cpet/openbsd-ports
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
119,709 Commits 1 Branch 0 Tags
Commit Graph

4544 Commits

Author SHA1 Message Date
espie
ec8f1d50db improvements to WANT_CXX 
- add clang to the list of keywords (to mean ports' clang)
- take MODGCC4_ARCHS/MODCLANG_ARCHS into account if they're defined,
so we can skip one compiler if it's not appropriate and go to the next
one.

This should take care of 95% of the remaining gcc4/clang users on "some
architectures"
 2017-05-22 08:06:25 +00:00
espie
9ea01bb815 get FETCH_PACKAGES to find things by default by using installpath. 
no incidence for people who haven't updated pkg_add yet, this used to
not work, so this will still not work for them :)
 2017-05-18 12:38:46 +00:00
espie
d5c35df595 various things should refer to BUILD_PACKAGES instead of MULTI_PACKAGES 2017-05-18 12:31:20 +00:00
mpi
7d4920a17e netinet/tcpip.h is going away. 2017-05-18 11:34:35 +00:00
espie
97c599aae8 add a FORCE_REGEN env option 2017-05-17 11:31:17 +00:00
espie
391f2f7916 expose all the package-args targets, as suitable ports introspection 
targets (useful to check if some wide-sweeping change will affect
built packages)
 2017-05-17 11:16:38 +00:00
espie
23a76e674f put the >=17 there, doesn't change anything in the generated packages 2017-05-16 13:38:36 +00:00
espie
66f2fb8254 allow register-plist to display a diff even without write access to 
the plist directory.
 2017-05-14 21:24:59 +00:00
espie
cad7fff9fc when requeuing, all infos are not defined for all paths. 
this is just a heuristics, so it doesn't matter, but the least
we can do is not crash
 2017-05-14 12:43:55 +00:00
espie
b312a313ed bogus fuzz for two lines cases 2017-05-13 19:40:22 +00:00
espie
c4a397b638 ouch, WANT_CXX should trigger modules.port.mk 
didn't run into that because my test cases had other modules
 2017-05-13 12:34:01 +00:00
espie
0f3a27918d normalize just in case 2017-05-13 09:56:03 +00:00
espie
6c87e147d9 forgot to commit CXX11 list 2017-05-12 18:19:42 +00:00
espie
5605711123 support WANT_CXX and LIBECXX in sqlports 2017-05-12 18:18:09 +00:00
espie
9f13678609 fix logic for !clang 2017-05-12 16:56:59 +00:00
espie
bd2f7dd873 crude mechanism to choose gcc4 or base clang automatically 
ports clang turns out to be more complicated... we'll see what we do with
it later
 2017-05-12 16:36:01 +00:00
espie
eb05af05bc sanitize DIST_SUBDIR as well, which is actually already clean thru 
the ports tree
 2017-05-12 15:34:42 +00:00
espie
18ea6cc3d9 forbid any of the WRK* variables to end with a / 
discussed with sthen@
various things are cleaner that way
 2017-05-12 15:15:25 +00:00
espie
af335f4b66 meh. need a LIBECXX for ports that depend loosely on estdc++ on !clang 2017-05-11 21:19:42 +00:00
espie
77b31d80a4 if I want to enforce rules on WRKDIST, I'd better respect them first 2017-05-09 14:18:28 +00:00
tsg
20062d8e89 Set CONFIGURE_STYLE = modbuild tiny for p5 ports using Module::Build::Tiny. 
While at it, fix indentation.
 2017-05-09 13:37:08 +00:00
juanfra
3a84791192 Add aarch64 to ALL_ARCHS. OK espie@. 2017-05-09 11:59:38 +00:00
espie
3418da4814 filter any distfile name thru a normalize function, so that comparisons 
are more accurate and we can get normal unix filesystem semantics.
 2017-05-07 16:50:22 +00:00
espie
946422c956 make NO_CHECKSUM an option, as it causes too many problems on shared 
dist farms.
 2017-05-07 14:50:14 +00:00
espie
aa03c14e94 keep a short form of the GH_COMMIT in the DISTFILES name, so that GH_COMMIT 
changes do not result in the distfile name
 2017-05-06 15:01:31 +00:00
espie
5429e33e9e lost one line I didn't intend to lose 2017-05-06 12:43:55 +00:00
espie
8a9b58fc6c do things correctly in case wrkdist is a symlink. 2017-05-06 12:42:01 +00:00
espie
758da384cc if DISTNAME is not defined, yield more reasonable default for WRKDIST 
and DISTFILES
okay sthen@
 2017-05-05 14:48:50 +00:00
espie
580352b684 WRKDIST shouldn't end in /, but just in case... 
noticed by jca@
 2017-05-05 13:38:13 +00:00
espie
5afe4fffea my bad, line at end of summary 2017-05-04 23:40:29 +00:00
espie
8b7f708146 document current 2017-05-04 11:21:37 +00:00
espie
43545a0546 need two separate matches if PATCHORIG finishes with .orig! 2017-05-03 09:11:26 +00:00
espie
fbecc64e44 warn just once per file 2017-05-02 22:32:29 +00:00
espie
bab6db12b8 parse and fuzz patches to remove CVS keywords. Yep, we lose lines of context 
but patch can cope.
okay sthen@
 2017-05-02 21:06:41 +00:00
tsg
2e722bdb86 There's no need for 'make fetch-all' when creating a new port. 2017-05-02 11:30:03 +00:00
espie
c47da7ab68 tweak things so that changes of PATCHORIG won't matter. 
-> always emit
--- stem.orig
+++ stem
lines regardless of PATCHORIG
-> prepend an Index: line like cvs does
-> add a blank line for comments.

okay aja@
 2017-05-02 09:26:30 +00:00
espie
538a08acb7 let's be slightly more helpful and show the dirs we know about along with 
the users
 2017-04-30 20:56:19 +00:00
espie
aa623134ed run ldconfig *after* unpopulate. Makes no sense to scan /usr/local/lib 
right before emptying it.
 2017-04-30 18:49:14 +00:00
espie
d24f1b6fd6 new update-patches. 
that one deals with .beforesubst files, so it should be nicer with
SUBST_CMD handling. It also doesn't care nor emits timestamps markers
in diff, so less garbage to think about.

Please report any issue, things work better so far, but it's a work
in progress and change to make it better.

'let's finish this in tree' okay aja@
 2017-04-29 14:49:38 +00:00
espie
e2b9922e1b draft for compiler auto-selection, not ready yet. 
move arch-defines.mk up so it can be used by it to load modules
 2017-04-25 10:23:09 +00:00
sthen
4912413d0c add comments to describe CLANG_ARCHS and LLVM_ARCHS 2017-04-24 09:08:06 +00:00
espie
b437d12dc7 white line at end of summary isn't very visible. 
if we restart the build for whatever reason, stuff gets appended as
usual, but the difference between old summary and new summary should
be more visible.
 2017-04-19 15:14:45 +00:00
espie
1c3befe8b9 cosmetic change: report the hostname again in case of failure. 
I routinely read thru logs looking for the error, and forget about which
machine it was. Having it conveniently located at the end instead of the
start saves me a few seconds.
 2017-04-19 15:13:12 +00:00
espie
22d7bcaa17 have both log watching jobs actually share the timeout code. 
If I had done this before, I would have fixed kill just once after privsep
and not inadvertently broken the builder stuck timeout...

problem noticed by naddy@
 2017-04-14 16:43:40 +00:00
espie
422ed0448d reorg the logic behind permanent_log so that it *is* the permanent log. 
fixes -DNO_BUILD_STATS as reported by naddy@ (predates rewriting buildinfo,
and I plain forgot about it obviously)
 2017-04-14 16:39:32 +00:00
fcambus
e732ccd914 Remove aviion from architecture constants, the OpenBSD/aviion port has 
been discontinued after the 5.7 release.

OK sthen@
 2017-04-13 11:15:15 +00:00
sthen
a3387388c7 Give ftpsesame its own uid; it was sharing "proxy" (removed from base). 
Reported by Olivier Regnier.
 2017-04-12 21:49:32 +00:00
espie
2442a27883 forgot synopsis 2017-04-11 16:02:15 +00:00
espie
90922bae85 explain the -S option and how it's currently used 2017-04-11 16:00:53 +00:00
espie
4d6ed68975 add cached packages to the list of cleaned packages when make clean=packages 
I agree with rpe@, this was odd.

Also remove an extra / in constructed package names.
 2017-04-11 15:36:56 +00:00