Commit Graph

1002 Commits

Author SHA1 Message Date
jtc e4cdcb8d79 soft-float accidently turned off the 68060 bit 1996-05-17 01:56:13 +00:00
phil a964fa9b57 Back out the change to include libio and libiostream. 1996-05-16 21:03:24 +00:00
oki c9d31dfe60 Oops, ${MACHINE} == "x68k" , not ${MACHINE_ARCH} == "x68k" 1996-05-16 15:55:38 +00:00
pk 46d44c05b0 Document environment variables (PR#2415).
Also correct spelling and a couple of format macros.
1996-05-15 23:46:00 +00:00
phil 0534b81c29 Install includes from libio and enable building and installation of libio and
libiostream.  Some functions are duplicated in libio, libiostream and
libstdc++.  If needed, we can remove duplicated functions later.
(Fixes PR 2390)
1996-05-15 22:56:13 +00:00
oki 03ca21bc0d Add x68k to the list of ports using gcc instead of gcc2. 1996-05-13 14:53:21 +00:00
jtc 3c02c6068b cheezy hack that adds -m68060 as a synonym for -m68040 1996-05-12 18:55:53 +00:00
jtc a7dc4ab369 Pass -m68060 through to the assembler 1996-05-12 18:52:22 +00:00
mycroft 16255ae0be Change the segment mov constraint to WordReg|WordMem, to disallow 8-bit memory
access, and enable a (somewhat) common register usage.
1996-05-11 13:39:26 +00:00
mycroft f0c3c199fa Update for removal of cexp.c. 1996-05-11 11:37:59 +00:00
mycroft 152a87dd9d This is generated automatically. 1996-05-11 11:30:58 +00:00
pk 8a3cf4d3fe Align BSS sections on machine dependent boundaries. The compiler/assembler
seem to not always do this for us.
1996-05-09 23:51:21 +00:00
jtc 2678c87a47 Add port-x68k category 1996-05-08 23:29:54 +00:00
pk fa57b905a9 kcorelow.c: use specified executable name (PR#2375). 1996-05-05 19:58:01 +00:00
pk f685c6dd5d Extend memory class of the `mov' instruction (PR#2370). 1996-05-03 09:33:49 +00:00
jtc 7d85884ca8 Added support for -R option.
This patch has been accepted by the FSF, and should appear in future
gcc releases.
1996-04-25 18:23:37 +00:00
jtc c63d7d5d2e Use 'includes' instead of 'beforeinstall' target to install headersUpdate libg++2netbsd to match 1996-04-25 01:16:07 +00:00
jtc 7bf481ecb2 Integrate change from the gcc development sources:
Tue Feb 13 17:59:03 1996  Lee Iverson  <leei@Canada.AI.SRI.COM>

	  * gcc.c (DEFAULT_SWITCH_TAKES_ARG): New macro, from SWITCH_TAKES_ARG.
	  (SWITCH_TAKES_ARG): Use it.

This change alows us to add support for the 'R' option in a way that
will minimize differences when it comes to merging a future FSF gcc
release into the NetBSD sources.
1996-04-24 02:35:29 +00:00
pk 390bcef256 md_parse_arg: `-K' implies `-k'. 1996-04-18 19:03:02 +00:00
pk be19763acb Accept `-K' as equivalent to `-k' (except the ns32k which already handles -K). 1996-04-14 11:31:11 +00:00
phil 3f09e4a7e4 Sync with gcc2netbsd. (and do the right thing.) 1996-04-11 21:22:42 +00:00
phil d012169278 Fix problems with which and where includes are installed. 1996-04-08 02:45:55 +00:00
jtc 3c36209f8c Re-enable weak alias and sysv pragma support that was lost when we
upgraded to gcc-2.7.  These changes have been submitted to the gcc
maintainers.  With luck, we won't have this problem again...
1996-04-06 00:51:08 +00:00
phil 912d727afa Fix for hangs on the form 0fInf. (Fix from PR 2279) 1996-04-05 16:21:14 +00:00
phil 208dd8b0cc Adding librx to libg++ tree. (Closes PR 2271 and 2298.) 1996-04-04 23:33:45 +00:00
christos bd9f45b014 Add -D__KPRINTF_ATTRIBUTE__ 1996-04-04 20:17:16 +00:00
chuck 3c87aee01a Switch mvme68k to `gcc'. 1996-04-03 21:20:52 +00:00
phil ab818facfa Sync with change to Makefile.cc1. 1996-04-03 08:18:16 +00:00
phil 5e11f509b9 Add a dependency that will cause a relink for anything that uses libcc1.a. 1996-04-03 08:15:38 +00:00
pk 5a2722dd1e Add missing argument to error() call. 1996-04-01 22:41:16 +00:00
pk 7426e189fb Recognise ".double 0fInf" as suggested by Matthias Pfaller (PR#2280). 1996-04-01 10:57:35 +00:00
christos a323f3fc97 - Add a kprintf format specifier, so that it includes only the formats valid
for kernel printf functions
- understand the db_printf %n,%r,%z format specifiers
- understand the kernel printf %: format specifier
- Be more permissive to %b arguments: accept any integer type, not only
  unsigned ints.
1996-03-30 04:35:04 +00:00
cgd d4ea3a96dc accept any tar header which begins with "ustar" (i.e. those 5 characters) as
a ustar-format archive.  POSIX specifies that ustar-format archives have magic
strings consisting of "ustar<nul>", however our PAX accepts anything starting
with just the 5 "ustar" chars, and this and other version of GNU tar output
ustar archives with magic numbers of "ustar  <nul>".  Leave the output format
of this version of GNU tar the same, so that old versions of GNU tar can
use ustar files it outputs.  (diff sent to GNU tar maintainer.)
1996-03-26 00:48:49 +00:00
jtc 04b0ab5344 Be pedantic, main() returns int in Standard C. 1996-03-21 18:30:19 +00:00
pk 27a9f993bb Bleh, got %b arguments reversed in previous change. 1996-03-17 22:27:17 +00:00
pk 1e66a8828f Take a stab at recognising %b formats. 1996-03-17 03:13:17 +00:00
jtc efdbc16b61 Patch from J"orn Rennecke that fixes code generation bugs with long
long comparisons.  As of this date, this change hasn't made it into
the development sources.  We must consider this when it comes time
to integrate a newer gcc release.

Thu Mar  7 01:16:23 1996  J"orn Rennecke (amylaar@meolyon.hanse.de)

	* expmed.c (negate_rtx): Don't negate LONG_MIN if mode is wider
	than HOST_WIDE_INT.


gcc now produces worse code for this test case than gcc 2.5.8 when
invoked with -O , but it will optimize as good as gcc 2.5.8 (i.e. all
comparisons vanish) when invoked with -O2 , thus I think it doesn't
matter.  If anyone is interested in looking into this, the code in
expr.c, function expand_expr, case MINUS_EXPR, if-clause for if
(TREE_UNSIGNED (type) || TREE_OVERFLOW (negated)), will refuse to
convert the subtraction into an addition if there is an overflow in
the conversion or negation of the constant.  If both host and target
machine are binary computers with 2-complement representation, the
overflow should not matter.
1996-03-15 22:55:51 +00:00
veego b9f085f78f Restore a local change from our old libg++ to fix a compile warning 1996-03-13 23:07:35 +00:00
is 3725e6e117 Fix a typo and mention -P/--absolute-paths in the man page.
From Greg Earle, PR 1585.
1996-03-13 14:55:00 +00:00
phil 6867e1fda1 Sync with previous changes to Makefiles ... 1996-03-12 21:28:52 +00:00
phil d99c63563c Uncomment the %expect. 1996-03-12 19:01:05 +00:00
phil 2fe7ea6745 Uncomment the %expect. 1996-03-12 18:46:21 +00:00
phil 4de8058449 Added / in cmp path for /usr/include/g++/stl. (To match libg++2netbsd and
be correct.)
1996-03-11 17:22:42 +00:00
phil ebca468319 Update for -DGCC_NAME=... in Makefile for g++. 1996-03-11 06:53:27 +00:00
phil cf6c269b55 Change so g++ doesn't concat junk to the front of /usr/bin/cc for
the name of gcc.  (Closes pr 2188, different fix)
1996-03-11 06:48:33 +00:00
phil 180e9e53f6 Back out last change now that libstdc++ now exists. 1996-03-11 06:35:33 +00:00
thorpej e311fddd2a Fix installation of header files. 1996-03-11 04:58:22 +00:00
thorpej a039d62135 s/ALLSRC/IMPSRC/ so that we don't try to compile header files. From
Gary D. Duzan <gary@wheel.tiac.net>.
1996-03-11 02:38:44 +00:00
phil 14b81cb599 Removing old libg++. 1996-03-09 00:36:34 +00:00
phil 84a13b4cb6 Removal of old libg++. 1996-03-09 00:28:43 +00:00