4621 Commits

Author SHA1 Message Date
espie
fd096fa2e2 glue that invokes check-wrkdir optionally 2017-11-24 14:48:33 +00:00
espie
0d3c66c241 a new very optional script to check various things during build
(mostly permissions for _pbuild and files changed during fake)
2017-11-24 14:47:26 +00:00
espie
08ecf94982 finish killing PORTS_BUILD_XENOCARA remnants 2017-11-24 14:26:20 +00:00
espie
43a36a1093 save the full path used, eventually for better restart, but not yet 2017-11-24 14:24:04 +00:00
espie
10c8e68ebc - add FIX_EXTRACT_PERMISSIONS, for ports that have paranoid archives
so that building with _pbuild does not make unreadable.
- zap MTREE_FILE, no longer useful
- rectify fake tree permissions under "normal user" fake stage.
okay jca@, naddy@
2017-11-23 18:11:05 +00:00
jca
4593b2842d Remove powerpc from LLVM_ARCHS, it never built there.
This should unlock the ~235 ports that try to use ports-clang before
ports-gcc.

ok bcallah@ landry@
2017-11-23 14:33:21 +00:00
espie
22b9cc686e kill PORTS_BUILD_XENOCARA in dpb as well. 2017-11-23 12:36:25 +00:00
espie
d0be71e5d3 kill the experimental xenocara build
I still haven't managed to make it fast enough to be worth it,
and I'm unlikely to do so, so bye bye.

already discussed a bit during p2k17, where I said "I'll think about it".
Well...
2017-11-23 12:29:15 +00:00
espie
7879b88914 as we do the install-wrapper twice, don't copy it the second time if it's
identical
2017-11-23 12:10:36 +00:00
espie
b0ac698a0b group more stuff after post-patch
okay sthen@, aja@ and a full bulk build
2017-11-19 17:31:04 +00:00
naddy
9ec48599e0 remove the backward-compatible COMPILER values as they are no longer used 2017-11-17 00:24:01 +00:00
espie
03bb654c59 allow better integration of ports tree with dpb.
if PORTS_PRIVSEP=Yes, use FETCH_USER and BUILD_USER (defaults to
_pbuild/_pfetch) for common directories. Includes
- fetching stuff into DISTDIR
- building packages under PACKAGE_REPOSITORY
- also caching packages
- running register-plist

the extra complication is reasonably small, the largest part being the
gap between pkg_create to tmp file as regular user and moving into
_pbuild.

(regular pkg_create has to be the same user as make build, because of
ports like boost where the archive has no read rights for anyone but
the user active during extraction)

The full build is not privsep'd (yet, I have to see how complicated this
gets), but with this, using the same machine for regular ports build and
for dpb usage becomes easier.

(note that dpb proper sets PORTS_PRIVSEP=No since it's doing all the privsep
on its own)

as discussed with pirofti@, sthen@ during p2k17

okay pirofti@
2017-11-13 14:01:44 +00:00
espie
cfd0c2446d prepare dpb for better ports tree integration: if we got started as root,
we're already privsep'd so don't do it again
2017-11-13 13:51:21 +00:00
espie
05f9d086bf if we user a non-standard builduser/fetchuser, mark that in mk.conf
also deal with "fun" devno (I've had this in my tree for ages)
2017-11-11 13:57:04 +00:00
ajacoutot
cf74427e0d Enable fp_cv_const_EBADMSG and fp_cv_const_EPROTO now that we have proper
definitions.

bulk tested
ok naddy@ sthen@
2017-11-10 07:10:02 +00:00
naddy
570e9d2a8f MODGETTEXT_LIB_DEPENDS, MODGETTEXT_RUN_DEPENDS, MODGETTEXT_WANTLIB are no
longer used
2017-11-06 20:43:22 +00:00
espie
e721a43980 define COMPILER_LINKS early so that modules can test it easily 2017-11-05 09:10:13 +00:00
landry
a2c4435970 reserve uid/gid 794 for _burp user/group 2017-11-03 11:49:58 +00:00
naddy
73f098c47c <net/slip.h> has been removed 2017-10-31 14:39:10 +00:00
naddy
fcc553acf5 Add missing ARCH and MACHINE_ARCH values to ALL_ARCHS.
Only use MACHINE_ARCH values for GCC[34]_ARCHS.  This seems to be consistent
with the other entries.
2017-10-26 19:40:32 +00:00
espie
96de49bdff use pkg_create -S so that port sigs and package sigs are handled
identically.

Along with removing the XXX in out-of-date, this also fixes dpb update
modes.
2017-10-11 14:27:20 +00:00
espie
d38cac9c54 don't die if we can't print this error message.
Seems that some people are heavy handed on the build cluster, as
experienced by phessler@
2017-10-05 22:41:32 +00:00
espie
bd1fb6f1ca oops, ports without depends have @version at the end. 2017-09-28 19:10:09 +00:00
espie
9a82bf35d3 have out-of-date ignore the @version part for now.
properly handling that is more complicated and can wait until after 6.2.

problem pointed out by naddy@
2017-09-28 10:30:35 +00:00
tb
08a014ca9c Display a clear error message indicating that WRKOBJDIR must be on a
wxallowed filesystem for USE_WXNEEDED ports instead of erroring out
unhelpfully.  Prompted by numerous questions about Python builds failing
with "configure: error: cannot run C compiled programs." on ports@

Input, tested in bulks and ok sthen; ok danj; no veto from espie.
2017-09-18 16:32:06 +00:00
espie
f0a5685b2c activate @version 2017-09-18 15:20:54 +00:00
espie
fbc2afa76d scaffolding for -V option, the only missing part is activating it in
bsd.port.mk
2017-09-18 13:02:34 +00:00
espie
86839ea99b don't count @version as relevant when registering plists. 2017-09-18 13:01:43 +00:00
robert
f00a70a056 no more _zarafa user 2017-09-13 10:45:20 +00:00
robert
60db48d44c add _kopano user 2017-09-13 10:17:45 +00:00
sthen
b0bb52643e Add LLD_ARCHS, listing arches using LLVM's ld.lld (rather than binutils'
ld.bfd) and a PROPERTIES matcher. Suggested by phessler@, "go for it" espie@
2017-09-09 10:22:45 +00:00
espie
f92d7101b9 turns out we already copy checked packages from cache into the normal
repository. keep that model for now, but make FETCH_PACKAGES more consistent
with respect to signature handling (e.g., use TRUSTED_PKG_PATH for the
local repo, and pass TRUSTED_PKG_PATH thru)

should help naddy
2017-09-05 14:50:28 +00:00
espie
11296357e7 always pass state thru 2017-08-29 14:22:10 +00:00
espie
f96c91f68c a few more bits 2017-08-24 13:25:38 +00:00
sthen
619d6c2b0d Unbreak. 2017-08-22 19:40:59 +00:00
tj
f586485c58 prioritize https over http/ftp for distfiles.
ok aja jasper for gnome, aja for gnutls, giovanni for gimp, sthen for the rest
2017-08-22 18:03:51 +00:00
espie
7737872aef rework COMPILER yet again. new version should be easier to grasp 2017-08-22 10:27:33 +00:00
espie
5690c3546f centralize arch compiler support so we cant teach COMPILER to
set ONLY_FOR_ARCHS
2017-08-21 09:12:47 +00:00
espie
1f0907b562 even though those names shouldn't be strange, doesn't hurt to protect against
-something names.
2017-08-21 09:10:52 +00:00
ajacoutot
c99f19ce29 Fix UID/GID of the mcastproxy user and register it in the user db.
from Paul de Weerd (maintainer)
2017-07-31 08:03:40 +00:00
jasper
aa7bfe242b make an appointment for _orthanc 2017-07-29 18:48:44 +00:00
espie
6ee8bff8bd switch ports as well 2017-07-26 21:25:32 +00:00
sthen
0c3a4d7394 Poison autopoint if gettext-tools is not a listed dependency, similar to
existing poison for msgfmt/msgcat/msginit. Makes sense to naddy@
2017-07-26 14:21:22 +00:00
espie
9952cda353 use -F pthread, so that COMPILER_LIBCXX doesn't trigger "extra" warnings 2017-07-23 09:27:15 +00:00
espie
c114d7057b add pthread to COMPILER_LIBCXX.
white lie, but it allows clang and gcc to be more similar
bump accordingly.
2017-07-23 09:26:25 +00:00
sthen
174ac67aa2 Switch to v1 API URL, minimal patch to fix portgen as the v0 API has been
removed - v1 appears to be compatible enough so far.  ok tsg@
2017-07-20 20:42:41 +00:00
espie
9b2efb46ff direct inclusion of MODULES=clang on a clang arch triggers double links
so protect against these as well
2017-07-20 07:31:48 +00:00
espie
a76fd1148f anchor regexp (pedantic fix, I'd have to craft a very specific example
to make this break)
2017-07-19 14:17:20 +00:00
espie
9113bc87c7 add sugar to have WANTLIB contain the right mix of
COMPILER_LIBCXX/LIBCXX/LIBECXX in most cases.

... and  future support for pthread variations.

okay sthen@
2017-07-19 14:16:13 +00:00
zhuk
b5c8dd973c Catch QML-specific errors, inspired by recently discovered breakage. 2017-07-14 07:42:51 +00:00