summaryrefslogtreecommitdiffstats
path: root/scripts/makepkg.sh.in
AgeCommit message (Collapse)AuthorFilesLines
2008-05-29No error in makepkg when removing deps failsAllan McRae1-2/+9
Catches error from when pacman is unable to remove dependencies after successfully building package and prints warning. Fixes FS#10039. Signed-off-by: Allan McRae <mcrae_allan at hotmail.com> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-05-29Use chown 0:0 instead of root.rootSebastian Nowicki1-2/+2
On BSD systems using a dot as a separator is not allowed. On Mac OSX it is deprecated. A colon should be used instead. BSD systems also use the "wheel" group instead of "root" to indicate the "super user" group. Both groups use the id of 0. Signed-off-by: Sebastian Nowicki <sebnow@gmail.com> Acked-by: Aaron Griffin <aaronmgriffin@gmail.com> Acked-by: Dan McGee <dan@archlinux.org> Signed-off-by: Xavier Chantry <shiningxc@gmail.com> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-05-29Use "legacy mode" on Mac OSX (for file command)Sebastian Nowicki1-0/+3
For some reason `file` on Mac OSX has different arguments than BSD and Linux; -i no longer prints out the mime strings. With the environment variable COMMAND_MODE set to "legacy", `file` behaves more like it does on Linux and BSD, i.e., `file -i` prints the mime type. Signed-off-by: Sebastian Nowicki <sebnow@gmail.com> Acked-by: Dan McGee <dan@archlinux.org> Acked-by: Aaron Griffin <aaronmgriffin@gmail.com> Signed-off-by: Xavier Chantry <shiningxc@gmail.com> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-05-29Allow an unprivileged user to create packages without fakerootSebastian Nowicki1-2/+2
When fakeroot was not in BUILDENV and the user was not root, makepkg still tried to use fakeroot for building packages. BUILDENV is now checked to see if fakeroot is enabled. If it is not enabled the package can still be built, but root will not have ownership of files. This is useful when users want to make packages for personal use and don't care about ownership. Closes FS#10450. Signed-off-by: Sebastian Nowicki <sebnow@gmail.com> Signed-off-by: Xavier Chantry <shiningxc@gmail.com> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-05-29du -b is not available on BSD, use du -k instead.Xavier Chantry1-1/+1
This fixes FS#10459. There is apparently no portable ways to get the apparent size of a file, like du -b does. So the best compromise seems to get the block size in kB, and then convert that to byte so that we keep compatibility. Signed-off-by: Xavier Chantry <shiningxc@gmail.com> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-05-29Replace usage of "cp -s" with the more portable "ln -s"Sebastian Nowicki1-3/+6
The "-s" argument does not exist on BSD, and neither does "--remove-destination". This patch replaces the calls to "cp -s --remove-destination" with the equivalent "rm -f" and "ln -s" calls, in order to increase portability. Signed-off-by: Sebastian Nowicki <sebnow@gmail.com> Signed-off-by: Xavier Chantry <shiningxc@gmail.com> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-05-13Remove --builddeps from makepkgDan McGee1-40/+3
This really should be in an external script, as it is not makepkg's job to rebuild your system. Signed-off-by: Dan McGee <dan@archlinux.org>
2008-05-10makepkg: Unify start and end messagesDan McGee1-2/+2
I've always found it odd why the package version is shown at the start but not the end of the package build. Fix it, and while we are at it, add the $CARCH variable to the display too. Signed-off-by: Dan McGee <dan@archlinux.org>
2008-04-30Merge branch 'maint'Dan McGee1-4/+6
2008-04-27makepkg: fix strip section to allow spaces in pathsDaenyth Blank1-4/+6
Inside tidy_install, change the section which strips libraries to use find | while read rather than for foo in `find`. This should allow whitespaces in filenames to still be processed correctly. This fixes FS#10294. Signed-off-by: Daenyth Blank <Daenyth+git@gmail.com> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-03-16Add zipman makepkg option.Chantry Xavier1-30/+34
All other steps in tidy_install function were already controlled by an option in makepkg.conf, so this patch adds an option for the man page compression step too. This will allow to keep man pages uncompressed, which is required for some special meta man page, like the zshall one (see FS#4580). Ref: http://www.archlinux.org/pipermail/pacman-dev/2008-March/011472.html Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
2008-03-11makepkg: Support for resuming source downloadsRoman Kyrylych1-2/+19
Download to *.part and remove the suffix if successful, thus md5sums will be checked only for complete files, and next time makepkg will resume downloading of partial file instead of complaining about invalid md5sums. Old DLAGENTS format is still supported. Late fix: Do not try to download %u first The output of grep was not suppressed, thus the output of get_downloadcmd was wrong because of extra line at the beginning that contained an unsubstituted (with %o and %u) entry from DLAGENTS. Signed-off-by: Roman Kyrylych <roman@archlinux.org> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-02-27Remove small remnant of old force=y optionDan McGee1-1/+1
Signed-off-by: Dan McGee <dan@archlinux.org>
2008-02-21Fix two missing BUILDSCRIPT variables in makepkg.Mark Constable1-2/+2
Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
2008-02-18scripts: allow usage when gettext is not installedDan McGee1-0/+9
Address the issue of our scripts not working so great when gettext is not available. This has come up in multiple bug reports, and is relatively easy to address by adding a simple check and a stub function if gettext was not found that simply echos the original message. Addresses concerns from FS#9214 and FS#9607. Signed-off-by: Dan McGee <dan@archlinux.org>
2008-02-12makepkg : set sane umask before source extraction.Chantry Xavier1-3/+3
This fixes FS#9242 and FS#9362. Set umask before the sources are extracted, and after the /etc/profile.d/* files are sourced in handledeps. This sourcing of profile.d files is why umask was moved to the build function in the first place, as can be seen in commit ac965ed401. Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
2008-02-08makepkg: do a quick check for the VCS binariesDan McGee1-0/+7
Fixes FS#9230. Ensure we have the VCS binaries available when doing the development version number checking. Signed-off-by: Dan McGee <dan@archlinux.org>
2008-02-03add 'force' to PKGINFO, have repo-add respect it.Chantry Xavier1-0/+3
Ref: http://www.archlinux.org/pipermail/pacman-dev/2008-January/011023.html Also see FS#9347 and FS#9349. Signed-off-by: Chantry Xavier <shiningxc@gmail.com> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-01-28makepkg: ensure binaries in /opt/* are strippedDan McGee1-1/+1
The path selection for stripping binaries was slightly off, so any binaries in subdirectories of opt/ were missed. Fixes FS#9342. Signed-off-by: Dan McGee <dan@archlinux.org>
2008-01-27makepkg : improve determination of svn revision.Chantry Xavier1-1/+1
The previous sed command matched every line starting with r. For example, with mpd-svn package in aur, the svn log output was rather big, and there were several lines starting with r (the actual revision : r7155, but also other lines starting with reverting and run), so this broke makepkg: > makepkg ==> Determining latest svn revision... -> Version found: 7155 everting un sed: -e expression #1, char 27: unterminated `s' command To make the sed command more bullet proof, I added the -q option of svn log, which produces a quieter output, without the log. And I changed the sed command to only match numbers for the revision. Signed-off-by: Chantry Xavier <shiningxc@gmail.com> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-01-20workaround for a gettext string starting with --.Chantry Xavier1-1/+1
Workaround found in Advanced Bash-Scripting Guide, localization section. "added a \0 (NULL) at the beginning of the sentence." Signed-off-by: Chantry Xavier <shiningxc@gmail.com>
2008-01-20makepkg: add check to ensure we have non-URL files in build dirDan McGee1-0/+6
Pacman 3.0 printed the following message if a file could not be found: ERROR: xxx was not found in the build directory and is not a proper URL. We lost this logic in 3.1 when moving to the DLAGENT type stuff, so a not-found file got passed all the way to the download logic where it failed with a odd error message. Bring back some logic to ensure only files with URLs get past a certain point, and fail if the file is not available. Fixes FS#9208. Signed-off-by: Dan McGee <dan@archlinux.org>
2008-01-14makepkg: Oops! Lost the colon after forcever in opts parsingDan McGee1-1/+1
I screwed this up here: 916e226b7839e38d9245ecbef396eb5ba5b6a224 Signed-off-by: Dan McGee <dan@archlinux.org>
2008-01-14makepkg: Introduce --holdver option to prevent bumped pkgrelScott Horowitz1-8/+24
This allows us to keep the --forcever option internal to makepkg. Fixes FS#9194. Signed-off-by: Dan McGee <dan@archlinux.org>
2008-01-10makepkg: Ensure we include all manpage directories in compression lineDan McGee1-1/+1
Signed-off-by: Dan McGee <dan@archlinux.org>
2008-01-10makepkg: add package name to output when installing packageKarolina Lindqvist1-2/+1
Signed-off-by: Dan McGee <dan@archlinux.org>
2008-01-08makepkg: Fix issues with --asroot and building dependenciesDan McGee1-3/+4
The --asroot option was not passed to recursive calls of makepkg. In addition, a 'cd' call was done only on the fakeroot branch instead of both fakeroot and root branches. Move it outside the conditional. Noticed-by: Karolina Lindqvist <karolina.lindqvist@kramnet.se> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-01-05makepkg: add application/zip to the recognized typesDan McGee1-1/+1
The recent upgrade of magic.mime in the file program changed the mime type output to application/zip instead of application/x-zip as it is a registered type. Unfortunately we then just skipped zip file extraction becuase of this change. Signed-off-by: Dan McGee <dan@archlinux.org>
2008-01-05Remove mentions of ABS from makepkgDan McGee1-11/+0
Kill off some of the $ABSROOT stuff that was still hanging around in there. Signed-off-by: Dan McGee <dan@archlinux.org>
2007-12-29Remove .FILELIST generation from makepkg (and elsewhere)Dan McGee1-10/+1
This is something pacman can do on its own straight from the archive, and we will reduce the chance of problems occurring becuase of inproper FILELIST generation as we have had in the past with special characters in filenames. Once we remove it from makepkg. we can remove any usage of it from all of our other tools, including pacman, pactest, and contrib/ utilities. Note that removing it from pacman uncovered a few other bugs anyway, so this was probably a good move. Signed-off-by: Dan McGee <dan@archlinux.org>
2007-12-29Makepkg missing check for empty pkgnameK. Piche1-0/+4
I thought it was wierd that makepkg didn't do a sanity check for the $pkgname. This makes for a simple first try at a GIT patch. Signed-off-by: K. Piche <kevin@archlinux.org> Signed-off-by: Dan McGee <dan@archlinux.org>
2007-12-21Do not move man pages to a non-FHS locationDan McGee1-8/+0
Fixes Arch Linux bug FS#8839. Signed-off-by: Dan McGee <dan@archlinux.org>
2007-12-17Set the default packager to "Unknown Packager"Dan McGee1-1/+1
There is no real reason to make it look like someone that can't bother to set the packager variable is official in any way, so remove the Arch Linux reference from makepkg. Signed-off-by: Dan McGee <dan@archlinux.org>
2007-12-11Update GNU GPL boilerplate and copyright datesDan McGee1-3/+1
Update the GPL boilerplate to direct people to the GNU website for a copy of the license, as well as bump all of Judd's copyrights to 2007. Signed-off-by: Dan McGee <dan@archlinux.org>
2007-12-05Compress hard linked man pagesAllan McRae1-12/+28
This fixes FS#5392. If hard links are present for a man page, all other hard linked files are removed, the man page is zipped and the hard links are updated to the newly compress man page. Signed-off-by: Allan McRae <mcrae_allan@hotmail.com> [Dan: use find -samefile option, don't fail if one of mandirs is nonexistent] Signed-off-by: Dan McGee <dan@archlinux.org>
2007-11-11makepkg: remove .pacsave files when uninstalling dependenciesRoman Kyrylych1-2/+2
Signed-off-by: Roman Kyrylych <roman@archlinux.org> Signed-off-by: Dan McGee <dan@archlinux.org>
2007-11-08Correctly handle version comparisons for SVN/CVS/etc PKGBUILDS.Scott Horowitz1-3/+4
Retrieve SVN/CVS/etc revision number before checking if the package has already been built. This allows building a newer version of a developmental package without having to use -f. Now -f will only be needed if the latest SVN/CVS/etc revision is the same as the built package. Signed-off-by: Scott Horowitz <stonecrest@gmail.com> Signed-off-by: Dan McGee <dan@archlinux.org>
2007-11-06Fix mercurial complaining about directory already existingSebastian Nowicki1-2/+1
Signed-off-by: Dan McGee <dan@archlinux.org>
2007-11-04Add support for arch='any'Roman Kyrylych1-0/+4
Signed-off-by: Roman Kyrylych <roman@archlinux.org> Signed-off-by: Dan McGee <dan@archlinux.org>
2007-10-29makepkg: add explicit dir to find callsDan McGee1-3/+3
Some of the find calls did not have a path, which isn't valid in some versions of the find utility. Add a . for compatibility. Signed-off-by: Dan McGee <dan@archlinux.org>
2007-10-27Integrate versionpkg into makepkgScott Horowitz1-4/+86
If PKGBUILDs use the correct svn/cvs/etc variables, determine the latest revision number and force this as the pkgver for building. Signed-off-by: Scott Horowitz <stonecrest@gmail.com> Signed-off-by: Dan McGee <dan@archlinux.org>
2007-10-27makepkg: readd -p option that was lost in getopt conversionDan McGee1-0/+1
The -p option was accidentally dropped in commit 54b71f0427e87e6d525423df06f8a06f2b71c518. Readd it. Signed-off-by: Dan McGee <dan@archlinux.org>
2007-10-19scripts/makepkg.sh.in: Strip bins/libs in all {bin, sbin, lib} directories.Andrew Fyfe1-1/+1
In one of the original clean up patches[1] I changed the search path for stripping binaries and libraries. This resulted in only usr/{bin,sbin,lib} being searched. This patch reverts that change. [1] 721ceee1e2c9b18425d84cf39f6541b2f04072b3 Signed-off-by: Andrew Fyfe <andrew@neptune-one.net>
2007-10-15Manually check xdelta exit codes.Nathan Jones1-9/+20
This is needed because 'xdelta delta' returns 1 even when the delta creation succeeds. This was causing makepkg to exit after the command due to bash's -e option. Some information from the xdelta man page: The delta command exits with status 0 to indicate that no differences were found, with status 1 to indicate that some differences were found, and with status 2 to indicate an error of some kind. Signed-off-by: Nathan Jones <nathanj@insightbb.com> Signed-off-by: Dan McGee <dan@archlinux.org>
2007-10-05makepkg: Fix typo to make xdelta creation work.Nathan Jones1-1/+1
Signed-off-by: Nathan Jones <nathanj@insightbb.com> Signed-off-by: Aaron Griffin <aaronmgriffin@gmail.com>
2007-09-28makepkg: Fix output color codes for realDan McGee1-5/+5
Make sure we reset both the bold attribute and the color attributes when it is necessary. Signed-off-by: Dan McGee <dan@archlinux.org>
2007-09-28makepkg: When we add an option, we should probably make it workDan McGee1-2/+2
The -A,--ignorearch option was added and given help text, but never actually listed as valid to the option parser. Fix this. Signed-off-by: Dan McGee <dan@archlinux.org>
2007-09-28Use ANSI 'foreground' color instead of hardcoding 'white'Aaron Griffin1-5/+5
Signed-off-by: Aaron Griffin <aaronmgriffin@gmail.com>
2007-09-28Support for localized times in metadataAaron Griffin1-1/+1
Packages and DBs now support using the UNIX epoch (seconds since Jan 1, 1970) for use in builddate and installdate. This will only affect newly built packages. Old existing packages with the text format are still supported, but this is deprecated. In the case of removal of text time support, this code will fail gracefully, returning the start of the epoch for broken packages. Signed-off-by: Aaron Griffin <aaronmgriffin@gmail.com>
2007-09-26Preliminary support for optdependsDan McGee1-1/+4
Add some alpm functions for getting optdepends, have makepkg include them in the PKGINFO file, and have a pacman -Qi operation display the raw string as stored by libalpm. Signed-off-by: Dan McGee <dan@archlinux.org>