(c) + [[ $cur == *([^u])u*([^u]) ]] && s+=(u) + [[ $cur == *([^y])y*([^y]) ]] && s+=(y) ;; + esac + list=("${list[@]/#/$cur}" "${s[@]/#/$cur}") + ;; + + # Use long array instead + -*) + eval "list=($(command sed 's,.*,\${&[0]},' <<<"${m#* }"; IFS=' '; echo ${common[0]}))" + for (( w=0; w<${#COMP_WORDS[@]}-1; w++)); do + if [[ ${COMP_WORDS[w]} == --@(help|version) ]]; then + return 0 + fi + for o in "${!list[@]}"; do + if [[ ${COMP_WORDS[w]} == ${list[o]} ]]; then + unset 'list[o]'; break + fi + done + done + ;; + + # The word isn't an operand; use pacman or _filedir + *) + case ${m% *} in + # The weak extglobs are due to a previous match already + # confirming validity + Q) + case $COMP_LINE in + @(* -*([^- ])g*|* --groups *)) _pacman_pkg Qgq sort ;; + @(* -*([^- ])o*|* --owns *)) ;; + @(* -*([^- ])p*|* --file *)) _filedir 'pkg.tar.gz' ;; + @(* -*([^- ])u*|* --upgrades *)) _pacman_pkg Quq ;; + *) _pacman_pkg Qq ;; + esac ;; - R) - _installed_pkgs - return 0 + R) + _pacman_pkg Qq ;; - S) - if _instring $mod l; then - _available_repos - else - _available_pkgs - fi - return 0 + S) + case $COMP_LINE in + @(* -*([^- ])l*|* --list *)) _pacman_pkg Sl sort ;; + @(* -*([^- ])g*|* --groups *)) _pacman_pkg Sg ;; + *) _pacman_pkg Slq ;; + esac ;; - esac - fi - - rem_selected + U) + _filedir 'pkg.tar.gz' + ;; + esac + _arch_rem_selected + return 0 + ;; + esac + COMPREPLY=($(compgen -W "$(echo "${list[@]}")" -- "$cur")) } -complete -o filenames -F _pacman pacman + +complete -F _makepkg -o default -o filenames -o plusdirs makepkg +complete -F _pacman -o default -o filenames -o plusdirs pacman -- 1.6.6

4 6

[pacman-dev] Release plans
by Andres P 08 May '10

08 May '10

On 05/05/10 at 18:31:22 CEST, Dan McGee <dpmcgee at gmail.com> wrote: >>> 8*?) Tidy up bash_completion stuff? (FS#16630) > The poster disappeared and he tried to sneak a license change in > there, so that just hit a bit of a roadblock. I'd like to improve this > so I might just sit down and hack at it myself. "Tried to sneak a license change"? >From your own ml: > On Tue, Jan 19, 2010 at 11:48 PM, Andres Perera <andres...(a)gmail.com> wrote: >> On Wed, Jan 20, 2010 at 1:07 AM, Dan McGee <dpmc...(a)gmail.com> wrote: >>> Is this strictly necessary? You are changing one piece of code in the >>> entire codebase to require v3 or later and as weird as it sounds, I am >>> not going to let that fly. I won't be merging this unless it is v2 or >>> you give some darn good reasons why we should move to v3. >> >> Not really necessary. You can change it to v2, or I can do that if it >> requires a commit from me. > > Nah that's good enough for me. I'll try to give this whole thing a > workout in the next day or two and see if I can come up with anything > else that we might want to tweak, however. > > -Dan http://www.mail-archive.com/pacman-dev@archlinux.org/msg04127.html Maybe I was a little harsh? Anyway, I don't care if you put a random bash script I wrote ages ago under GPL, Public Domain, BSD... lol, Andres P

2 1

Re: [pacman-dev] [arch-general] Package signing
by Linas 08 May '10

08 May '10

I wrote about this topic ~1 month ago. You don't need PKCis or distribute the keyrings themselves. GPG supports transitive trust. The pacman keyring would be installed by default trusting on whatever keys a pacman root signature has signed (there could also be a different master key for community developers). The basic idea here is that you are not trusting the repository, but the individuals themselves. The master key -which can be kept offline and is only used when a developer joins/part- provides a basic default (people we generally trust) but a power user could reconfigure it to not accept packages signed by Pierre, because he distrusts him :), or he can add additional trusted people (a much more likely scenario) by just adding that person key to its keyring. A developer asks you to try a beta package -> You need to do nothing since you already have its key. Fetching from personal repo from a developer -> You need to do nothing since you already have its key. That really smart guy from the forum whose advice you follow blindly. -> Add its key. Packages built by you -> Add your own key. Sysadmin -> Add its key to the computers he administers. Paranoic sysadmin -> Add its key and remove the developers one. Then resign all packages after verifying that they work. I think this is more or less what Florian Pritz proposes, althouugh I think he was thinking in terms of the user gpg, while I would go using a different, pacman gpg. Denis A. Altoé Falqueto wrote: > The central point is the management of the keyring. > > We could have a single key that all developers share, but this is > troublesome as it would require some trusted channel to exchange the > password (the first time and in the event of it changing). This > approach is what others big distributions do, but the have a different > structure than ours. > With this structure the developers never share the keys. You could even remove the master key and consider someone a developer when n developers signed him. But I find the master key method clearer. > We could have a keyring with all the trusted keys of the developers, > but this could bring some problems as the keyring should be very well > synchronized with the official one. Including a rule to upgrade the > keyring package before others, as pacman. This requires intervention > on user`s systems, so I think that it should be avoided, if possible. > There are two cases where the keyring could change due to distro changes. a) Add a new developer. b) Remove a developer. In case (a), pacman would find an unknown signature, an download [on user approval] the public key from the key servers. Case (b) is trickier, since no matter how it is done, I think there will always be a race condition on a roque developer. Luckily, the shouts of such case would make it difficult to have people holding outdated keys. I would probably make the key updating part of pacman -y to ensure that users are not using a revoked key. > My idea tries to give the best of both worlds. The management of > trusted devs would be made by a few admins and users would just be > exposed to an official Arch key that would change just a few times, if > it is not compromised. > Your approach is more insecure, since you are removing the developer key (end-to-end signing) and putting instead a generic one. With the other schema, in case a developer key was compromised, you would only need to untrust that key, not the whole master key which signs everything. Plus, your key is used continously, so it is more exposed (the signing server is compromised, it signs a fake request...). IMHO trusting the individuals also facilitates forking. A derivative distribution could share arch packages by trusting the same devvelopers. With a master key owning everything, the child distro would need to have a second key resigning everything. __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

8 31

[pacman-dev] [PATCH 2/2] Switch Contributor line with Maintainer line.
by David Campbell 06 May '10

06 May '10

When someone is creating a new PKGBUILD he will most likely be the maintainer not a contributor. --- PKGBUILD-split.proto | 2 +- PKGBUILD.proto | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/PKGBUILD-split.proto b/PKGBUILD-split.proto index 53a3e8f..f549cfd 100644 --- a/PKGBUILD-split.proto +++ b/PKGBUILD-split.proto @@ -3,7 +3,7 @@ # see 'man PKGBUILD'. NOTE: Please fill out the license field for your package! # If it is unknown, then please put 'unknown'. -# Contributor: Your Name <youremail(a)domain.com> +# Maintainer: Your Name <youremail(a)domain.com> pkgname=('pkg1' 'pkg2') pkgbase="" pkgver=VERSION diff --git a/PKGBUILD.proto b/PKGBUILD.proto index 93da0fc..59b9acb 100644 --- a/PKGBUILD.proto +++ b/PKGBUILD.proto @@ -3,7 +3,7 @@ # NOTE: Please fill out the license field for your package! If it is unknown, # then please put 'unknown'. -# Contributor: Your Name <youremail(a)domain.com> +# Maintainer: Your Name <youremail(a)domain.com> pkgname=NAME pkgver=VERSION pkgrel=1 -- 1.7.0.3

2 1

[pacman-dev] [PATCH] compute package download size outside _alpm_sync_prepare
by Jonathan Conder 06 May '10

06 May '10

Thanks to Dan for pointing out a mistake I made. --- lib/libalpm/db.h | 3 ++- lib/libalpm/sync.c | 14 +++++++++++++- 2 files changed, 15 insertions(+), 2 deletions(-) diff --git a/lib/libalpm/db.h b/lib/libalpm/db.h index 9b78ad4..1851b5c 100644 --- a/lib/libalpm/db.h +++ b/lib/libalpm/db.h @@ -34,7 +34,8 @@ typedef enum _pmdbinfrq_t { INFRQ_FILES = (1 << 3), INFRQ_SCRIPTLET = (1 << 4), INFRQ_DELTAS = (1 << 5), - /* ALL should be sum of all above */ + INFRQ_DSIZE = (1 << 6), + /* ALL should be info stored in the package or database */ INFRQ_ALL = 0x3F } pmdbinfrq_t; diff --git a/lib/libalpm/sync.c b/lib/libalpm/sync.c index 4996920..13d39aa 100644 --- a/lib/libalpm/sync.c +++ b/lib/libalpm/sync.c @@ -356,6 +356,7 @@ static int compute_download_size(pmpkg_t *newpkg) off_t size = 0; if(newpkg->origin == PKG_FROM_FILE) { + newpkg->infolevel |= INFRQ_DSIZE; newpkg->download_size = 0; return(0); } @@ -392,6 +393,7 @@ static int compute_download_size(pmpkg_t *newpkg) _alpm_log(PM_LOG_DEBUG, "setting download size %jd for pkg %s\n", (intmax_t)size, alpm_pkg_get_name(newpkg)); + newpkg->infolevel |= INFRQ_DSIZE; newpkg->download_size = size; return(0); } @@ -649,6 +651,9 @@ cleanup: */ off_t SYMEXPORT alpm_pkg_download_size(pmpkg_t *newpkg) { + if(!(newpkg->infolevel & INFRQ_DSIZE)) { + compute_download_size(newpkg); + } return(newpkg->download_size); } @@ -850,7 +855,14 @@ int _alpm_sync_commit(pmtrans_t *trans, pmdb_t *db_local, alpm_list_t **data) if(files) { EVENT(trans, PM_TRANS_EVT_RETRIEVE_START, current->treename, NULL); - if(_alpm_download_files(files, current->servers, cachedir)) { + errors = _alpm_download_files(files, current->servers, cachedir); + + for(j = trans->add; j; j = j->next) { + j->infolevel &= ~INFRQ_DSIZE; + j->download_size = 0; + } + + if (errors) { _alpm_log(PM_LOG_WARNING, _("failed to retrieve some files from %s\n"), current->treename); if(pm_errno == 0) { -- 1.7.1

3 4

[pacman-dev] [PATCH] compute package download size outside _alpm_sync_prepare
by Jonathan Conder 05 May '10

05 May '10

Hi Here is a patch to resolve FS#18769. I created a new INFRQ/infolevel to describe package download size, so that it could be lazily calculated like most of the other package info. This differs slightly from most of the other infolevels, which correspond to information actually read by _alpm_db_read and alpm_pkg_load. However, I felt it better to do it this way than setting the download size to -1 for invalid, as this would make it the only field in pmpkg_t that needs to be initialized to something other than 0. It is also possible that the download size will be reported incorrectly if a package download size is read, and then that same package is downloaded with alpm_fetch_pkgurl. However, this is better than the present situation where download sizes aren't reported at all, and furthermore this is unlikely since alpm_fetch_pkgurl would probably not be used to download a package that is in the repositories. Thanks for your time, Jonathan

3 4

[pacman-dev] [GIT] The official pacman repository branch, master, updated. v3.3.3-166-gd485c03
by dan＠archlinux.org 05 May '10

05 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 d485c0370f554758a28eff2cc32b2cc9c72a4604 (commit) via 9528c36452a6ca8b3718ef5895f31f46c16b17fa (commit) via f9fa822ccb0a077cf4cfa2cbbdac773187b88e20 (commit) via ac9dde072c56b9dbc3fe515dc3702c2c992f492f (commit) via 6b6eb6345bb3d9f59d967144968328c9a4b7873c (commit) via eb6af031ec83c4c0c1d7e12d1cc8f585197c025f (commit) via 4f80993933a3feee68758c27c58e9cf5d84bf20e (commit) via 2a6f3f06528149200e4fc8996c0c1e6d2bfa5480 (commit) via 8a6bbed98e11be3f619ef31e9e753d31081d11ee (commit) from 90aca75cb92c12419f2998cf54caa8ef5fe59f99 (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 d485c0370f554758a28eff2cc32b2cc9c72a4604 Author: Dan McGee <dan(a)archlinux.org> Date: Wed May 5 11:18:27 2010 -0500 Improve documentation of -k/--dbonly We had the long option wrong in some places and its behavior wasn't documented at all with regards to -U/--upgrade. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 9528c36452a6ca8b3718ef5895f31f46c16b17fa Author: Dan McGee <dan(a)archlinux.org> Date: Wed May 5 10:59:20 2010 -0500 Add two new pactests for --dbonly It caught me by surprise that: 1. These weren't being tested at all 2. The --dbonly combined with -U not only "works" but is also completely undocumented. It also has some weird behavior on install vs. upgrade that may need addressing. Add some tests which will hopefully provoke some discussion. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit f9fa822ccb0a077cf4cfa2cbbdac773187b88e20 Author: Nagy Gabor <ngaba(a)bibl.u-szeged.hu> Date: Tue Mar 16 19:40:29 2010 +0100 Remove unused 'z' option from getopt_long's optstring In addition, I permuted shortopts to make it more readable. Signed-off-by: Nagy Gabor <ngaba(a)bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit ac9dde072c56b9dbc3fe515dc3702c2c992f492f Author: Nagy Gabor <ngaba(a)bibl.u-szeged.hu> Date: Wed May 20 21:46:23 2009 +0200 Introduce -D, --database The request of FS#12950 is implemented. On the backend side, I introduced a new function, alpm_db_set_pkgreason(), to modify the install reason of a package in the local database. On the front-end side, I introduced a new main operation, -D/--database, which has two options, --asdeps and --asexplicit. I documented this in pacman manual. I've created two pactests to test -D: database001.py and database002.py. Signed-off-by: Nagy Gabor <ngaba(a)bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 6b6eb6345bb3d9f59d967144968328c9a4b7873c Author: Dan McGee <dan(a)archlinux.org> Date: Tue May 4 23:43:43 2010 -0500 Fix up the cross-compilation patch library lookup Don't explicitly add things to the list that might not need to be there, and get the fallback list of libraries correct. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit eb6af031ec83c4c0c1d7e12d1cc8f585197c025f Author: Dan McGee <dan(a)archlinux.org> Date: Tue May 4 23:42:01 2010 -0500 Build vercmp without needing link to libalpm Include the object file directly from the libalpm version comparison code as it is the only thing we need. This drops the dependency of vercmp on libalpm and all of the stuff we know it drags in. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 4f80993933a3feee68758c27c58e9cf5d84bf20e Author: Dan McGee <dan(a)archlinux.org> Date: Tue May 4 23:38:52 2010 -0500 Remove call to function logger It isn't really necessary here and it helps us get rid of some link pollution so we can have a slim vercmp binary. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 2a6f3f06528149200e4fc8996c0c1e6d2bfa5480 Author: Dan McGee <dan(a)archlinux.org> Date: Mon Apr 5 14:56:32 2010 -0500 Move vercmp code into a separate file This will facilitate using this object file on its own in the vercmp tool which will be done in a future commit. The net impact on the generated binaries should not be noticeable after this commit. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 8a6bbed98e11be3f619ef31e9e753d31081d11ee Author: Dan McGee <dan(a)archlinux.org> Date: Tue Apr 27 13:08:09 2010 -0500 Update PKGBUILD.proto to include a package() function Relevant after we deprecated `makepkg -R` without a package() function being present in the PKGBUILD. Signed-off-by: Dan McGee <dan(a)archlinux.org> ----------------------------------------------------------------------- Summary of changes: PKGBUILD.proto | 3 + configure.ac | 4 +- doc/pacman.8.txt | 21 ++++- lib/libalpm/Makefile.am | 3 +- lib/libalpm/alpm.h | 16 ++-- lib/libalpm/db.c | 38 ++++++++ lib/libalpm/package.c | 171 ------------------------------------ lib/libalpm/version.c | 195 ++++++++++++++++++++++++++++++++++++++++++ pactest/tests/database001.py | 11 +++ pactest/tests/database002.py | 11 +++ pactest/tests/database010.py | 17 ++++ pactest/tests/database011.py | 17 ++++ pactest/tests/database012.py | 31 +++++++ src/pacman/Makefile.am | 1 + src/pacman/conf.h | 3 +- src/pacman/database.c | 90 +++++++++++++++++++ src/pacman/pacman.c | 24 ++++-- src/pacman/pacman.h | 2 + src/pacman/util.c | 2 + src/util/Makefile.am | 2 +- src/util/vercmp.c | 6 +- 21 files changed, 475 insertions(+), 193 deletions(-) create mode 100644 lib/libalpm/version.c create mode 100644 pactest/tests/database001.py create mode 100644 pactest/tests/database002.py create mode 100644 pactest/tests/database010.py create mode 100644 pactest/tests/database011.py create mode 100644 pactest/tests/database012.py create mode 100644 src/pacman/database.c hooks/post-receive -- The official pacman repository

1 0