Commit Graph

54 Commits

Author SHA1 Message Date
pascal
c648bb387c Deal with hidden dependencies libmagic and libexecinfo, found by Mikolaj
Kucharski:
- disable detection of magic.h
- record build dependency on devel/libexecinfo.  Note that this affects only
  profiling features of libstdc++, enabled by defining _GLIBCXX_PROFILE.

ok espie@
2012-07-01 18:28:40 +00:00
sthen
49a55c3e0f bump for plist changes 5.1 -> 5.2 2012-06-21 13:56:57 +00:00
pascal
a51e30b3c9 Fix build with Brad's sysconf(3) diff, found by naddy@ in his bulk build. 2012-05-14 23:39:39 +00:00
pascal
21fda32a64 Don't hardcode GC_nprocs to 1 in boehm-gc, but get it from sysconf instead,
following a similar change in devel/boehm-gc.
From Brad.

ok kurt@
2012-05-07 10:34:33 +00:00
pascal
46702013ca Do not let different versions of libstdc++ conflict with each other. At the
same time, make sure that g++ is never installed alongside a higher version
of libstdc++, so that stuff will always link to the correct library.

feedback espie@ and kurt@, ok kurt@
2012-05-04 08:47:55 +00:00
pascal
b15c7fe082 - define _REENTRANT, not _POSIX_THREADS for -pthread
- sync kprintf formats with base

looks ok to kurt@
2012-04-23 14:16:34 +00:00
pascal
3f726ecffb Enhance this module to provide ${MODGCC4_CPPLIBDEP} and ${MODGCC4_CPPWANTLIB}.
ok robert@
2012-04-18 18:41:22 +00:00
pascal
0e321f98b8 Defaulting to march=i486 in the compiler doesn't help much if the driver
overrides that setting.  Sync both to use i486.  Fixes libreoffice and chromium
on i386, tested by both me and naddy@.
2012-04-09 11:55:53 +00:00
pascal
589e02ed65 Fix previous to work when building java, but not go. Noticed and patch by
Andre Zepezauer on ports@ (thanks).

ok sthen@
2012-04-08 12:40:51 +00:00
pascal
52678a89c6 Also build libffi on arches where java isn't supported, it's needed for libgo.
tested on powerpc and ok aja@
2012-04-06 07:50:32 +00:00
pascal
826b4aa211 Apply the same fix as to base gcc wrt -fident.
ok espie@
2012-04-02 15:28:04 +00:00
pascal
c263e2ec14 Previous commit actually fixed gcc 4.6 on sparc64; just fix one stupid typo.
Tested on one of weerd@'s machines (thanks).

Also, update TODO.

ok espie@
2012-04-02 08:51:30 +00:00
pascal
7b9d7a0837 forgot to cvs rm 2012-03-31 20:06:40 +00:00
pascal
e7f261aa15 Update to gcc 4.6.3.
Also, sync a bit with base gcc:
- disable unsafe builtins
- disable -Wpointer-sign in -Wall
- move -fstrict-aliasing to -O3, disable -fstrict-overflow and
  -fdelete-null-pointer-checks
- enable -Wsystem-headers
- default to march=i486 on i386
- fix crtsavres on powerpc
- sync sparc64 config
- implement -Wvariable-decl, -Wstack-larger-than-N (alpha, amd64, i386, sparc,
  sparc64), -Wtrampolines, -ftrampolines
- default to -fno-ident
- fix -rdynamic by passing -export-dynamic to ld(1)

some bits from Brad, others from me.

tested by myself and jsg@, adastrap now hosted by sthen@ (thanks).

ok jsg@ sthen@
2012-03-31 20:02:03 +00:00
kurt
d4c1807932 enable __cxa_atexit. okay pascal@ 2012-03-07 20:39:37 +00:00
pascal
bdd5b86d65 fix boehm-gc with rthreads for gcc 4.6. based on kurt@'s work.
ok espie@
2012-02-28 20:00:16 +00:00
pascal
fe79b79bb5 Fix build of libgo on powerpc by:
- backporting a fix for the ppc backend from upstream
- telling libgo configure not to try to link to libffi_convenience on ppc

issue brought up and fix tested by ajacoutot@, "go ahead" kettenis@
2012-02-26 22:32:36 +00:00
pascal
378165bc9a Fix -ada build by providing a dummy fname_as_string() in gnat, missed in the
previous commit when enabling ssp.

Reported and fix tested by naddy@.
2012-02-21 08:09:30 +00:00
pascal
72bbe18ad2 - modify stack protection to work with our native libc routines, passing the
function name to __stack_smash_handler.  Only supported for "c-family"
  languages at the moment.
- enable it per default on archs where FRAME_GROWS_DOWNWARD
- remove obsolete comment about adastrap needing a specific libc
- -java actually needs a runtime dep on -main for -lgcc

ok espie@
2012-02-16 21:30:15 +00:00
landry
878a92f4a1 Mark gcc 4.6 as BROKEN-sparc64, gengtype loops infinitely.
ok espie@ sthen@
2012-02-03 17:01:17 +00:00
pascal
be59426cab - bump libestdc++ following the bump in gcc 4.2
- deal with ldconfig shitz being run at fake install everywhere
- fix sprintf/strcpy usage in libestdc++ so we don't get a warning every time
  we compile a C++ program

Everything brought up by espie@.
2012-01-26 13:17:22 +00:00
espie
7ac6808f15 fix for !i386, sorry 2012-01-24 19:19:27 +00:00
espie
64971ac124 as discussed with pascal, make build more reproducible, by enforcing
CONFIG_SITE even if the FSF says no.

Also, finally get rid of the stage3 build/comparison, since apart from
taking build time, it's not really THAT useful unless you're specifically
debugging gcc.
2012-01-23 17:04:10 +00:00
espie
b7fe4f0430 like for 4.2, bundle a libc with adastrap.
cheat a bit: bundle the libc from 5.0, and pretend it's the one we built
with... (to be cleaned next time someone bootstraps ada)
2012-01-23 15:34:01 +00:00
pascal
1e01a1282f This patch is no longer needed, noticed by Brad. 2012-01-10 17:18:18 +00:00
robert
9d3dc83a2a missing endif 2011-12-14 18:16:51 +00:00
pascal
3afb4d8443 Fix configure test for wstring support, reported by robert@. 2011-12-13 13:46:45 +00:00
pascal
d16fa4f7fa don't pick up gmkdir 2011-12-11 21:11:37 +00:00
pascal
c20fdc65e8 don't need gsed patch anymore 2011-12-10 13:38:45 +00:00
pascal
74d633cf69 - fix fixincludes to work with our sed
- mention bootstrap comparison failure on alpha
2011-12-09 22:45:18 +00:00
pascal
d17a2a6954 ejar -> egjar 2011-12-08 19:59:36 +00:00
pascal
5c685202be missed two files in i386 plist 2011-12-08 17:42:10 +00:00
pascal
5833a9b4c7 - fix openbsd threads in boehm-gc; fixes egcj-dbtool (or, for that matter,
any gcj-compiled program)
- link boehm-gc with -pthread
- download and install ecj.jar (required by gcj)
- plist fixes
2011-12-08 17:40:55 +00:00
pascal
458d0a917c - be sure to always use sed from base
- adjust gfortran WANTLIB version in gcc4.port.mk
2011-12-06 01:51:37 +00:00
pascal
ac4ecf5e4e zap pthread from WANTLIB-go. nothing but libgo needed it. 2011-11-23 20:55:11 +00:00
pascal
5ed828f4e4 revert small mistake in previous 2011-11-22 21:47:47 +00:00
pascal
362ff059e3 - update and use OpenBSD threads in boehm-gc, from myself and kurt@. gcj-dbtool
still doesn't work.
- update gcc4.port.mk for gcc 4.6: always use libestdc++
2011-11-22 21:46:39 +00:00
pascal
2e3d081bf6 - PCH support using kurt@'s implementation does work, if you actually enable it
by not making a stupid typo.  From kurt@ (thanks).
- respect version of liblto_plugin
- add %%PIC%% in PLIST-go
2011-11-20 11:10:36 +00:00
pascal
51962ef0a2 Some plist fixes to reduce cruft from update-plist, add forgotten %%SHARED%% to
PLIST-java.
2011-11-20 10:32:30 +00:00
pascal
298b7b381a s/Free/Open/ 2011-11-19 22:49:36 +00:00
pascal
76c4d2cea0 - add Go support (only tested on amd64, but I'd be surprised if it broke
anywhere else.  The only thing not supported yet is tracing using ptrace(2).
- clean up Makefile and restructure a bit
- VMEM_WARNING when -java is built
- some minor PLIST tweaks

ok espie@
2011-11-19 22:19:53 +00:00
pascal
e75074c148 typo causing version of libgcj-tools to not be respected 2011-11-19 02:53:11 +00:00
sthen
5749b5be10 rename MODGCC{3,4}_ARCHES to MODGCC{3,4}_ARCHS, suggested by naddy,
phessler and jasper agree
2011-11-17 18:03:43 +00:00
pascal
b4395d05e9 disable on alpha. It will only build if/when it switches to gcc4, due to
missing builtins.
2011-11-15 15:48:13 +00:00
pascal
add7ae3c42 remove unnecessary patch 2011-11-15 13:38:49 +00:00
pascal
50003e9486 move down the post-install in order to be able to tell if we're a PIC arch,
so that I can fix the ownership of some libraries. Also, get rid of a trace
of m68k support.
2011-11-15 00:42:52 +00:00
pascal
b53a6cbc01 make no_f95 pseudo-flavour work 2011-11-14 22:55:10 +00:00
pascal
8e058a43c0 add powerpc support, fix java, ppc and i386 plists 2011-11-14 22:37:23 +00:00
pascal
e053729f20 new (now working) adastrap 2011-11-14 18:50:25 +00:00
pascal
e3178f7a55 bump major library versions, as discussed with espie@.
Disable on vax and m68k (I don't think it makes sense there ...), and sort archs
2011-11-14 17:34:35 +00:00