March 2012 - Pacman-dev - lists.archlinux.org

[pacman-dev] [PATCH] libalpm: add pkg-config file
by Dave Reisner 20 Mar '12

20 Mar '12

Borrowed heavily from libarchive. Signed-off-by: Dave Reisner <dreisner(a)archlinux.org> --- Throw a bone to the few projects who link against us... configure.ac | 1 + lib/libalpm/Makefile.am | 3 +++ lib/libalpm/libalpm.pc.in | 11 +++++++++++ 3 files changed, 15 insertions(+) create mode 100644 lib/libalpm/libalpm.pc.in diff --git a/configure.ac b/configure.ac index 12a6e65..9093a90 100644 --- a/configure.ac +++ b/configure.ac @@ -382,6 +382,7 @@ AC_DEFINE_UNQUOTED([SCRIPTLET_SHELL], "$SCRIPTLET_SHELL", [The shell used to run AC_CONFIG_FILES([ lib/libalpm/Makefile lib/libalpm/po/Makefile.in +lib/libalpm/libalpm.pc src/pacman/Makefile src/pacman/po/Makefile.in src/util/Makefile diff --git a/lib/libalpm/Makefile.am b/lib/libalpm/Makefile.am index 61dcb87..17d9a58 100644 --- a/lib/libalpm/Makefile.am +++ b/lib/libalpm/Makefile.am @@ -23,6 +23,9 @@ if ENABLE_GNU89_INLINE_CC AM_CFLAGS += -fgnu89-inline endif +pkgconfigdir = $(libdir)/pkgconfig +pkgconfig_DATA = libalpm.pc + libalpm_la_SOURCES = \ add.h add.c \ alpm.h alpm.c \ diff --git a/lib/libalpm/libalpm.pc.in b/lib/libalpm/libalpm.pc.in new file mode 100644 index 0000000..12bea28 --- /dev/null +++ b/lib/libalpm/libalpm.pc.in @@ -0,0 +1,11 @@ +prefix=@prefix@ +exec_prefix=@exec_prefix@ +libdir=@libdir@ +includedir=@includedir@ + +Name: libalpm +Description: Arch Linux package management library +Version: @VERSION@ +Cflags: -I${includedir} +Libs: -L${libdir} -lalpm +Libs.private: @LIBS@ -- 1.7.9.4

2 1

[pacman-dev] Debug symbol packages bikeshedding
by Allan McRae 20 Mar '12

20 Mar '12

I have been looking at the patches for automatic creation of packages with debug symbols. I think they are good starts, but before I polish them we need to decide on a few things. My preferences are highlighted with a (*) 1) the option in makepkg.conf and PKGBUILD options=() array - debug (*) - debugpkg - splitdbg - splitdebug 2) the naming of the package with debug symbols - foo-dbg - foo-debug (*) - foo-debuginfo - foo-debug-symbols (Note -dbg is Debian style, -debuginfo is RPM style, and -debug potentially conflicts with a few packages in the AUR but none in the Arch repos...) 3) what information the .PKGINFO file has: pkgdesc="Debugging symbols for $pkgname" Must be included: - pkgname - pkgver/pkgrel/epoch/arch as in parent package - depends=(parent=$epoch:$pkgver-$pkgrel) - license Should be included? - url - replaces - conflicts Do not include: - groups - provides - backup - install - changelog Allan

2 1

[pacman-dev] The great VCS packages overhaul of 2015
by Allan McRae 18 Mar '12

18 Mar '12

It is well know that the VCS package support in makepkg is subpar... The number of bugs about this is into double figures. There have also been some patches submitted attempting to be able to use the source=() array in PKGBUILDs to specify a VCS repo and remove a lot of the repetitive crap involved for repo checkout/updating in the current PKGBUILDs. To get the ball rolling again, I think we should pick one VCS system and flesh out what we need and what the prototype PKGBUILD would look like. Then we can move on to the other VCS systems and finally implement it. I guess by the time all the bikeshedding is finished, this will all be done by 2015 and hence the subject line. :) I am going to start with git. Current: pkgver = the date the package was built _gitroot = the url of the git repo _gitname = the name of the directory we check out the repo into... According to the man page, _gitname is supposed to be a branch or tag but that is all lies. What would we like to have for a flexible git package implementation? - url of the repo - be able to specify the branch/tag/commit to use (and appropriate combinations) - a decent pkgver 1) URL: There were previous patches to the mailing list that never really got finished, but I think we were fairly happy with this syntax: source=(git://projects.archlinux.org/pacman.git) source=(git@@http://projects.archlinux.org/git/pacman.git) Does it make any sense to allow the "::" syntax here? i.e. source=(git@@dirname::http://projects.archlinux.org/git/pacman.git) where dirname is the name of the directory it checked out into? I am thinking we should probably do the checkout into $vcsdir=$startdir/vcs, so this would only be needed if #1 - we supported multiple VCS checkouts in one PKGBUILD there were two that wanted to used the same name... #2 - another source file conflicted #2 is readily dealt with and I am not sure we should allow #1 (see below) so I would vote to skip it until a genuine need is shown. 2) Specifying commit to work with: I think that this is the difficult bit... the syntax with the source array is already convoluted enough, so I do not think they should be added there. So that suggests we go with assigning them to variables like _git_branch, _git_commit, _git_tag... etc. But what if we have two git sources? For example, say pacman allowed building against an internal copy of libarchive if a folder named libarchive was found in its root directory. So: source=(git://projects.archlinux.org/pacman.git git://github.com/libarchive/libarchive) build() { cd $srcdir/pacman ln -s ../libarchive ./autogen.sh ... It might seem a somewhat convoluted example, but (e.g.) gcc does allow in source tree building of many of its dependencies. The question is should we consider this outside the realms of the reasonable and state one VCS repo should be one package. I'd say 99.999% of VCS PKGBUILDs (at a lower bound) would never use two VCS sources... and the ones that do need this could do manual checking out of the non-main source within the PKGBUILD anyway. 3) pkgver Use output of "git describe" (with a s/-/_/) and fall back to "git rev-list HEAD | wc -l" (with a trailing commit id added) if there are no tags in the repo. PROTOTYPE: pkgname=pacman-git pkgver=AUTO ... source=(git://projects.archlinux.org/users/allan/pacman.git) _git_branch="working" _git_commit="8c857343" build() { cd $srcdir/pacman ./autogen.sh ... } What makepkg does: 1) goes into $vcsdir, checks for the pacman directory - if not present, do the git checkout - if present, enter and do a "git pull" unless --holdver is specified 2) enters $srcdir, and does the appropriate clone of the repo in $vscdir to be at the required branch/tag/commit 3) starts build() etc... Oh wow... you are still reading this? Well then, you are now qualified to comment on the proposal. Please stick to just concerns with the git usage at the moment, unless you see something monumental that we will not be able to support using other VCS systems using this approach. Allan

6 24

[pacman-dev] [PATCH 1/3] Pass a file descriptor to getcols and flush_term_input
by Dan McGee 16 Mar '12

16 Mar '12

This makes these methods a bit more flexible. Signed-off-by: Dan McGee <dan(a)archlinux.org> --- src/pacman/callback.c | 4 ++-- src/pacman/util.c | 31 ++++++++++++++++--------------- src/pacman/util.h | 2 +- 3 files changed, 19 insertions(+), 18 deletions(-) diff --git a/src/pacman/callback.c b/src/pacman/callback.c index 9fd9cc4..61f517c 100644 --- a/src/pacman/callback.c +++ b/src/pacman/callback.c @@ -384,7 +384,7 @@ void cb_progress(alpm_progress_t event, const char *pkgname, int percent, int len, wclen, wcwid, padwid; wchar_t *wcstr; - const unsigned short cols = getcols(); + const unsigned short cols = getcols(fileno(stdout)); if(config->noprogressbar || cols == 0) { return; @@ -544,7 +544,7 @@ void cb_dl_progress(const char *filename, off_t file_xfered, off_t file_total) const char *rate_label, *xfered_label; int file_percent = 0, total_percent = 0; - const unsigned short cols = getcols(); + const unsigned short cols = getcols(fileno(stdout)); if(config->noprogressbar || cols == 0 || file_total == -1) { if(file_xfered == 0) { diff --git a/src/pacman/util.c b/src/pacman/util.c index a5242ae..15cd6fd 100644 --- a/src/pacman/util.c +++ b/src/pacman/util.c @@ -130,10 +130,10 @@ int check_syncdbs(size_t need_repos, int check_valid) } /* discard unhandled input on the terminal's input buffer */ -static int flush_term_input(void) { +static int flush_term_input(int fd) { #ifdef HAVE_TCFLUSH - if(isatty(fileno(stdin))) { - return tcflush(fileno(stdin), TCIFLUSH); + if(isatty(fd)) { + return tcflush(fd, TCIFLUSH); } #endif @@ -142,24 +142,24 @@ static int flush_term_input(void) { } /* gets the current screen column width */ -unsigned short getcols(void) +unsigned short getcols(int fd) { const unsigned short default_tty = 80; const unsigned short default_notty = 0; unsigned short termwidth = 0; - if(!isatty(fileno(stdout))) { + if(!isatty(fd)) { return default_notty; } -#ifdef TIOCGSIZE +#if defined(TIOCGSIZE) struct ttysize win; - if(ioctl(1, TIOCGSIZE, &win) == 0) { + if(ioctl(fd, TIOCGSIZE, &win) == 0) { termwidth = win.ts_cols; } #elif defined(TIOCGWINSZ) struct winsize win; - if(ioctl(1, TIOCGWINSZ, &win) == 0) { + if(ioctl(fd, TIOCGWINSZ, &win) == 0) { termwidth = win.ws_col; } #endif @@ -259,7 +259,7 @@ void indentprint(const char *str, size_t indent) wchar_t *wcstr; const wchar_t *p; int len, cidx; - const unsigned short cols = getcols(); + const unsigned short cols = getcols(fileno(stdout)); if(!str) { return; @@ -622,7 +622,7 @@ int table_display(const char *title, const alpm_list_t *header, totalwidth = table_calc_widths(header, rows, padding, totalcols, &widths, &has_data); /* return -1 if terminal is not wide enough */ - if(totalwidth > getcols()) { + if(totalwidth > getcols(fileno(stdout))) { pm_printf(ALPM_LOG_WARNING, _("insufficient columns available for table display\n")); return -1; @@ -660,7 +660,7 @@ void list_display(const char *title, const alpm_list_t *list) if(!list) { printf("%s\n", _("None")); } else { - const unsigned short maxcols = getcols(); + const unsigned short maxcols = getcols(fileno(stdout)); size_t cols = len; const char *str = list->data; fputs(str, stdout); @@ -1377,7 +1377,7 @@ int multiselect_question(char *array, int count) break; } - flush_term_input(); + flush_term_input(fileno(stdin)); if(fgets(response, response_len, stdin)) { const size_t response_incr = 64; @@ -1440,7 +1440,7 @@ int select_question(int count) break; } - flush_term_input(); + flush_term_input(fileno(stdin)); if(fgets(response, sizeof(response), stdin)) { size_t len = strtrim(response); @@ -1463,6 +1463,7 @@ static int question(short preset, char *fmt, va_list args) { char response[32]; FILE *stream; + int fd_in = fileno(stdin); if(config->noconfirm) { stream = stdout; @@ -1489,7 +1490,7 @@ static int question(short preset, char *fmt, va_list args) } fflush(stream); - flush_term_input(); + flush_term_input(fd_in); if(fgets(response, sizeof(response), stdin)) { size_t len = strtrim(response); @@ -1499,7 +1500,7 @@ static int question(short preset, char *fmt, va_list args) /* if stdin is piped, response does not get printed out, and as a result * a \n is missing, resulting in broken output (FS#27909) */ - if(!isatty(fileno(stdin))) { + if(!isatty(fd_in)) { fprintf(stream, "%s\n", response); } diff --git a/src/pacman/util.h b/src/pacman/util.h index 0df46e8..fc97707 100644 --- a/src/pacman/util.h +++ b/src/pacman/util.h @@ -47,7 +47,7 @@ int trans_init(alpm_transflag_t flags, int check_valid); int trans_release(void); int needs_root(void); int check_syncdbs(size_t need_repos, int check_valid); -unsigned short getcols(void); +unsigned short getcols(int fd); int rmrf(const char *path); const char *mbasename(const char *path); char *mdirname(const char *path); -- 1.7.9.4

1 2

[pacman-dev] [GIT] The official pacman repository branch, master, updated. v4.0.2-185-ge6f72c6
by dan＠archlinux.org 16 Mar '12

16 Mar '12

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 e6f72c61a098b52ea29e54b8eb4739a2ff81e6b0 (commit) via f988aa6b32503f5d4003b1402089df74adf8b485 (commit) via 573bd80f147ce90f21ef51fb8ca6f71e89cc3a74 (commit) via 44de3183ff42dc4e5972a9d563ebc7125c3eb4dd (commit) via 03465ad6b6719e0afb20ca8829581190a2b7658e (commit) via d521f30f071e2bd2c39c05f39df34be867dfe572 (commit) via 9e1a20022d8186eba0005e20b8b6db16cf7dec64 (commit) via 56f821b72a6b56cdcbc727515e43c16affd39f08 (commit) via 930e93b8925874a3c6d3460d2c0cebcda5fc0d09 (commit) via 89f4dd88d784632fed134ae88284074511cf4567 (commit) via 83e42dcccef08233134a0a5da7e3da70638e6c97 (commit) via f41037358af167424403d30a3b972b1fd733c755 (commit) via 952ee5e6cb8bb1849c1bceedae38744935c82fca (commit) via 47d0df6c101455c989887551bd0f0149a0f00de7 (commit) via 5bda38196fbd6ad389fd829924736b007732e2f1 (commit) via ff58e5cb2d6aca0e1c8ed0e77770f3c5d2442c46 (commit) from 8de0631edbd1fd7681f6a924f5e166621328c13f (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 e6f72c61a098b52ea29e54b8eb4739a2ff81e6b0 Merge: 573bd80 f988aa6 Author: Dan McGee <dan(a)archlinux.org> Date: Fri Mar 16 12:08:10 2012 -0500 Merge branch 'maint' Conflicts: lib/libalpm/sync.c commit 573bd80f147ce90f21ef51fb8ca6f71e89cc3a74 Author: Allan McRae <allan(a)archlinux.org> Date: Fri Mar 9 17:27:43 2012 +1000 makepkg: abort when failing to create BUILDDIR Signed-off-by: Allan McRae <allan(a)archlinux.org> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 44de3183ff42dc4e5972a9d563ebc7125c3eb4dd Author: Allan McRae <allan(a)archlinux.org> Date: Fri Mar 9 17:15:28 2012 +1000 makepkg: prevent issues with files starting with a hyphen Most places in makepkg deal with full file paths, but a few use the file name only. Protect from potential issues when a file name starts with a hyphen. Signed-off-by: Allan McRae <allan(a)archlinux.org> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 03465ad6b6719e0afb20ca8829581190a2b7658e Author: Allan McRae <allan(a)archlinux.org> Date: Fri Mar 9 16:01:28 2012 +1000 makepkg: append pkgbase to BUILDDIR if specified This means multiple packages can be build using the same BUILDDIR without their source files all ending up extracted in the same directory. Fixes FS#28417. Signed-off-by: Allan McRae <allan(a)archlinux.org> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit d521f30f071e2bd2c39c05f39df34be867dfe572 Author: Dan McGee <dan(a)archlinux.org> Date: Tue Mar 13 06:33:56 2012 -0500 pactest: make OPTDEPEND rule look at non-description only Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 9e1a20022d8186eba0005e20b8b6db16cf7dec64 Author: Dan McGee <dan(a)archlinux.org> Date: Wed Mar 14 22:19:13 2012 -0400 Always display download size if it applies Currently, we try to do a bunch of funkyness constraining download size to print only when doing a -S/--sync operation. However, it is possible we try to download packages on a -U/--upgrade operation, and we currently won't show any itemized download sizes. Fix this ommission by always including the download size stuff in the built table rows; this column will be completely omitted anyway if there are no values due to prior work in commit 33bb7dbd35. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 56f821b72a6b56cdcbc727515e43c16affd39f08 Author: Dan McGee <dan(a)archlinux.org> Date: Wed Mar 14 22:18:37 2012 -0400 Handle printing NULL correctly in table display Treat this value as the empty string. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 930e93b8925874a3c6d3460d2c0cebcda5fc0d09 Author: Dan McGee <dan(a)archlinux.org> Date: Fri Mar 9 14:24:48 2012 -0600 Prefer fputs/fputc over fprintf when writing local database Using fputs should be faster as no format string parsing is required. It also prevents silly errors related to unescaped '%' signs, and removes the need to double them up in a lot of places. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 89f4dd88d784632fed134ae88284074511cf4567 Author: Dan McGee <dan(a)archlinux.org> Date: Fri Mar 9 14:10:04 2012 -0600 Extract a write_deps function from local database writing This reduces a lot of code duplication in the write function, which cleans it up a bit. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 83e42dcccef08233134a0a5da7e3da70638e6c97 Author: Dan McGee <dan(a)archlinux.org> Date: Fri Mar 9 13:40:55 2012 -0600 Remove unnecessary path check from local database This removes a call to _alpm_local_db_pkgpath() as well as an access() call when reading the local database. This appears to be code from 2006 that has stuck around. We don't need it because: 1) We never use this path except to check it via access(); however, we are already in a readdir() loop so it exists, or at least did at the time of the call. 2) The fopen() and other calls will fail on accessing the database files anyway, and we need to check those for errors. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 5bda38196fbd6ad389fd829924736b007732e2f1 Author: Dan McGee <dan(a)archlinux.org> Date: Mon Mar 12 21:47:29 2012 -0500 Minor format-string related cleanups We had one stubbed out so we didn't require a translation update, and the other is more a code style issue. Signed-off-by: Dan McGee <dan(a)archlinux.org> ----------------------------------------------------------------------- Summary of changes: doc/index.txt | 17 +++++-- lib/libalpm/be_local.c | 106 ++++++++++++++++++------------------------------ lib/libalpm/dload.c | 5 ++ lib/libalpm/sync.c | 10 +++- scripts/makepkg.sh.in | 25 ++++++++---- src/pacman/sync.c | 17 ++++++++ src/pacman/util.c | 34 +++++++-------- test/pacman/pmrule.py | 7 ++- 8 files changed, 118 insertions(+), 103 deletions(-) hooks/post-receive -- The official pacman repository

1 0

[pacman-dev] [GIT] The official pacman repository branch, maint, updated. v4.0.2-25-gf988aa6
by dan＠archlinux.org 16 Mar '12

16 Mar '12

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, maint has been updated via f988aa6b32503f5d4003b1402089df74adf8b485 (commit) via f41037358af167424403d30a3b972b1fd733c755 (commit) via 952ee5e6cb8bb1849c1bceedae38744935c82fca (commit) via 47d0df6c101455c989887551bd0f0149a0f00de7 (commit) via ff58e5cb2d6aca0e1c8ed0e77770f3c5d2442c46 (commit) from 1a8c3e52d70bfa21ba108aa77179adf90401949d (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 f988aa6b32503f5d4003b1402089df74adf8b485 Author: Allan McRae <allan(a)archlinux.org> Date: Fri Mar 9 17:32:11 2012 +1000 makepkg: quote removed filename as it can have spaces Signed-off-by: Allan McRae <allan(a)archlinux.org> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit f41037358af167424403d30a3b972b1fd733c755 Author: Dave Reisner <dreisner(a)archlinux.org> Date: Wed Mar 14 23:06:24 2012 -0400 lib/sync: avoid unwanted unlink after download fail In case we have a mirror failure, unlink_on_fail would remain set, causing an interrupt in a successive download attempt to be wrongly unlinked. This also fixes a memory leak in the url member, as we would allocate over the previous, unfreed URL. Signed-off-by: Dave Reisner <dreisner(a)archlinux.org> commit 952ee5e6cb8bb1849c1bceedae38744935c82fca Author: Dave Reisner <dreisner(a)archlinux.org> Date: Wed Mar 14 23:04:43 2012 -0400 dload: reset payload filename members before download To avoid conflicts on reusing a payload after a failed download, ensure that we reset the filename hints in the payload struct prior to the download operation. Signed-off-by: Dave Reisner <dreisner(a)archlinux.org> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 47d0df6c101455c989887551bd0f0149a0f00de7 Author: Allan McRae <allan(a)archlinux.org> Date: Fri Mar 9 23:16:40 2012 +1000 Skip special files when cleaning package cache Ignore *.sig, *.db*, and *.src.tar* when cleaning the package cache. Fixes FS#25166. Signed-off-by: Allan McRae <allan(a)archlinux.org> commit ff58e5cb2d6aca0e1c8ed0e77770f3c5d2442c46 Author: Dan McGee <dan(a)archlinux.org> Date: Mon Mar 12 21:16:14 2012 -0500 Add 3.0.x versions to doc/index.txt These got lost when I switched to the new Asciidoc tables syntax in commit 1ebe5dc1979e. Signed-off-by: Dan McGee <dan(a)archlinux.org> ----------------------------------------------------------------------- Summary of changes: doc/index.txt | 17 ++++++++++++----- lib/libalpm/dload.c | 5 +++++ lib/libalpm/sync.c | 2 ++ scripts/makepkg.sh.in | 2 +- src/pacman/sync.c | 17 +++++++++++++++++ 5 files changed, 37 insertions(+), 6 deletions(-) hooks/post-receive -- The official pacman repository

1 0

[pacman-dev] [PATCH 1/2] dload: reset payload filename members before download
by Dave Reisner 15 Mar '12

15 Mar '12

To avoid conflicts on reusing a payload after a failed download, ensure that we reset the filename hints in the payload struct prior to the download operation. Signed-off-by: Dave Reisner <dreisner(a)archlinux.org> --- lib/libalpm/dload.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/lib/libalpm/dload.c b/lib/libalpm/dload.c index 0598806..b3ac024 100644 --- a/lib/libalpm/dload.c +++ b/lib/libalpm/dload.c @@ -390,6 +390,11 @@ static int curl_download_internal(struct dload_payload *payload, CURL *curl = get_libcurl_handle(handle); handle->pm_errno = 0; + /* make sure these are NULL */ + FREE(payload->tempfile_name); + FREE(payload->destfile_name); + FREE(payload->content_disp_name); + payload->tempfile_openmode = "wb"; if(!payload->remote_name) { STRDUP(payload->remote_name, get_filename(payload->fileurl), -- 1.7.9.4

1 1

[pacman-dev] [PATCH 1/2] Handle printing NULL correctly in table display
by Dan McGee 15 Mar '12

15 Mar '12

Treat this value as the empty string. Signed-off-by: Dan McGee <dan(a)archlinux.org> --- src/pacman/util.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/src/pacman/util.c b/src/pacman/util.c index 7be3dc5..a96aa62 100644 --- a/src/pacman/util.c +++ b/src/pacman/util.c @@ -500,6 +500,9 @@ static void table_print_line(const alpm_list_t *line, short col_padding, } value = curcell->data; + if(!value) { + value = ""; + } /* silly printf requires padding size to be an int */ cell_padding = (int)widths[i] - (int)string_length(value); if(cell_padding < 0) { -- 1.7.9.4

1 1

[pacman-dev] [PATCH 1/2] Revert "makepkg: calculate exact total file size"
by Dave Reisner 15 Mar '12

15 Mar '12

btrfs's cow snapshots seem to do very strange things, causing du to not count data still in filesystem buffers. Repackaging the same build of handbrake within a chroot on brynhild, I witnessed (via bash xtrace) du reporting a different $pkgdir size every time the write_pkginfo() function ran. Unfortunately, replacing du with stat has its own slew of problems, mostly due to hard links (e.g. git, with 106 hardlinks to the same file). Working around this is neither fun, nor practical. As it turns out, all we needed here all along was a simple call to sync to flush writes to disk before calling du. This reverts commit b264fb9e9ddcc31dc8782390309421965e507383. Signed-off-by: Dave Reisner <dreisner(a)archlinux.org> --- configure.ac | 1 + scripts/Makefile.am | 1 + scripts/makepkg.sh.in | 7 ++++++- 3 files changed, 8 insertions(+), 1 deletion(-) diff --git a/configure.ac b/configure.ac index 12a6e65..52f8ad4 100644 --- a/configure.ac +++ b/configure.ac @@ -282,6 +282,7 @@ esac AM_CONDITIONAL([CYGWIN], test "x$host_os_cygwin" = "xyes") AM_CONDITIONAL([DARWIN], test "x$host_os_darwin" = "xyes") +AC_PATH_PROGS([DUPATH], [du], [du], [/usr/bin$PATH_SEPARATOR/bin] ) AC_SUBST(SIZECMD) AC_SUBST(SEDINPLACE) AC_SUBST(STRIP_BINARIES) diff --git a/scripts/Makefile.am b/scripts/Makefile.am index 727de25..328fbff 100644 --- a/scripts/Makefile.am +++ b/scripts/Makefile.am @@ -53,6 +53,7 @@ edit = sed \ -e 's|@BUILDSCRIPT[@]|$(BUILDSCRIPT)|g' \ -e 's|@SIZECMD[@]|$(SIZECMD)|g' \ -e 's|@SEDINPLACE[@]|$(SEDINPLACE)|g' \ + -e 's|@DUPATH[@]|$(DUPATH)|g' \ -e 's|@SCRIPTNAME[@]|$@|g' \ -e 's|@configure_input[@]|Generated from $@.sh.in; do not edit by hand.|g' diff --git a/scripts/makepkg.sh.in b/scripts/makepkg.sh.in index 89cd118..c2941b2 100644 --- a/scripts/makepkg.sh.in +++ b/scripts/makepkg.sh.in @@ -1156,7 +1156,8 @@ write_pkginfo() { else local packager="Unknown Packager" fi - local size="$(find . -print0 | xargs -0 @SIZECMD@ | awk '{ sum += $1 } END { print sum }')" + local size="$(@DUPATH@ -sk)" + size="$(( ${size%%[^0-9]*} * 1024 ))" msg2 "$(gettext "Generating %s file...")" ".PKGINFO" echo "# Generated by makepkg $myver" @@ -1264,6 +1265,10 @@ create_package() { PKGARCH=$CARCH fi + # flush buffers to disk. This avoids problems with du calculating the wrong + # installed package size, particularly on btrfs CoW snapshots. + sync + write_pkginfo $nameofpkg > .PKGINFO local comp_files=('.PKGINFO') -- 1.7.9.4

2 4

[pacman-dev] [PATCH 0/6] VCS versioning modifications
by dgbaley27＠0x01b.net 13 Mar '12

13 Mar '12

From: Matthew Monaco <matthew.monaco(a)0x01b.net> I've been bothered by makepkg automatically adjusting pkgver if a package doesn't have the correct -git/svn/etc suffix. The first 4 patches might be useful regardless. They allow for multiple sets of _vcs variables in a PKGBUILD and use the package's name to clear up any ambiguity. The 5th patch changes the current semantics of makepkg. The 6th patch provides a new option to force the old semantics. An alternative might be to sanitize the --forcever option so that it may be used from the command line as well as internally. Matthew Monaco (6): makepkg: remove unused variable makepkg: devel_check(): determine vcs makepkg: devel_check(): consolidate common code makepkg: devel_check(): set newpkgver (cleanup) makepkg: devel_check(): only work for -vcs pkgnames makepkg: add --devver option for SCM packages doc/PKGBUILD.5.txt | 3 +- doc/makepkg.8.txt | 4 ++ scripts/makepkg.sh.in | 126 +++++++++++++++++++++++++++---------------------- 3 files changed, 75 insertions(+), 58 deletions(-) -- 1.7.9.3

5 22