summaryrefslogtreecommitdiff
path: root/mk
AgeCommit message (Collapse)AuthorFilesLines
2009-12-20Fix libtool under OS X: require a fixed version of f2c, and thus bumpjmmv2-4/+4
PKGREVISION of libtool to 3 and require it as a dependency.
2009-12-19Introduce PKG_OPTIONS for database/tokyocabinet to select compression supportobache1-0/+2
(zlib/bzip2/lzo/lzma and enabled by default for former two). LZO and LZMA supports are exclusive, optional and should be selected by user's use-case (priority of data compression ratio, compression speed, or so on). Furthermore, Tokyocabinet uses archives/lzmalib for lzma support, but lzmalib conflicts with archivers/xz. It means if tokyocabinet is installed with lzma, failed to install xz, and result in failure of "make extract" for packages using .lzma/.xz distfiles. This issue also resolved by optionify lzma support and disabled by default. This changes also make tokyocabinet ABI changes, so bump PKGREVISION and recursive bump. OKed to commit in freeze by wiz@.
2009-12-19SUA 6.0 have patch command sufficent to use with pkgsrc.obache1-1/+4
Workaround of PR#42379, broken nbpatch package on SUA 6.1.
2009-12-19fix commenttnn1-2/+2
2009-12-19default X11_TYPE to modulartnn1-1/+6
2009-12-16Force libtool 2.2 on all platforms.joerg4-8/+5
2009-12-15Update to libtool-2.2.6b. This switches libtool to the new main branch.joerg1-11/+20
Changes are many, important for pkgsrc: - much higher speed (up to 30% faster for the overall build of libX11) - less magic for detection of the tag The pkgsrc integration of f2c has been changed to be more sane. libtool itself now depends on the f2c frontend to use proper auto-configuration and only fakes up the library. This part is encapsulated in mk/compiler/f2c.mk.
2009-12-13Update tool paths and arguments for SUA 6.0obache1-1/+28
XXX: If someone have SUA 5.2 and confirmed that it have same tools for SUA 6.0, XXX: please replace Interix-[0-5].*-* with Interix-[0-3].*-*.
2009-12-13Interix>=5.2 have libc.so.5.2, but no higer revision exists.obache1-5/+9
So gave up to detect OS_VERSION with shlib version if libc.so.5.2 exists.
2009-12-13treat interix>3 same as interix3.obache1-3/+3
2009-12-12Describe "openmpi-vampirtrace" option.asau1-0/+1
2009-12-07Linux's rpcgen has the location of cpp compiled in. Make it use thedholland1-1/+3
pkgsrc wrapper by adding -Y ${WRAPPER_BINDIR:Q} when OPSYS is Linux. This will fix obscure/subtle build problems caused by bypassing the wrappers; it will also fix pkgsrc builds on improperly configured Linux distros (such as in PR pkg/41359) whose rpcgen has the wrong location for cpp compiled into it. While there are at least two main branches of rpcgen floating around the Linux world, both of them support '-Y cpp-path'.
2009-12-04- clean up liferea mozilla optionstnn1-2/+1
- fix build with mozilla option
2009-12-03Adjust to follow f2c/libf2c split: build-depend on f2c, buildlink libf2c.asau1-3/+4
2009-11-23Make wrapper generation code work properly if there is a directory withtron1-4/+4
a whitespace in the command search path.
2009-11-21Clean up path list for ghostscript.joerg1-5/+2
2009-11-19Add description for gd optionsborrill1-0/+1
2009-11-17remove backslash.zafer1-2/+2
2009-11-17general clean-up. update and remove dead ones.zafer1-53/+10
2009-11-17Fix GNOME path on ftp.tuwien.ac.at.wiz1-2/+2
2009-11-10add xjdic-localedogcow1-0/+1
2009-11-05Don't bother with pkg_create -L. @src is essentially unused and thejoerg1-2/+1
combination of -p and -I provides all the information needed.
2009-11-03Create the work directory itself with umask 077 to prevent access tojoerg1-2/+3
files with broken permissions (e.g. world-writable files).
2009-11-02mention xrandr option (used by wm/echinus).obache1-0/+1
2009-10-31Avoid echoing commands at bin-install.asau1-2/+2
2009-10-29Make sure that MV is properly defined. From PR 42247.joerg1-1/+2
2009-10-29Add description for guile optionsborrill1-0/+1
2009-10-25Add 2-clause-bsd to DEFAULT_ACCEPTABLE_LICENSES.wiz1-2/+2
2009-10-25recognize compiler newer than 5.9, toosno1-2/+2
2009-10-21revert previous - better do a bulk build with this change first to seetnn1-1/+4
if/how much stuff is dependent on it.
2009-10-21remove lib/charset.alias from CHECK_FILES_SKIPtnn1-4/+1
2009-10-20Allow USE_JAVA=build (build dependency on jre).obache1-5/+11
2009-10-14Support lang/openjdk7 and make it the default onwiz1-5/+17
NetBSD-[4-9]*-i386 and NetBSD-[5-9]*-x86_64.
2009-10-11explain snort-preludehubertf1-1/+1
2009-10-11Fix tuwien GNU path.wiz1-2/+2
2009-10-11don't assume we have /bin/zcattnn1-1/+3
2009-10-06Use the correct variable names.joerg1-4/+4
2009-10-05Fix indentation of show-buildlink3, came from a typo.obache1-2/+2
2009-10-02Fix typo in comment.obache1-2/+2
2009-09-29* Add readline which select GNU readline explicitly.taca1-2/+1
* Remove squirrelmail-japanese and squirrelmail-lite.
2009-09-25Describe the "cvs" option. End "svn" description with a dot.schmonz1-1/+2
2009-09-24Unconditionally set the default ABI to 32-Bit for Mac OS X Snow Leopard.tron1-6/+11
There are too many third party applications which don't build using the 64-Bit ABI at the moment. Approved by Alistair Crooks.
2009-09-24The 64-Bit version of "libiconv" is not compatible with the GNU versiontron1-1/+5
and therefore rejected by e.g. the "glib2" package. Use the "libiconv" package to fix this. Approved by Alistair Crooks.
2009-09-22Workaround for amd64 vs x86_64 mess on DragonFly as well.hasso1-1/+5
2009-09-22Try again to fix Mac OS X Snow Leopard ABI issue (32-bit vs. 64-bit):tron1-1/+13
1.) Always set the "ABI" variable. 2.) Default to 64-bit mode if the machine is able to run 64-bit binaries. This seems to match the default behavior of Xcode (Apple's toolchain). 3.) Set "LOWER_ARCH" so it automatically evalutes to the correct value based on the ABI we are compiling for, not on based on the kernel the machine is running. This even works properly if the ABI is set in "/etc/mk.conf" or on the command line. Thanks a lot to OBATA Akio for providing the crucial hint to get this working properly.
2009-09-21Revert last change. It doesn't work properly because the "ABI" settingtron1-19/+1
is not defined that early as we have included neither "bsd.own.mk" nor "mk.conf" yet. We unfortunately cannot make these adjustments later because "bsd.own.mk" and "mk.conf" would be processed with an incorrect value for "MACHINE_ARCH".
2009-09-21add an option to openvpn to enable using certificates on USB sticksspz1-0/+1
or cards (etc) that are using the PKCS11 protocol
2009-09-21Try to fix Mac OS X Snow Leopard ABI issue (32-bit vs. 64-bit):tron1-1/+19
1.) Always set the "ABI" variable. 2.) Default to 64-bit mode if the machine is able to run 64-bit binaries. This seems to match the default behavior of Xcode (Apple's toolchain). 3.) Set "LOWER_ARCH" based on the ABI we are compiling for, not on based on the kernel the machine is running. The "gnupg" package now builds for the 64-Bit API without extra tricks.
2009-09-20Fix syntax. Mea culprajoerg1-2/+2
2009-09-20Avoid hardcoding sandbox location in a shell script.wiz1-2/+2