- Pacman-dev - lists.archlinux.org

[pacman-dev] [PATCH] bash-completion: fix leaking "files" array into shell environment
by Eli Schwartz 03 Sep '16

03 Sep '16

Signed-off-by: Eli Schwartz <eschwartz93(a)gmail.com> --- This time with a signoff... contrib/bash_completion.in | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/contrib/bash_completion.in b/contrib/bash_completion.in index 6f99681..06963c4 100644 --- a/contrib/bash_completion.in +++ b/contrib/bash_completion.in @@ -97,7 +97,7 @@ _pacman_pkg() { } _pacman() { - local common core cur database prev query remove sync upgrade o + local common core cur database files prev query remove sync upgrade o COMPREPLY=() _get_comp_words_by_ref cur prev database=('asdeps asexplicit') -- 2.9.3

2 1

[pacman-dev] [PATCH] Fix memory leak in remove_notify_needed_optdepends
by Allan McRae 03 Sep '16

03 Sep '16

Also add pactest which captures this leak when run under valgrind. Reported-by: Sergey Petrenko Signed-off-by: Allan McRae <allan(a)archlinux.org> --- lib/libalpm/remove.c | 1 + test/pacman/tests/TESTS | 1 + .../pacman/tests/remove-optdepend-of-installed-package.py | 15 +++++++++++++++ 3 files changed, 17 insertions(+) create mode 100644 test/pacman/tests/remove-optdepend-of-installed-package.py diff --git a/lib/libalpm/remove.c b/lib/libalpm/remove.c index 45f7c2f..173dbc6 100644 --- a/lib/libalpm/remove.c +++ b/lib/libalpm/remove.c @@ -189,6 +189,7 @@ static void remove_notify_needed_optdepends(alpm_handle_t *handle, alpm_list_t * }; EVENT(handle, &event); } + free(optstring); } } } diff --git a/test/pacman/tests/TESTS b/test/pacman/tests/TESTS index bd5a0b6..2d87796 100644 --- a/test/pacman/tests/TESTS +++ b/test/pacman/tests/TESTS @@ -109,6 +109,7 @@ TESTS += test/pacman/tests/querycheck002.py TESTS += test/pacman/tests/querycheck_fast_file_type.py TESTS += test/pacman/tests/reason001.py TESTS += test/pacman/tests/remove-assumeinstalled.py +TESTS += test/pacman/tests/remove-optdepend-of-installed-package.py TESTS += test/pacman/tests/remove-recursive-cycle.py TESTS += test/pacman/tests/remove001.py TESTS += test/pacman/tests/remove002.py diff --git a/test/pacman/tests/remove-optdepend-of-installed-package.py b/test/pacman/tests/remove-optdepend-of-installed-package.py new file mode 100644 index 0000000..4973df5 --- /dev/null +++ b/test/pacman/tests/remove-optdepend-of-installed-package.py @@ -0,0 +1,15 @@ +self.description = "Remove packages which is an optdepend of another package" + +p1 = pmpkg("dep") +self.addpkg2db("local", p1) + +p2 = pmpkg("pkg") +p2.optdepends = ["dep: for foobar"] +self.addpkg2db("local", p2) + +self.args = "-R %s" % p1.name + +self.addrule("PACMAN_RETCODE=0") +self.addrule("!PKG_EXIST=%s" % p1.name) +self.addrule("PKG_EXIST=%s" % p2.name) +self.addrule("PACMAN_OUTPUT=%s optionally requires %s" % (p2.name, p1.name)) -- 2.9.3

1 0

[pacman-dev] Memleak
by Sergey Petrenko 03 Sep '16

03 Sep '16

libalpm/remove.c: in remove_notify_needed_optdepends(): char *optstring = alpm_dep_compute_string(optdep); isn't freed after being used. Ought to thank Valgrind for this one. -- With wish of constant improvement and unstoppable creativity. Sergey Petrenko

2 1

[pacman-dev] [PATCH] bash-completion: fix leaking "files" array into shell environment
by Eli Schwartz 02 Sep '16

02 Sep '16

--- contrib/bash_completion.in | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/contrib/bash_completion.in b/contrib/bash_completion.in index 6f99681..06963c4 100644 --- a/contrib/bash_completion.in +++ b/contrib/bash_completion.in @@ -97,7 +97,7 @@ _pacman_pkg() { } _pacman() { - local common core cur database prev query remove sync upgrade o + local common core cur database files prev query remove sync upgrade o COMPREPLY=() _get_comp_words_by_ref cur prev database=('asdeps asexplicit') -- 2.9.3

1 0

[pacman-dev] [PATCH] [RFC] Provide source files for useful debug packages
by Allan McRae 01 Sep '16

01 Sep '16

Debug packages are fairly useless currently because the soucre files needed for stepping through code etc are not packaged with them. This patch adds the needed source files to the debug package and adjusts the debug info to look at the /usr/src/debug/ directory for them rather than the build location. This requires using the "debugedit" program which is provided as part of the RPM sources. Signed-off-by: Allan McRae <allan(a)archlinux.org> --- This is a first pass at making debug packages useful... Some things I want comments on: 1) It feels dirty depending on an RPM tool for adjusting the paths inside ELF binaries, but I don't see any other choice. Suggestions welcome 2) Getting the list of src files from the binary involves an awk that works... but I don't know how! Can someone better at awk than me take a look and confirm I have not done something wrong 3) split packages... I need to figure out how to support these as source files will overlap. My current idea is to order the packaging by dependency chain so that e.g gcc-libs comes before gcc which comes before gcc-fortran. Then gcc-debug will not include files from gcc-libs-debug, and gcc-fortran-debug will not include files in either of those. I hope that would remove most file conflicts across packages. scripts/libmakepkg/tidy/strip.sh.in | 20 +++++++++++++++++++- 1 file changed, 19 insertions(+), 1 deletion(-) diff --git a/scripts/libmakepkg/tidy/strip.sh.in b/scripts/libmakepkg/tidy/strip.sh.in index 1c7aacf..bd5c79a 100644 --- a/scripts/libmakepkg/tidy/strip.sh.in +++ b/scripts/libmakepkg/tidy/strip.sh.in @@ -35,6 +35,11 @@ build_id() { LANG=C readelf -n $1 | sed -n '/Build ID/ { s/.*: //p; q; }' } +source_files() { + LANG=C readelf $1 --debug-dump | \ + awk '/DW_AT_name +:/{name=$8}/DW_AT_comp_dir +:/{print $8 "/" name}' +} + strip_file() { local binary=$1; shift @@ -50,6 +55,18 @@ strip_file() { return fi + # copy source files to debug directory + local f t + for f in $(source_files "$binary"); do + t=${f/$srcdir/$dbgsrc} + mkdir -p "${t%/*}" + cp "$f" "$t" + done + + # adjust debug symbols to point at sources + debugedit -b "${srcdir}" -d /usr/src/debug/ -i "$binary" &> /dev/null + + # copy debug symbols to debug directory mkdir -p "$dbgdir/${binary%/*}" objcopy --only-keep-debug "$binary" "$dbgdir/$binary.debug" objcopy --add-gnu-debuglink="$dbgdir/${binary#/}.debug" "$binary" @@ -89,7 +106,8 @@ tidy_strip() { if check_option "debug" "y"; then dbgdir="$pkgdir-@DEBUGSUFFIX@/usr/lib/debug" - mkdir -p "$dbgdir" + dbgsrc="$pkgdir-@DEBUGSUFFIX@/usr/src/debug" + mkdir -p "$dbgdir" "$dbgsrc" fi local binary strip_flags -- 2.9.3

1 0

[pacman-dev] [PATCH] Fix typo in pacman-db-upgrade usage message
by Allan McRae 01 Sep '16

01 Sep '16

Signed-off-by: Allan McRae <allan(a)archlinux.org> --- scripts/pacman-db-upgrade.sh.in | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/pacman-db-upgrade.sh.in b/scripts/pacman-db-upgrade.sh.in index 94085ba..79e1c59 100644 --- a/scripts/pacman-db-upgrade.sh.in +++ b/scripts/pacman-db-upgrade.sh.in @@ -40,7 +40,7 @@ usage() { printf -- "$(gettext "Usage: %s [options]")\n" "$0" echo printf -- "$(gettext "options:")\n" - printf -- "$(gettext " -b, --dbpath <path> set an alternate database location")\n" + printf -- "$(gettext " -d, --dbpath <path> set an alternate database location")\n" printf -- "$(gettext " -h, --help show this help message and exit")\n" printf -- "$(gettext " -r, --root <path> set an alternate installation root")\n" printf -- "$(gettext " -V, --version show version information and exit")\n" -- 2.9.3

1 0

[pacman-dev] [PATCH] Replace CURLOPT_PROGRESSFUNCTION with CURLOPT_XFERINFOFUNCTION
by ivy.foster＠gmail.com 31 Aug '16

31 Aug '16

From: Ivy Foster <ivy.foster(a)gmail.com> Curl 7.32.0 added CURLOPT_XFERINFOFUNCTION, which deprecates CURLOPT_PROGRESSFUNCTION and means less casting doubles to size_ts for alpm. This change has no user-facing nor frontend-facing effects. Signed-off-by: Ivy Foster <ivy.foster(a)gmail.com> --- lib/libalpm/dload.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) diff --git a/lib/libalpm/dload.c b/lib/libalpm/dload.c index f4e6a27..dc57c92 100644 --- a/lib/libalpm/dload.c +++ b/lib/libalpm/dload.c @@ -90,8 +90,8 @@ static void inthandler(int UNUSED signum) dload_interrupted = ABORT_SIGINT; } -static int dload_progress_cb(void *file, double dltotal, double dlnow, - double UNUSED ultotal, double UNUSED ulnow) +static int dload_progress_cb(void *file, curl_off_t dltotal, curl_off_t dlnow, + curl_off_t UNUSED ultotal, curl_off_t UNUSED ulnow) { struct dload_payload *payload = (struct dload_payload *)file; off_t current_size, total_size; @@ -106,7 +106,7 @@ static int dload_progress_cb(void *file, double dltotal, double dlnow, return 1; } - current_size = payload->initial_size + (off_t)dlnow; + current_size = payload->initial_size + dlnow; /* is our filesize still under any set limit? */ if(payload->max_size && current_size > payload->max_size) { @@ -119,9 +119,9 @@ static int dload_progress_cb(void *file, double dltotal, double dlnow, return 0; } - total_size = payload->initial_size + (off_t)dltotal; + total_size = payload->initial_size + dltotal; - if(DOUBLE_EQ(dltotal, 0.0) || payload->prevprogress == total_size) { + if(dltotal == 0 || payload->prevprogress == total_size) { return 0; } @@ -134,7 +134,7 @@ static int dload_progress_cb(void *file, double dltotal, double dlnow, * x {x>0}, x: download complete * x {x>0, x<y}, y {y > 0}: download progress, expected total is known */ if(current_size == total_size) { - payload->handle->dlcb(payload->remote_name, (off_t)dlnow, (off_t)dltotal); + payload->handle->dlcb(payload->remote_name, dlnow, dltotal); } else if(!payload->prevprogress) { payload->handle->dlcb(payload->remote_name, 0, -1); } else if(payload->prevprogress == current_size) { @@ -142,7 +142,7 @@ static int dload_progress_cb(void *file, double dltotal, double dlnow, } else { /* do NOT include initial_size since it wasn't part of the package's * download_size (nor included in the total download size callback) */ - payload->handle->dlcb(payload->remote_name, (off_t)dlnow, (off_t)dltotal); + payload->handle->dlcb(payload->remote_name, dlnow, dltotal); } payload->prevprogress = current_size; @@ -303,8 +303,8 @@ static void curl_set_handle_opts(struct dload_payload *payload, curl_easy_setopt(curl, CURLOPT_FILETIME, 1L); curl_easy_setopt(curl, CURLOPT_NOPROGRESS, 0L); curl_easy_setopt(curl, CURLOPT_FOLLOWLOCATION, 1L); - curl_easy_setopt(curl, CURLOPT_PROGRESSFUNCTION, dload_progress_cb); - curl_easy_setopt(curl, CURLOPT_PROGRESSDATA, (void *)payload); + curl_easy_setopt(curl, CURLOPT_XFERINFOFUNCTION, dload_progress_cb); + curl_easy_setopt(curl, CURLOPT_XFERINFODATA, (void *)payload); curl_easy_setopt(curl, CURLOPT_LOW_SPEED_LIMIT, 1L); curl_easy_setopt(curl, CURLOPT_LOW_SPEED_TIME, 10L); curl_easy_setopt(curl, CURLOPT_HEADERFUNCTION, dload_parseheader_cb); -- 2.9.3

4 5

[pacman-dev] [PATCH] Remove AC_FUNC_MALLOC check.
by Alastair Hughes 31 Aug '16

31 Aug '16

We weren't supplying the rpl_malloc function needed if this failed, and didn't check for realloc, so just remove. Signed-off-by: Alastair Hughes <hobbitalastair(a)gmail.com> --- The build will currently fail if cross compiling or if ac_cv_func_malloc_0_nonnull=no is set, because we don't supply a rpl_malloc function anywhere. As we don't use AC_FUNC_REALLOC either, just assume that we will never encounter a buggy system and remove the check. The other option appears to be to provide a malloc definition somewhere, but in that case it appears that we should also check for realloc. I'm unsure which one should be used - opinions? configure.ac | 1 - 1 file changed, 1 deletion(-) diff --git a/configure.ac b/configure.ac index dd4ac04..f2d3b75 100644 --- a/configure.ac +++ b/configure.ac @@ -308,7 +308,6 @@ PATH_MAX_DEFINED AC_FUNC_FORK AC_FUNC_GETMNTENT AC_FUNC_LSTAT_FOLLOWS_SLASHED_SYMLINK -AC_FUNC_MALLOC AC_FUNC_MKTIME AC_FUNC_STRCOLL AC_CHECK_FUNCS([dup2 getcwd getmntinfo gettimeofday memmove memset \ -- 2.9.3

2 1

[pacman-dev] I'm back (mostly)
by Allan McRae 30 Aug '16

30 Aug '16

Hi all, I had a low presence over the last couple of months due to purchasing a house, moving and the subsequent delay of internet connection. But I am back! Ping any patches that I have missed in my flurry tonight and I will review them. Cheers, Allan

1 0

[pacman-dev] Typo in pacman manpage
by Luca Weiss 30 Aug '16

30 Aug '16

Hey, just a quick "typo" report in the pacman manpage: In the category DATABASE OPTIONS there is no space between --asexplicit and <package>. Thanks!

3 2