Commit Graph

57 Commits

Author SHA1 Message Date
espie
414ffa3078 Patching g77 configure.in to eradicate mkstemp looks like a good idea. 1999-11-28 00:35:24 +00:00
espie
e8ce5474b4 Fix g77 mktemp patch, now that the testsuite checks for it.
Remove extraneous warnings from the build process... for some strange
reason, the protoize patch was integrated by the Cygnus folks, but
the libiberty part is still hanging in no-man's land... go figure.
1999-11-25 22:02:30 +00:00
espie
f844940867 Argh. I was SURE I had repaired that port WEEKS ago.
So, here goes.

- retrieve a recent snapshot of egcs, to justify the port's existence,
- remove outdated fixes, now that we're synching with 2.6,
- update patches to deal with the current snapshot,
- get a description that matches better what's going on,
- add fixed include files to match the new varargs handling (complexity
is now within gcc, instead of arch-dependent include files).

Builds on i386, at least.
1999-11-24 00:15:04 +00:00
espie
917ed4d379 Update to 990907, including i386 test results,
- use contrib/egcs-update instead of rolling our own,
- autoreconf does not perform as I expected, autoconf does.
1999-09-09 14:25:11 +00:00
espie
adcfdcf547 Long-standing empty patch... 1999-09-06 13:57:36 +00:00
espie
faf101e462 Update experimental snapshot as well. 1999-08-22 16:53:37 +00:00
espie
e7e2b2f4c5 Move from egcs-stable and update to gcc-2.95.1 1999-08-22 15:50:38 +00:00
espie
378673528e Bump to 990616
PLIST not verified yet
1999-06-17 17:06:39 +00:00
espie
d69d390b8a Patch configure.in to disable texinfo instead of removing the dir. 1999-05-14 01:15:29 +00:00
espie
3cdae74b42 Necessary stuff for the use_collect2 alternative to handle shared library
ctors.
1999-05-12 15:12:03 +00:00
niklas
0e6c6b9e8f Make sure --save-temps will not clobber intermediate files from -fmain=foo 1999-05-08 20:33:35 +00:00
espie
dbb6f244c5 Update to 990428... remove patches that are now in the official version. 1999-04-29 16:54:19 +00:00
espie
061f7bc5ed Add forgotten patch 1999-04-28 15:41:53 +00:00
espie
01e1343880 Update to 19990418 snapshot
A few odds and end fixed with PLIST.template

Framework to detect the gas i386 problem

Removed bogus patch
1999-04-28 15:40:26 +00:00
espie
ca0ad472a7 ... and you need this one to prevent the same problem when you build C++.
Phew !
1999-04-23 21:33:22 +00:00
espie
22a97fceeb ... and this patch had the wrong name. 1999-04-23 13:45:22 +00:00
espie
712de16f65 Testsuite wanted installed libstdc++, because libstdc++.a link was not done
properly.

Clean a few odds & ends of main configuration.

Fix a long-standing bug that let libgcc.a get rebuilt again and again,
e.g., once for make build, once for make install
1999-04-23 13:44:32 +00:00
espie
d7aa20b123 Upgrade to 19990412 snapshot
Add a few things to simplify testing still more.

Introduce SUPDISTFILES/SUPPATCHFILES to use by a pending bsd.port.mk...
1999-04-18 20:54:38 +00:00
espie
77b9479f08 19990314 snapshot, builds okay on i386,
patch-core-weak has been integrated.
1999-03-15 02:27:27 +00:00
espie
26668457e3 - 19990307 snapshot
- consistent version number for libstdc++.
1999-03-07 14:15:33 +00:00
espie
4c161a8e6b Same work-around as mainline. 1999-02-28 18:19:54 +00:00
espie
26318df0ad Synch with 19990228 snapshot 1999-02-28 18:17:06 +00:00
espie
c6fc2f31fa Then upgrade to 980221 snapshot 1999-02-23 09:28:23 +00:00
espie
fc986ab32a Kill files that are now part of the latest snapshot 1999-02-23 09:25:52 +00:00
espie
77df149f53 Update to 990214 snapshot.
This snapshot disables recent loop changes, so the regressions from
990131 -> 990208 should no longer be there, I hope.

Removed patches have been folded back to mainstream.
This is a slight problem, as someone who wants to use an older snapshot
will have to track the CVS history of this port to find out what he
needs...

Fixed small PATH problem with autoreconf, thanks pefo.

New mips tweaks not yet in.
1999-02-14 19:33:31 +00:00
espie
bb89b14a54 Add generic xm* fragments to prevent nasty surprises, such as being
mistaken for systemV :) (and real nasties, too)
1999-02-13 00:53:48 +00:00
espie
acd112f21e Kill x-mips 1999-02-11 23:54:48 +00:00
espie
af4c16ca18 Add needed stuff to rs6000/xm-openbsd.h, configure this,
correct patch-core-mkstemp,
describe dejagnu snapshot more correctly in files/tests
1999-02-10 00:20:41 +00:00
espie
14117cab73 Use mt fragment ONLY when threads are defined, this avoids spurious
IO_flockfile and fixes a regression...

Complete mt_fragment with posix define...
1999-02-09 10:01:15 +00:00
espie
113d425e1c Revert previous change since egcs guys fucked up and didn't use their
cvs tree properly, sigh...
(took me 3 tries to get this right)
1999-02-08 22:47:53 +00:00
espie
2985d84f98 Update to 990208 snapshot: mainly kill all patches that I've got into egcs,
synch configure.in line numbers
and add a symbolic MASK_SNAKE I had forgotten for hppa.
1999-02-08 20:34:26 +00:00
espie
e8e4c15ae3 Preliminary support for m88k. 1999-02-08 17:51:03 +00:00
espie
8212c948d1 Kill bogus rs6000 generic xmakefile. 1999-02-07 12:37:01 +00:00
espie
e4101d6079 Recognition for native openbsd hppa. 1999-02-07 12:34:44 +00:00
espie
feefda1505 ... use those symbols 1999-02-06 21:38:47 +00:00
espie
8d3240fe35 Get symbolic constants from cvs-current, in order to prepare openbsd
configuration for inclusion into egcs mainstream.
1999-02-06 21:34:44 +00:00
espie
7a613f1639 Thread support, approximation... 1999-02-03 18:02:53 +00:00
espie
033355874a Remove ONLY_FOR_ARCHS, makes little sense now...
rs6000 is special, as it incorporates everything into one file...
1999-02-02 01:19:36 +00:00
espie
f02d753a4e Clean up mkstemp patch mess. 1999-01-31 23:19:24 +00:00
espie
6270519121 Stupid mistake... this is the real Mc Coy,
Thanks niklas !
1999-01-25 14:46:10 +00:00
espie
e6bd2e83d8 Update to 19990124 snapshot.
Adjust PLIST.template as it was missing some directories.
Rebuild configure.in patch to only add collect2 when needed, plus typo fix.
1999-01-24 21:36:11 +00:00
espie
ec4539d95a Upgrade to 19990117. Fix a few details for integrating pa & alpha.
First shot at mips.
1999-01-18 10:23:25 +00:00
espie
1770daac5a - fixed up i386/*VTABLE once more... this time for good.
- clean up openbsd.h comments
- add patch to auto-detect mkstemps. Accordingly, need to run
autoreconf in gcc and in libiberty.
1999-01-12 11:27:39 +00:00
espie
204aea09bf Oops 1999-01-12 00:41:09 +00:00
espie
7d42d0d4e1 partial archives are no longer bogus.
i386/openbsd.h does include unix.h which yields a perfectly good definition
of the *THUNK macro.

trivial fix to a bug in gcc.c.
1999-01-12 00:19:49 +00:00
espie
1237398b0a Some cleaning up, triggered by alpha.
- TARGET_CPU_DEFAULT belongs in configure.in, not in various .h files.
If it goes out wrong, it's the output of uname that needs to be fixed
(or config.guess, e.g., even pentium gives out i386, as intended). Makes
it possible to configure a i586-unknown-openbsd system if one really WANTS
to output only pentium specific code.

- ASM_FILE_START put into alpha/openbsd.h. Used to be in alpha.h, but
elf & osf diverged...

- set xfrag=none for alpha, as the default config is bogus and supplies
a dec-osf specific xalpha.
1999-01-11 21:23:25 +00:00
espie
e4628fe129 Update to 990103 snapshot.
Inform user of the warnings file in case of a bootstrap crash.

Major change: add alpha & pa configuration, and clean up ALL
openbsd configuration files thoroughly... this is probably going to
break a few things.
1999-01-10 02:50:02 +00:00
espie
5f1c9b66a3 and accordingly, this is no longer necessary... 1999-01-09 16:56:19 +00:00
espie
646bb8a000 Update to 19981213 snapshot.
- use new `enable-languages' mechanism,
- depend i386/openbsd.h on flag_pic
- remove temp fix for m68k
1998-12-16 17:52:29 +00:00
espie
0e4277a594 Forgot to clean up those for 19981130 1998-12-05 16:15:53 +00:00