cpet/openbsd-ports
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
98,674 Commits 1 Branch 0 Tags
Commit Graph

3979 Commits

Author SHA1 Message Date
zhuk
12b19d3aa2 Catch Samba libraries not written as lib/samba/ as well as Qt and others. 2013-12-29 16:59:06 +00:00
zhuk
4263b12592 Unbreak detection of extra files in ports directory. 
The portcheck starts to really need having regression tests.
 2013-12-29 07:07:25 +00:00
zhuk
15c78efc69 Try to catch some obvious errors like: 
RUN_DEPENDS =
...
RUN_DEPENDS =

This won't catch such problems if they are spread in different files, but
should be helpful anyway now: idea from my recent typo in x11/kde4/runtime.
 2013-12-28 22:29:50 +00:00
jim
4fc0ebcd19 reserve uid for privsep in mail/greyscanner. 
ok phessler@
 2013-12-28 17:26:40 +00:00
espie
0a8e199101 future keyword, shouldn't be registered 2013-12-26 22:33:35 +00:00
zhuk
37a4b2b2f9 Warn about SHARED_LIBS inside .if statements: those could hide some 
problems, and make some other automated checks ineffective. And we
try to have are ports more declarative than imperative, anyway.

The idea was supported by espie@ a while ago.
 2013-12-21 11:47:20 +00:00
zhuk
4f3e41b9d0 Catch duplicated REVISION marks. Idea born after looking at result of 
cvs update devel/boost in my local tree.
 2013-12-21 08:59:30 +00:00
dcoppa
6f5174d059 Reserve a seat for '_sockd'. 2013-12-12 14:02:15 +00:00
zhuk
1d58015f11 Now that we have portcheck(1), start relying on it instead of doing 
the checks and pkgpath detection ourself in portimport(1).

Sitting on this diff since september, reminded by kirby@.

Tested in the wild by importing many ports, mostly KDE4-related.

Input and okay sthen@

(now committing the right file instead of ports/bin/portcheck)
 2013-12-11 16:11:09 +00:00
zhuk
1769aeb25d Oops, committed to portcheck instead of portimport. This reverts 
the accidental changes that need some discussion.
 2013-12-11 16:07:44 +00:00
zhuk
dab8a6b9ba Fix spurious "coprocess already exists" by adding some "wait" calls. 
From Bitrig via Martin Natano.
 2013-12-11 15:54:03 +00:00
zhuk
8dff51f298 Zap extra word in message. 
From Bitrig via Martin Natano.
 2013-12-11 15:53:07 +00:00
zhuk
12e075e9da Now that we have portcheck(1), start relying on it instead of doing 
the checks and pkgpath detection ourself in portimport(1).

Sitting on this diff since september, reminded by kirby@.

Tested in the wild by importing many ports, mostly KDE4-related.

Input and okay sthen@
 2013-12-11 13:55:34 +00:00
zhuk
ed6a8d0bf3 Catch missing multimedia/phonon in modules. Actual for upcoming KDE4 apps 
outside of KDE SC.
 2013-12-10 19:08:49 +00:00
espie
fe87b98440 preliminary scaffolding to handle tests: create specific task types 
for dependencies and for running tests.

fix PortInfo to know whether tests are required.
Amend PortBuilder to be able to launch tests.

(there are pieces missing in the engine, like a secondary queue with forced
affinity to run actual tests, and the portbuilder code is very incomplete
and needs love)
 2013-12-07 16:03:03 +00:00
espie
4283ce26aa work-around quote protected MASTER_SITES... 
brad and sthen didn't test thoroughly, so I get to pick up the broken pieces...
 2013-12-07 15:45:15 +00:00
espie
5e698a1f18 fix initial size match for DIST_SUBDIRS 2013-12-07 15:41:47 +00:00
naddy
e16dddbff1 toggle features that are going away with libcompat 2013-12-04 22:05:47 +00:00
espie
29e4606615 zap extra experiment, noticed by rpe@ 2013-12-02 21:47:37 +00:00
rpe
a6f54f2511 adapt to new DISTFILES format filename{url}sufx 
ok naddy@ sthen@
 2013-12-02 21:01:43 +00:00
espie
a69d8e1708 adjust distfiles handling to allow for filename{url}sufx 
to become
filenamesufx (stored name)
urlsufx (fetched url)

to be documented

okay sthen@
 2013-12-02 08:23:16 +00:00
espie
c944f3939b add "show-prepare-test-results" target, to be used by dpb for running tests 2013-11-30 16:06:56 +00:00
sthen
5037fde3bf sslsplit 2013-11-28 16:48:09 +00:00
zhuk
fce9ca5853 Oops, extra colon sneaked in previous commit. 2013-11-28 15:48:29 +00:00
zhuk
74577504a0 Some missing ${portref} in err() calls. Found while checking new KDE ports. 2013-11-28 13:38:01 +00:00
sthen
99396eb433 poison VMEM_WARNING, it's no longer used 2013-11-25 14:16:35 +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
4ef039a785 libdir based on experimental xenocara support 2013-11-17 09:43:09 +00:00
espie
54abc612eb make the data in permanent logs a bit more self-describing, so that we may add 
new fields in the future without atrocious hacks for bw compatibility.
(grows the files a bit, but quite acceptable).
 2013-11-16 16:39:28 +00:00
espie
8f841e2499 add/read timestamps to rolling logs, so that eventually old stuff gets 
removed.

alias -DSTARTUP/-DCLEANUP, since I keep mixing up the two.
 2013-11-16 13:06:00 +00:00
espie
8e8d495fb6 extra explaination of Core's role 2013-11-14 09:35:01 +00:00
espie
1b90ba4300 redo the discovery of tree yet once again, after issues found out by naddy@ 
following the bsd.port.arch.mk change. Suddenly, some paths make their way
into the build where they should not !

This actually comes from setting up wantbuild/wantinfo straight when building
path lists.

So reorganize stuff:
- don't ask for more scans during straight dump-vars reading.
- perform the "BUILD_PACKAGES may vanish from MULTI_PACKAGES" early in
merge_depends.
- do add appropriate wantbuild/wantinfo during actual dependency registration
there, which ensures depends that got ignored will be built.

This should restore proper behavior and fix long-standing semi-random
behavior.
 2013-11-12 22:11:50 +00:00
zhuk
7774a27c62 Remove check for files/directories named "core", they are not a problem 
anymore.

Pointed out by sthen@.
 2013-11-12 11:24:21 +00:00
naddy
9b97317187 spelling; ok zhuk@ 2013-11-10 14:56:41 +00:00
espie
8a3feb1a92 turns out that dpb is happy with BUILD_PACKAGES being wildy different 
from MULTI_PACKAGES, so there's no reason left to tweak MULTI_PACKAGES
directly for pseudo-flavors.

saner.
 2013-11-10 12:27:50 +00:00
naddy
394ea9b42a fix the check for long lines, simplify; ok zhuk@ 2013-11-09 23:27:30 +00:00
espie
95fbf58182 rebuild_info has a mix of pkgpath and distfiles, so we need this to 
exist.
 2013-11-08 10:48:27 +00:00
ajacoutot
5a2b6c9606 SECURITY files are gone. 
ok zhuk@
 2013-11-07 08:18:16 +00:00
juanfra
ddf83dc013 Add lzip support to the ports framework. 
ok espie@ sthen@
 2013-11-07 01:23:19 +00:00
sthen
9566cffdca new uid for _rrdcached (from net/rrdtool) 2013-11-05 17:55:43 +00:00
espie
15995508bc fix some subtle case of BULK=auto (for instance, non-regression tests in 
lang/icon/interp): needs more information than just "we're solving deps"
to know whether to clean things or not.

-> similar problems as locking oneself, similar solution, replace
_SOLVING_DEP with _DEPENDENCY_STACK that tracks the dependencies.

okay mbalmer^Waja
 2013-11-03 15:45:00 +00:00
zhuk
3b4d47508a Add checks for maximum line length in some port files. Feel free to correct 
actual numbers if needed.

Suggested by bcallah@ on ports@.
 2013-11-01 10:43:40 +00:00
sebastia
b851ec1000 reserve uid for www/opengroupware daemon user 2013-10-29 06:12:13 +00:00
espie
3bbccd56dd rearrange a few things to get direct access to whoami. 2013-10-27 09:26:01 +00:00
espie
ae15ebfef7 affinity and memory will work MUCH better for error correction if we don't 
forget to mark things in memory too, and not just in disk markers.
 2013-10-27 09:21:38 +00:00
espie
b7b1d97ddc #!#% cvs 
why did not you not show this in my diff ?
 2013-10-26 06:51:28 +00:00
zhuk
bcc21ce9ca Add a check that we do not end up with WANTLIB in arch-independent package. 2013-10-25 23:32:08 +00:00
espie
c8f6b0a82f mark most opened files as close-on-exec, since the commands we run don't care 
about them. Heck, they shouldn't even be able to access them... Ouchie.
 2013-10-25 16:54:27 +00:00
espie
4cb98fbf25 split up the fetch code for readability. 
don't bother passing logger around, just keep it in the fetch object,
the distfile objects, and use it directly, that's simpler.
 2013-10-25 13:40:42 +00:00
jasper
468e05ba33 s/beagle/armv7/ 
ok abieber@
 2013-10-22 13:19:40 +00:00