sater
|
7e31dee840
|
changed EM_WSIZE to TEM_WSIZE
|
1984-10-23 15:10:15 +00:00 |
|
sater
|
a4a5fc4bfa
|
changed references to EM_?SIZE to TEM_?SIZE
|
1984-10-23 15:04:24 +00:00 |
|
sater
|
62f42d18f7
|
changed EM_WSIZE to TEM_WSIZE
|
1984-10-23 15:02:03 +00:00 |
|
sater
|
79ea4bdea3
|
Changed to use -h flag to cgg and better way of preserving old tables.h
if not changed. Nice huh, Andy?
|
1984-10-23 13:29:41 +00:00 |
|
sater
|
ab0db1c5fb
|
added rcsid(again). Who took that away?
|
1984-10-16 13:45:16 +00:00 |
|
sater
|
5d5a09a5d0
|
added code to generate code for ACK assembler.
Works if ACK_ASS is defined.
|
1984-10-16 13:31:44 +00:00 |
|
em
|
94e069fde1
|
Initial revision
|
1984-10-09 10:36:29 +00:00 |
|
keie
|
6cbb37051b
|
Initial revision
|
1984-10-08 14:11:42 +00:00 |
|
keie
|
550d1fa79e
|
New installation procedure:
1 - New place for binaries and libraries in the ACK tree.
2 - The ranlib action is moved to ../../install which tries it at
every action.
|
1984-10-08 14:09:28 +00:00 |
|
bal
|
46dc90f366
|
bug fixed: generate "add.l #65536,a1" rather than "lea 65536(a1),a1".
The latter form is only generated for constants that fit in a word.
|
1984-10-08 12:45:38 +00:00 |
|
sater
|
13d6e6815e
|
added missing cast at call to regsave, in ms_gto case
|
1984-10-02 11:17:57 +00:00 |
|
sater
|
92aecb175c
|
changed definition of myalloc to extern
|
1984-10-02 10:31:06 +00:00 |
|
sater
|
a3f62d0255
|
added external definition for myalloc()
|
1984-10-02 10:30:44 +00:00 |
|
bal
|
ebed9d4804
|
unlk a6 added to restr().
|
1984-09-11 15:44:14 +00:00 |
|
keie
|
4c265a7cc6
|
Initial revision
|
1984-09-03 13:49:40 +00:00 |
|
sater
|
2bf16c5b2f
|
Variable nregvars had a name clash with the array nregvar[] in the
file tables.c. That was different in the 8'th character (!"#$%&'()_).
These variables should actually be static.
|
1984-08-23 17:07:53 +00:00 |
|
sater
|
5689e6ab64
|
Increased the amount of information passed through error routines
to printf. The routine badassert, running on a 2/4 machine lost it's
last argument on the way.
Anyone know of a better way to solve this problem?
|
1984-08-17 15:35:28 +00:00 |
|
sater
|
ce098c78cb
|
added patterns to prevent stacking in *p++ cases
|
1984-08-15 16:05:54 +00:00 |
|
sater
|
e682a47599
|
added patterns for inc/dec/adp of local and external offsetted
This makes the getc() macro from C very efficient, among others
|
1984-08-15 14:31:54 +00:00 |
|
sater
|
e41876f32a
|
reworked patterns for autoincrement stores, now that peephole optimizer
rearranged it.
|
1984-08-07 10:22:49 +00:00 |
|
sater
|
6f8e6faa1f
|
added setcc calls to some incrementing ops
added patterns for indirect local and external increments
added patterns for loe ine and such
|
1984-08-03 16:33:40 +00:00 |
|
sater
|
ff9d23e4ec
|
A move from a nonreg to a reg did not have the effect that all
registers clashing with the destination were erased. Erase added.
|
1984-08-02 16:31:25 +00:00 |
|
bal
|
6e519cf02b
|
bug fixed: added "fprintf(codefile,""rts\n");" to restr()
|
1984-07-30 12:47:54 +00:00 |
|
bal
|
f2e52892ac
|
bugs fixed (syntax errors, patterns moved to front,
"ldl ldl adp sdl sti" requires DATAREG instead of ANY1 on
fakestack)
|
1984-07-30 11:16:41 +00:00 |
|
bal
|
233b4aeee4
|
patterns added to generate autoincrement and to generate
a "move X,d0" instead of a "move X,-(sp)" before storing into X
|
1984-07-27 16:53:19 +00:00 |
|
sater
|
0b07016f0b
|
added PR* routines
|
1984-07-26 13:15:21 +00:00 |
|
sater
|
5894adfd67
|
Changed procedure prolog to use PR* routines.
|
1984-07-26 13:05:22 +00:00 |
|
keie
|
14a3683f85
|
*** empty log message ***
|
1984-07-23 11:49:00 +00:00 |
|
sater
|
eebc753306
|
removed all calls to blm~ from table
duplicated some patterns for in[lc] to deal with de[lc]
extra rule for adi 2 to deal with source1 cleverly
|
1984-07-23 11:29:59 +00:00 |
|
sater
|
978dd0191a
|
removed blm.s, not used by table anymore
|
1984-07-20 16:20:54 +00:00 |
|
sater
|
5b20e4fed2
|
edited to make reentrant, no more globals in use
|
1984-07-20 16:20:28 +00:00 |
|
sater
|
eb823929a1
|
A sti 8 of a regdef8 token went into panic mode. Fixed.
|
1984-07-20 10:32:48 +00:00 |
|
sater
|
ef248ee65e
|
to check long >= 0 or long < 0 it is enough to test high order word. Added.
anding and oring of long constants is much better now
dvu 2 and rmu 2 by positive constants is now done inline.
|
1984-07-19 16:50:27 +00:00 |
|
sater
|
e20ab23a4b
|
added rcsid
|
1984-07-19 15:03:30 +00:00 |
|
sater
|
5fb767966b
|
Initial revision
|
1984-07-19 15:01:43 +00:00 |
|
sater
|
15e9fd8cd5
|
removed #ifdef UNTESTED around logical instructions without arguments,
seems they were used ( which implicitly means tested ) after all
|
1984-07-19 14:03:49 +00:00 |
|
sater
|
73d8f05fb5
|
Initial revision
|
1984-07-19 13:52:18 +00:00 |
|
sater
|
3ca5d2fcb6
|
added rcsid
|
1984-07-19 13:34:12 +00:00 |
|
sater
|
b95f4c8bfd
|
Initial revision
|
1984-07-19 13:20:15 +00:00 |
|
sater
|
801226a45c
|
old code put back, now register handling is different again
actually old code never was removed, due to a misunderstanding
about comment conventions.
|
1984-07-19 13:12:03 +00:00 |
|
sater
|
f16fdf4e7f
|
makefile adapted to new style archive handling, RCS and all that
|
1984-07-19 13:11:36 +00:00 |
|
sater
|
d082a2b0a2
|
a setl instruction was changed to a setd
setl was just plain wrong
|
1984-07-19 13:09:53 +00:00 |
|
sater
|
d5aa088876
|
Initial revision
|
1984-07-19 12:51:00 +00:00 |
|
bal
|
8defe7ce53
|
rcsid added
|
1984-07-19 11:50:29 +00:00 |
|
bal
|
69584eb0ca
|
Initial revision
|
1984-07-19 11:50:28 +00:00 |
|
bal
|
da4092cbdf
|
New implementation of GTO (without register save masks)
several optimizations added
|
1984-07-19 11:15:28 +00:00 |
|
bal
|
c1fd560106
|
New implementation of GTO (without register save masks)
|
1984-07-19 11:07:15 +00:00 |
|
bal
|
17caf7129e
|
rcsid added
several optimizations added
|
1984-07-19 11:02:06 +00:00 |
|
bal
|
a520ccdbbc
|
rcsid added
clr.w -(sp) generated in prolog if register save mask is 0
|
1984-07-19 10:57:19 +00:00 |
|
bal
|
eeeff10a51
|
Initial revision
|
1984-07-19 10:03:10 +00:00 |
|