109 Commits

Author SHA1 Message Date
naddy
7652ff02cd bump for plist changes 5.5 -> 5.6 2014-07-16 14:35:38 +00:00
tobiasu
3e0322844a Add gnat for alpha. Because we can.
ok pascal@
2014-05-01 23:08:26 +00:00
tobiasu
c5c900db71 Provide hppa adastrap for GNAT. Hehe... 2014-04-27 21:50:55 +00:00
pascal
484f25ad69 Fix after gets() removal (note that this is only built on alpha and hppa).
ok landry@
2014-04-23 11:34:49 +00:00
espie
86a0a064dc document what's actually going on 2014-03-10 21:33:17 +00:00
landry
63af4a755e Only build 4.6 on alpha/hppa. Lots of cpu cycles saved on other archs! 2014-03-10 21:11:49 +00:00
tobiasu
fa5f89b880 Repair time/delay/duration time_t issues in gnat. All direct imports of
C functions have been modified to use same-size types on both sides.
There are no checks in gcc to make sure the types match, and it happily
trashes the stack or reads bogus data without this fix.

The C shim in gnatlib is still casting down to long (by design), thus
gnat is not 2038-proof on 32 bit architectures. Fixing this is an
ongoing project, but I'm running out of time for this release.

Getting the timezone offset is now supported; the unused usleep
procedure that snuck in from FreeBSD support got removed.

Tested on all supported architectures, plus new bootstraps due to the
stack trashing, just in case...

Conceptually fine with brad@
deraadt@ is unhappy with the partial solution, fixing this remains a
high priority todo
2014-02-07 01:27:21 +00:00
brad
adb9b36d26 Enable PIE by default on alpha to match the base compiler.
ok pascal@ sthen@
2014-01-15 20:16:03 +00:00
pascal
001e0489ab Switch ports gcc to PIE on i386.
ok sthen@
2014-01-12 22:16:02 +00:00
naddy
09999a5515 bump for plist changes 5.4 -> 5.5 2014-01-12 14:58:30 +00:00
zhuk
0999226a47 Make ccache work with Clang and GCC 4.6 & 4.8.
tested and okay juanfra@; also okay sthen@ for Clang part
2013-12-01 20:05:41 +00:00
zhuk
0da83f0322 Force NO_CCACHE for non-base compilers. Before that, ports system tried to
enable it through symlinks in ${WRKDIR}/bin/, which were effectively
overriden by gcc4.port.mk or clang.port.mk - thus lying that ccache was
used while it was not.

Still investigating for a good solution to actually enable ccache support
for non-base compilers.

okay juanfra@, sthen@
2013-11-27 20:42:08 +00:00
sthen
701c91e413 kill VMEM_WARNING 2013-11-25 14:16:22 +00:00
brad
19bc641c6e As was commited to the base GCC have the compiler use collect2 when
linking instead of calling the linker directly.

ok pascal@
2013-11-22 19:58:40 +00:00
tobiasu
a6e0c5595f New sparc64 adastrap 2013-09-19 21:14:56 +00:00
sthen
1ec0ac3ed5 new i386 bootstraps 2013-09-16 21:16:17 +00:00
espie
0862814d0b new adastrap for amd64 WITHOUT the fixed-includes
(and moved an rm out of a loop)
2013-09-16 10:55:00 +00:00
espie
e9a158ab37 fix gcc build, make sure it doesn't rely on bootstrapped old
fixed includes
2013-09-16 10:28:15 +00:00
tobiasu
eb47b19fef New adastraps for sparc64. 2013-08-16 07:37:18 +00:00
sthen
ebf9dd03db new bootstraps for amd64/i386, on a temporary mirror for now. 2013-08-14 07:56:25 +00:00
tobiasu
18f2e1d247 * Use SJLJ exections instead of ZCX in the hope of fixing
intermittent build problems. Exception style affects Ada only.
 * Add better tasking support and sync OpenBSD specific changes from 4.8.
 * Call __errno() from librthread when using pthreads (Tero Koskinen)
 * Enable shared library support for OpenBSD (Tero Koskinen)

bulk tested by landry@ and sthen@
ok pascal@

sorry for the delay guys
2013-07-18 16:54:08 +00:00
naddy
6b7bf0fa1b bump for plist changes 5.3 -> 5.4 2013-07-07 21:25:27 +00:00
sthen
b6cab70bad regen PLIST to unbreak following kerberosV update; com_err.h no longer
needs to be "fixed."  ok pascal@
2013-06-18 20:24:29 +00:00
landry
42351afa43 Fix gcc's on sparc64 by using a regen'ed ada boostrap. From tobiasu@ 2013-06-13 06:44:08 +00:00
pascal
ca64a01d65 Update to GCC 4.6.4. tested by and adastrap-sparc64 built by tobiasu@ 2013-06-07 07:54:23 +00:00
pascal
18196e1d85 Don't pass -g in CFLAGS; it increases the size of ${WRKBUILD} to over 4G and
thus prevents building it in tmpfs under this threshold.

prompted by/ok espie@
2013-05-18 19:45:11 +00:00
tobiasu
12a99dc798 Take another stab at debugging the sporadic build failures in the gnat
build tools. Split compiling and linking/binding into two parts and add
some ls in between to spot the disappearing object file...

Reverts previous attempt running gnatmake with -j1 which turned out to
be insufficient.

ok pascal@ sthen@
2013-05-03 23:23:26 +00:00
tobiasu
277833d422 Sync 4.8 changes to 4.7 and 4.6
- Absolute path for STRIP
- Enable ncpu detection in gnatlib
- Force sequential "make install" target
2013-05-03 23:13:49 +00:00
pascal
d4d059a765 Attempt to fix longstanding ada build failure when building in parallel by
forcing gnatmake to use -j1.

ok sthen@
2013-04-11 08:34:13 +00:00
ajacoutot
58f1a6f9f6 USE_LIBTOOL=Yes is the default now. 2013-03-21 08:45:11 +00:00
espie
0662a4e9d6 PERMIT_* / REGRESS->TEST sweep 2013-03-11 11:20:26 +00:00
pascal
c77b1f0cae sync hppa LINK_SPEC; ok brad@
no bumps for 4.7 as it doesn't build there
2013-02-04 11:37:09 +00:00
naddy
cbeb31112b bump for plist changes 5.2 -> 5.3 2013-02-01 16:34:36 +00:00
pascal
5d1e11d3f5 Don't hide error messages from gnatmake to give us a chance to debug this when
it fails.

ok tobiasu@
2013-01-28 11:42:33 +00:00
tobiasu
7bf586c62e Remove sparc, m68k, and vax from the list of working architectures.
Their respective gcc backend ports are unfinished and broken.

ok pascal@
2013-01-28 01:03:05 +00:00
pascal
3f1c4aa554 rm TODO, now obsolete. 2013-01-23 18:28:03 +00:00
pascal
f5b212fcc5 Add alpha support (without -go). Takes 14h to build.
tested by and ok naddy@
2013-01-23 18:26:55 +00:00
tobiasu
5610f716f0 "Fix" parallel make install by forcing gmake to only run 1 job. It's not
pretty and we would like to have a better fix, but for now it should suffice.

ok naddy@, sthen@
2013-01-20 02:23:40 +00:00
bcallah
6e2fc64bd9 More mips64el check-in. Add mips64el to NOT_FOR_ARCHS-go. ld fails: Bad final value
ok jasper@ pascal@
2013-01-18 19:09:58 +00:00
pascal
b368cf7d60 Add hppa support. Still a bit rough around the edges: Some optimisations
(-foptimize-sibling-calls and -freorder-blocks) cause segfaults during build,
so don't include them in -O2 on hppa.  Also, -go is not supported.

requested by and ok brad@ and tested on one of his machines
2013-01-16 19:28:48 +00:00
pascal
05fd655191 Bring in some WIP mips64 work; fails atm due to a binutils bug when linking
liblto_plugin.so:

/usr/bin/ld: not enough GOT space for local GOT entries
/usr/bin/ld: BFD 2.15 internal error, aborting at
/usr/src/gnu/usr.bin/binutils/bfd/elfxx-mips.c line 6483 in _bfd_mips_elf_relocate_section

with/tested by jasper@
2013-01-10 15:00:55 +00:00
espie
298824094f a little overkill on autoconf override magic... 2012-12-28 14:32:00 +00:00
naddy
814177890f do not run parallel fake until the race condition there has been fixed 2012-12-26 16:15:19 +00:00
tobiasu
18e9b90a33 Add Ada support on amd64 and sparc64
Also unbreak and improve adastrap generation
Cleanup bootstrap libc handling

ok pascal@
2012-11-06 00:57:43 +00:00
kurt
e6d0f75aab Fix boehm-gc and increase gcj default thread stack size. Makes gjc work
well enough to bootstrap devel/jdk/1.6. okay pascal@ jasper@
2012-10-17 19:58:21 +00:00
landry
5b823cb084 Sync the config/openbsd.h and config/sparc/openbsd64.h patches with
respectively r1.9 and r1.6 of base gcc ones adding #define
NO_IMPLICIT_EXTERN_C. While here add the missing #define
NO_UNSAFE_BUILTINS to gfortran/gcc 4.2, Bump all REVISIONs.
ok sthen@
2012-10-07 20:09:44 +00:00
pascal
a2552ed7a7 Unify MASTER_SITES for adastrap; they're now all hosted by Tobias Ulmer on
tmux.org.  Makes gnat bootstrapping work a little easier for him.
2012-09-28 15:02:28 +00:00
naddy
b59373b722 Add DPB_PROPERTIES=parallel for some big ports.
Required to make full use of the amd64 package building cluster.
2012-09-26 17:36:46 +00:00
pascal
f080554be1 Make gcc 4.[67] use __guard_local instead of __guard. From Brad.
ok matthew@
2012-09-03 13:24:58 +00:00
pascal
13f6a15732 Patch GCC and derived ports to use PIE by default on the same architectures it
is enabled on in base; except for gcc 3.3, which does not have PIE support and
therefore needs to pass -nopie to the linker.

Also, sync powerpc pthread specs with base.  From Brad.

ok naddy@
2012-09-01 00:00:32 +00:00