97 Commits

Author SHA1 Message Date
espie
0f39741cda backsubst COMPILER_LIBCXX there as well 2017-07-13 11:46:14 +00:00
espie
1b6cc5b9f4 add COMPILER_LINKS. Tweak USE_WXNEEDED to support a 3rd option (special) 2017-06-05 14:48:09 +00:00
espie
874436213e pass COMPILER/COMPILER_LANGS thru 2017-05-31 08:39:21 +00:00
espie
5605711123 support WANT_CXX and LIBECXX in sqlports 2017-05-12 18:18:09 +00:00
espie
c7ac6aae8b hack things so that LIBCXX doesn't get expanded for sqlports.
This will probably be revisited and cleansed later.

okay sthen@
2017-04-10 19:11:11 +00:00
sthen
8098de6469 add RDEP on sqlite3, ok espie@ 2017-01-30 16:39:22 +00:00
sthen
564c6c344a Cope with moving sqlite3 back to ports. 2016-09-23 09:16:52 +00:00
sthen
91c5ca2886 Add an USE_WXNEEDED flag for ports, to write an ld wrapper script in
${WRKDIR}/bin that adds -z wxneeded to linker command lines. It won't work
everywhere but provides an easy (and easily identifiable) way to add this
flag without fiddling with build systems.

Feedback/ok jca@ jasper@
2016-08-18 12:28:53 +00:00
naddy
4297cd347a drop SHARED_ONLY from the variables available in the database; ok jca@ 2016-03-20 20:03:07 +00:00
naddy
3b0c6364b3 remove NO_SHARED_ARCHS and the resulting unreachable code 2016-03-16 15:49:36 +00:00
jasper
58de9d15d6 store PORTROACH_COMMENT
ok sthen@ (previous bsd.port.mk commit was ok with stuart too)
2015-05-14 18:01:08 +00:00
espie
3ec5ab2238 track first dependency parent so that we can display it, instead
of the user having to figure out where the broken dependency comes from
by inspecting the partial database.
2015-04-19 12:08:02 +00:00
espie
ab547da7aa add support for MAKEFILE_LIST 2014-11-04 10:44:33 +00:00
jasper
4e48ba8d2c PORTSCOUT -> PORTROACH 2014-09-05 14:45:02 +00:00
jasper
9ea47b1ad0 hookup PORTSCOUT
ok espie@
2014-09-02 14:26:39 +00:00
espie
fed0ac4a66 support GH_* variables 2014-07-12 17:35:32 +00:00
espie
bd609c9781 fix $_ usage 2014-03-21 16:45:51 +00:00
sthen
3a2dfaa992 Remove support for VMEM_WARNING, as it's manually "maintained" it isn't
very accurate, and isn't sensitive to differences between arches.
ok ajacoutot dcoppa zhuk espie
2013-11-25 14:13:33 +00:00
espie
5e63d2c808 oops, missed the conflict. Oh well, rename it correctly, and bump to 3.0,
that's an important enough change in my opinion.
2013-11-01 17:19:18 +00:00
espie
8bd186ad57 add a 3rd package that's not really a database, but just the pkgpath list
used by pkglocatedb.

mkdblite is a very small version of mksqlitedb that does NOT use DBI at all,
and can create that list only.

allows pkglocatedb to build on arches without dbi, e.g., vax.

(real benefit: separate the code a bit more)
2013-11-01 14:52:47 +00:00
espie
004c046d9d have comment reflect reality 2013-11-01 14:50:29 +00:00
espie
b3fc396048 specialize FULLKPGNAME and STATIC_PLIST variable classes, to be able to
special-case them easily
2013-11-01 14:50:11 +00:00
espie
1836068583 move the TreeWalker code to its own file, no functional change 2013-11-01 11:46:19 +00:00
espie
e3341500e6 refactor TreeWalker code for later 2013-09-04 15:35:02 +00:00
zhuk
85e57ad8b7 Unbreak fetching of SHARED_LIBS and TARGETS in sqlports.
okay espie@
2013-08-27 07:51:03 +00:00
espie
412d1b4229 clean-up the business of variables we silently ignore.
no bump, doesn't affect the built packages at all.
2013-06-23 08:58:07 +00:00
espie
247caf9db6 also store PREFIX, which allows for useful statistics 2013-06-21 10:48:38 +00:00
espie
49240e0d2f needed for error messages 2013-04-06 12:52:35 +00:00
espie
5a9e827f05 add README support 2013-03-31 09:43:13 +00:00
espie
660518085b REGRESS -> TEST, more fun here 2013-03-11 11:53:51 +00:00
espie
3f680606a7 PERMIT_* / REGRESS->TEST sweep 2013-03-11 02:52:04 +00:00
espie
fcde8ce576 remove PERMIT_DISTFILES_CDROM from the table 2013-03-08 10:05:31 +00:00
espie
9f252ccdd2 buglet fix 2013-03-08 10:04:01 +00:00
espie
082db27e18 integrate DPB_PROPERTIES
(noticed by nigel@ if memory serves)
2013-01-14 13:09:54 +00:00
espie
38b020c19c make it work even if dump-vars doesn't record empty vars 2013-01-06 21:20:58 +00:00
espie
5c8ff2cca6 there's totally no need for FULLPKGNAMEs in there 2012-10-11 08:19:02 +00:00
espie
8c682ebea4 more stuff than the DESCR says, bump rev. 2012-07-06 10:22:17 +00:00
espie
dad2db2c22 clean-up ports-readmes special-case 2012-06-18 12:18:45 +00:00
espie
19aa179987 tweak: MULTI table now holds pointer to the corresponding subpackage path 2012-05-22 12:04:02 +00:00
espie
e79b763a4e more surgery:
- don't ask the database about existing pathkeys, it's simpler to have
our hash.
- in dependencies, don't like for non normalized pathkeys
- add CANONICAL column, and fill it.
2012-05-20 11:06:07 +00:00
espie
cea1d3a0f3 bite the bullet, reuse dpb's algorithm to make sure we get correct
PkgPath.

This implies restructuring the database code, and the start of a
better separation between variables handling, and actual insertion
in the database.

bump version to 2.0: this changes the paths that actually get recorded,
and pkglocatedb will need to be aware of that.
2012-05-18 12:11:28 +00:00
espie
6fb048701d if this pipe exits with an error, it means some subdirs could not be
scanned. That's not a warning, that's a hard error.
2012-05-08 17:22:00 +00:00
espie
bc5cab5e8f +VMEM_WARNING 2012-04-29 16:34:53 +00:00
espie
95fefd0c50 make sure we don't pass FLAVOR/SUBPACKAGE info thru subdirs.
count number of passes (purely cosmetic)
2012-02-17 07:31:26 +00:00
espie
c757bb306d fix description of DEPENDS table. 2012-01-28 11:40:34 +00:00
kili
0132b0aa29 Fill in the master_sites* columns of the sqlports.ports table.
From espie@, after a short discussion on icb.
2011-11-26 22:30:30 +00:00
espie
fcbd4ac01b better: use REPORT_PROBLEM_LOGFILE to have nicer error messages.
use new FULLPATH setting from bsd.port.subdir.mk to avoid having to fiddle
too much with options.
2011-09-24 07:37:34 +00:00
espie
8388e334df store PSEUDO_FLAVOR along. tweak descriptions a bit. 2011-09-10 08:48:15 +00:00
jasper
61f9ed9482 adjust after recent lesstif removal
ok aja@ sthen@
2011-04-10 18:00:23 +00:00
espie
59488755a7 tweak the way mksqlitedb work: keep looking at dependencies after we do
the main tree.

initial work by florent tribouilloy

a bit of clean-up from me.
2011-03-02 16:19:54 +00:00