cpet/openbsd-ports
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
62,038 Commits 1 Branch 0 Tags 554 MiB
2c0a53e285
Commit Graph

1252 Commits

Author SHA1 Message Date
ajacoutot
66223a71e6 Add _AUTOMAKE_SPEC for automake 1.10. 2010-07-22 19:14:09 +00:00
steven
715dacfb85 pass COMPILER_VERSION in MAKE_ENV instead of USE_GCC3. 
this removes the last instance of USE_GCC3 in ports.

from brad
 2010-07-21 17:11:48 +00:00
steven
952b3ac19a drop USE_GCC4. we don't use this and the module checks COMPILER_VERSION. 
from/discussed with brad
 2010-07-21 16:45:21 +00:00
steven
26f69c2f4b use COMPILER_VERSION instead of USE_GCC3 
from brad
 2010-07-21 16:42:52 +00:00
naddy
a2c9a774ff Remove redundant variables OPSYS and OPSYS_VER. 
OPSYS was always "OpenBSD", the ports tree doesn't cover other
operating systems.
OPSYS_VER was the same as OSREV.

ok espie@
 2010-07-18 18:49:01 +00:00
steven
ba3f871673 new WANTLIB,LIB_DEPENDS 2010-07-18 11:10:44 +00:00
espie
0af1c9e7da fix fix 2010-07-10 15:21:15 +00:00
espie
11d3fea7e4 some recursive scenarios would break, let _ARCH_OK be undefined, catch it 
later.
 2010-07-10 15:17:40 +00:00
espie
25c3ed92c3 always define _ARCH_OK${SUBPACKAGE} (to 0 by default), so 
that bsd.port.mk can rightfully complain about wrong subpackages.
 2010-07-10 13:29:48 +00:00
espie
188c8ed5f5 new support for REVISION/EPOCH 2010-07-10 09:11:10 +00:00
espie
d170ea5206 logic of ONLY_FOR_ARCHS was slightly broken: always define individual 
PACKAGE_COOKIE, so that print-plist-contents keeps working.
Only the packaging part (e.g., PACKAGE_COOKIES) should be affected.
Alter PKGNAMES while we're there, as it's only used for debugging, and
it's better that way.

(fixes pkgmklocatedb)
 2010-07-09 13:11:59 +00:00
espie
d9dd312bbf synch with src 2010-07-08 20:48:39 +00:00
espie
62995c89d4 make ONLY_FOR_ARCHS and NOT_FOR_ARCHS subpackage-dependent 2010-07-06 12:38:26 +00:00
espie
da03d678c9 print-plist-contents | grablibs appears often enough: combine them as 
print-plist-libs
 2010-07-06 12:09:55 +00:00
espie
45a2880841 allow libspecs to come from current ports internal dependencies (even if 
they're not installed yet)
 2010-07-06 12:02:35 +00:00
espie
e946b4c62d explicitly document that the .spec files are just there to force reevaluation 
of cookie during development
 2010-07-06 11:50:57 +00:00
espie
8ea8f2a482 don't bother computing pkgname if all we want is the plist from the pkgpath 2010-07-06 11:27:38 +00:00
espie
3a7ee8c966 SUBST_CMD is not subpackage dependent, so don't. 2010-07-05 09:00:28 +00:00
espie
d6f6f17460 some libs may be only static libs. we find them correctly, but passing these 
as -W makes no sense.
 2010-07-05 08:58:09 +00:00
espie
63e9ac2f46 tweak the way LIB_DEPENDS and WANTLIB interact. 
More precisely:
- for "older" libspec in LIB_DEPENDS, only compare them against
the libs from the LIB_DEPENDS.
- check LIB_DEPENDS against all libspecs for the package (both from WANTLIB
and LIB_DEPENDS) to turn them into LIB_DEPENDS.
- create a full list of installed libs and of 'inter-package' libraries to
solve all WANTLIB in one go.
This is actually faster than the old code, but a bit more complicated.
 2010-07-04 17:26:14 +00:00
espie
f8a41e2fb7 slightly cleaner logic structure for building packages. 
In particular, create a tmp package until we're certain it is okay
(thanks to register-plist). -> remove race condition in dpb3, which
assumes a package existing is enough for it to proceed.
 2010-07-04 08:45:02 +00:00
espie
132a204bdb activate locks by default, in /tmp/portslock 
from a strong suggestion by ajacoutot@, after discussion with various
people.
 2010-06-20 07:48:20 +00:00
william
eb2f81bc10 Remove the WRKDIR_LINKNAME symlink at 'make clean' if present 
help and ok espie@
 2010-06-17 03:31:33 +00:00
espie
5d22cfc900 zap old stuff (can't work, since -m is used for something else now) 2010-06-16 12:22:20 +00:00
espie
c1fa9429b1 remove PKG_CMD, set up PKG_ADD/PKG_CREATE/PKG_DELETE/PKG_INFO 2010-06-16 12:06:46 +00:00
kili
e0fb41c023 Move `-q' flag for pkg_info(8) from PKG_QUERY back to the two places 
where it's really needed. Fixes ${_UPDATE_COOKIE${_S}} (i.e. make
update).

"oh oops. yep. Fuck." espie@ :-)
 2010-06-15 11:42:21 +00:00
espie
bcaeea964d fix directory creation so that PLIST_DB can be a path, as should be. 2010-06-14 12:02:43 +00:00
espie
c347a26fa6 reorg pkg* thru _PKG*, so that we can add a PROGRESS_METER user settings 
(defaults to yes, as for ftp).
 2010-06-14 12:01:19 +00:00
espie
d26ccc25a2 simplify check-lib-depends run 2010-06-14 11:49:03 +00:00
espie
90ca9c763d BREAK the tree in a very obvious way for people who did not remove 
/usr/bin/f77 and /usr/bin/g77
 2010-06-05 16:38:47 +00:00
espie
e6e3083ecd if IGNORE'd ports are used as dependencies, write the IGNORE message in an 
_IGNORE_COOKIE file.

React accordingly in dependency handling, namely fault the port that tries
to use IGNORE'd stuff.
 2010-05-28 12:34:22 +00:00
espie
a7f09f8eca our mkdir -p is thread-safe, so use it (tested for a few months) 2010-05-28 10:42:09 +00:00
espie
5bdc8e686d amd64 and sparc64 have switched 2010-05-28 10:09:58 +00:00
espie
00ee958882 prepare old compiler too 2010-05-25 10:10:43 +00:00
espie
28eb674094 f77 no longer lives in source 2010-05-25 00:27:06 +00:00
espie
d932c34a34 also count libf2c as a build depends: it's needed for a working g77 
compiler, but some ports end up building libraries that do not depend on it
 2010-05-23 09:40:19 +00:00
espie
bfb7e8e53c fix 2010-05-23 09:28:19 +00:00
espie
7dd09beb4d only warn for empty patches, instead of erroring out. 
Simplify the task of testing patches that remove patches.
Don't forget to cvs rm though !
 2010-05-23 09:22:50 +00:00
espie
a2c438edf9 prepare the move 2010-05-23 08:59:36 +00:00
espie
5f9bf53f3f classify machines per compiler arch. May occasionally be useful if you're 
looking for something like this.

some help from miod@ for hw that's been forgotten by every human. ;-)
 2010-05-16 10:33:32 +00:00
espie
e00097ed08 also prevent SUBDIRLIST from escaping, as noticed by kili@ 2010-05-09 21:41:40 +00:00
espie
303bd978e6 pattern I use a lot: start right after a given dir. 
Easy to do.
 2010-05-01 10:27:56 +00:00
ajacoutot
056a331311 Override MAKE in CONFIGURE_ENV when USE_GMAKE is set to Yes. Some ports 
seem to need this more and more and it makes sense anyway.

tested in a bulk by jasper@, ok espie@
 2010-04-24 09:54:42 +00:00
espie
3f48860b4c fix and recommit, my bad. 2010-04-20 21:00:35 +00:00
landry
9a3be8dc0f Revert to r1.991, this temporarly broke the portstree.. 
agreed by kili@ and sthen@
 2010-04-20 19:58:09 +00:00
espie
09ed1dcb61 use _MAKE and _MAKESYS, simpler to read 2010-04-20 10:15:09 +00:00
espie
8565ca1737 provide a _MAKE internal that does the cd ${.CURDIR} && exec, 
use it to avoid the obj/ problem in /usr/ports.
provide a _MAKESYS internal for bsd.port.mk with SYSTRACE too
 2010-04-20 10:03:36 +00:00
espie
5b287b6cbf make sure SUBDIR doesn't escape into individual ports directory. 2010-04-20 10:02:39 +00:00
espie
8beda10cb7 fix logic, so that BROKEN-arch shows in the correct places 
when MULTI_PACKAGES is set.
 2010-04-17 10:16:11 +00:00
espie
51b217c20a forgot vax 2010-04-17 09:49:52 +00:00
espie
6d10bb1b49 add a list of ALL_ARCHS so that I can dump BROKEN-<arch> in dump-vars. 2010-04-12 13:08:20 +00:00
espie
39f6cafc40 fix buglet 2010-04-05 14:03:26 +00:00
espie
03931ebeb5 USE_GROFF framework for the mandoc switch 2010-04-05 14:02:50 +00:00
espie
c52b023249 don't know how this broke, but make it work again (and more sturdy against 
multiple commas)

noticed by aja
 2010-04-05 13:59:03 +00:00
espie
abbf59d6e3 typo, noticed by Andreas Kähäri <ak@ebi.ac.uk> 2010-03-22 20:19:12 +00:00
espie
a07ce53b70 use :Q to avoid any problem 2010-03-21 17:12:10 +00:00
espie
0a363398c3 ouch, pass TERM only if it's defined. Add TERMCAP while there. And do the same 
thing to http_proxy/ftp_proxy, please...
 2010-03-21 17:00:46 +00:00
ajacoutot
c4c7ece3ba Add a new APM_ARCHS variable. 
comments from espie@ naddy@ jasper@
"looks good" jasper@
 2010-03-21 11:34:38 +00:00
espie
cd50be5feb move to newer style -D for pkg_add. 2010-03-20 19:14:49 +00:00
espie
ae1f947e6d pass TERM to pkg_add so that Term::Cap will be able to use more capabilities 2010-03-20 19:14:06 +00:00
espie
feec1c7f7a introduce BROKEN-arch 2010-03-20 19:11:51 +00:00
espie
228149426e support for more stats in dpb, basically harmless 2010-03-05 07:49:29 +00:00
espie
a090480935 TRY_BROKEN, idea from FreeBSD (by popular demand, jasper + phessler) 2010-02-26 19:20:24 +00:00
espie
7a81866fd7 warn about FULLPKGNAME without FULLPKGPATH, as it is a bit difficult to spot 
in complicated ports
 2010-02-26 19:05:23 +00:00
espie
9e0b3d6c97 don't iterate on IGNORE, this breaks java's jrl message 2010-02-13 21:11:35 +00:00
espie
f7f5b4ff0a some important improvements to bsd.port.mk 
- simplify IGNORE handling, always define it, even when NO_IGNORE is set,
and only use it for the targets it's meant to influence.
- make it possible to multiply-IGNORE a port, so stuff that's BROKEN and
NOT_FOR_ARCH will show up as both.
- DESCRIBE_TARGETs can die, since we no longer need special IGNORE dance
for them.
- add an IGNORE_IS_FATAL tweak, so that ignored ports can actually error out.
- reorg dump-vars to choose what to show, including new IGNORE, and COMES_WITH.
- name an explicit prepare step, which does check build dependencies and
install them (useful for build timing purposes)
- move the do-fetch test out, so that we can add pre-extract and do-extract
there eventually.
 2010-02-12 12:00:19 +00:00
jasper
923a684afc - mention mips64el where applicable 
ok espie@ ajacoutot@
 2010-02-05 13:06:03 +00:00
espie
2a6cec5724 duplicate RUN_DEPENDS, it should be indexed only. 2010-01-28 10:09:06 +00:00
steven
191ff969ee allow USE_LIBTOOL to be set to 'gnu'; 'gnu' and 'yes' do the same for now. 
ok espie
 2009-10-14 13:01:03 +00:00
landry
fb2acb23e7 Change WRKOBJDIR (again...) to default to ${PORTSDIR}/pobj instead of obj, 
it confuses make mirror-maker (among others..)
req'd by espie@, agreed by several@ in budapest.
 2009-10-13 14:39:23 +00:00
simon
456bb1cca4 Finally add CONFIGURE_STYLE=modinst. This adds p5-Module-Install to 
BUILD_DEPENDS, and --skipdeps to CONFIGURE_ARGS.  Additionally it nukes
existing ${WRKSRC}/inc/Module/*Install* files in pre-configure to force
regeneration of bundled and often outdated Module::Install versions.

feedback/testing by msf@, jim@ and landry@
 2009-08-12 22:36:48 +00:00
espie
5552fa34d2 maintainer usability: some targets are usually not invoked recursively, 
but they can (sometimes) be useful: add DANGEROUS knob to prevent running
them accidentally, but allow people to run makesum/update-plist/update-patches
recursively.
 2009-07-26 12:14:05 +00:00
espie
ef63e25ff6 small mirror-maker tweaks: 
- there's no need for :: on pure dependencies, : works just fine for
accumulating (and :: is not 100% standard on all systems)
- append directly to _FETCH_MAKEFILE_NAMES. This avoids an untidy blank space
 2009-07-15 23:44:36 +00:00
espie
3ea520053a when using peek-ftp, I often start with make show=DISTFILES, so save myself 
the trouble.
 2009-07-13 12:21:44 +00:00
simon
e47a3e8294 let Module::Build tests behave like ExtUtils::MakeMaker ones 
ok landry@, sthen@
 2009-06-18 13:41:28 +00:00
landry
27251dbdc9 systrace isn't too happy when WRKOBJDIR contains a symlink, which can 
happen often (/usr/ports NFS mounted & /usr/ports/obj a symlink to a
local dir, /usr/ports as a symlink to /home/wherever..)
Use readlink -fn only if USE_SYSTRACE is set so less things are run when
bsd.port.mk is included. Should make everyone happy.

Issue reported by and original fix ajacoutot@, sounds reasonable to espie@
 2009-06-17 13:42:49 +00:00
sthen
df3fbaad7d set RECURSIVE_FETCH_LIST to No by default. ok sturm, ajacoutot, espie. 2009-06-12 17:26:51 +00:00
landry
861925b221 Provide saner defaults for WRKOBJDIR, using ${PORTSDIR}/obj. No more w-* 
dirs forgotten all around. One can still unset it in /etc/mk.conf to get
back to the older behaviour, or still override it with /usr/obj/ports or
/usr/wobj or whatever. At least, now it can be a separate partition.
Most of us were already overriding this value, OpenBSD is all about sane
& simple default settings.
${PORTSDIR}/obj suggestion from jakemsr@.
Discussed at length with several, 'Works for me' krw@ ok todd@ wcmaier@.
 2009-06-12 13:39:30 +00:00
espie
3cb8b96f11 don't export brutally everything from bsd.own.mk into each port's Makefile, 
but select some variables instead.

checked by naddy@ on a bulk-build.

(apart from making stuff less easy to break, it also means Makefiles stop
inheriting a WIDE set of .PHONY targets, which is a good thing)
 2009-06-09 17:46:58 +00:00
simon
f9808159d2 allow ports to set REGRESS_IS_INTERACTIVE=X11 in order to get rid of 
redundant makefile parts found at many places in the tree

DISPLAY and XAUTHORITY variables/checks to be nuked soon from many ports

ok ajacoutot@, sthen@, jasper@
 2009-05-16 22:18:50 +00:00
martynas
fb518f37e5 make "doesn't seem to exist on this system" and "Size matches for" 
implicit;  since they filled the whole screen making important stuff
unnoticable.  ok espie@, sthen@
 2009-05-05 20:58:38 +00:00
espie
3286c9fba0 remove trailing dot 2009-04-24 08:53:33 +00:00
espie
7112ac8117 add support for MATCHDIR: will only recurse into dirs that match the 
pattern...
 2009-04-11 14:56:31 +00:00
espie
b466c39b7a minor tweaks: 
- totally forbid overriding do-fetch
- slightly better diagnostics in a few cases
 2009-04-11 14:55:31 +00:00
ajacoutot
21e51b9ca9 - add old behaviour back where MODGNU_AUTOMAKE_DEPENDS is always set even 
when not using CONFIGURE_STYLE=automake
this fixes ports depending on a specific version of automake while e.g
using the gnu CONFIGURE_STYLE

ok espie@
 2009-03-14 10:46:56 +00:00
espie
5bcba4cec0 fix automake dependencies to read like automake->=1.4,<1.5 instead 
of automake-1.4.*
 2009-03-09 23:07:33 +00:00
espie
db7f3a147c say exactly WHAT was skipped, instead of giving the pattern 2009-01-24 11:32:43 +00:00
ajacoutot
126913162c - fix some unquoted shell expansions 
this fixes an old bug where package creation would fail if one has some
particular file under PKGPATH because it would give pkg_create the wrong
args

from Wim Lewis on ports@, thank you!
"looks right" sthen@, "sounds reasonable" landry@, ok jasper@ espie@
 2009-01-05 12:43:49 +00:00
espie
65e7dd9971 somewhat better error indications after some user feedback ;) 2008-11-18 11:45:41 +00:00
ajacoutot
3b893bbd83 - revert; USE_GCC* does not belong to the gcc* MODULES usage and 
the man page has always been wrong about it

requested by and ok steven@
 2008-10-28 13:29:05 +00:00
ajacoutot
c78568ec50 - reverse logic so that gcc3/4 is actually used when USE_GCC* is set 
- set USE_GCC* to Yes by default when using a gcc* MODULE which makes
much more sense

discussed with and ok robert@
 2008-10-28 12:26:39 +00:00
bernd
5f93c223a6 Add SUBPACKAGE to _ALL_VARIABLES. ok espie@ 2008-10-25 15:06:26 +00:00
simon
f06c5645e2 remove dependencies to ports now coming with perl 2008-09-29 22:55:46 +00:00
espie
79bdb256c9 add a possible convenience link to the WRKDIR 
*solely* as a facility for people working on the ports tree.
not to be used for anything at all in the ports tree itself !!!
since it's not protected against any kind of race.
 2008-09-19 13:00:30 +00:00
ajacoutot
13ef0005f6 - add ^TRUEPREFIX to the generic SUBST_VARS 
"go ahead" espie@
 2008-09-17 13:42:10 +00:00
simon
bd1e06cf76 Redirect regression output so we get the actual exit code instead of the 
last one in the pipe to make sure failed regression tests don't bake a
cookie when REGRESS_LOG is set (default).
Also redirect stderr to regress log and introduce REGRESS_STATUS_IGNORE
variable so post-regress targets like found in cpan.port.mk can do the
actual failure detection.

Problem noticed by bluhm@, latching of exit code suggested by steven@

Testing, feedback and ok on earlier version by bluhm@, 'go for it' espie@
 2008-08-20 10:33:50 +00:00
espie
9442611200 add clean=build: incentive to use SEPARATE_BUILD, as this allows you to not clean up the source, but only phases from configure up... 2008-08-20 08:56:53 +00:00
espie
409b0bc6fb NOT_FOR_ARCHS as dump-vars, to allow sqlports to perform better 2008-08-19 23:49:32 +00:00
espie
7d70a8b03f typo 2008-07-29 17:54:52 +00:00
espie
7c4a60e481 we have too many old distfiles -> too many links. 
instead of storing links directly as
sha1/digest
allow them to be in
by_digest/sha1/di/digest

so the number of links will not grow that fast
 2008-07-29 11:25:44 +00:00