pierre
fec0db7ba0
* some new bugs
2000-05-02 15:11:00 +00:00
pierre
a41bd45aad
* 0**n where n<>0 is 0; 0**0 generates RTE 207
2000-05-02 10:37:50 +00:00
sg
19bd9af1af
* Fixed a small typo (GI_BM_TYPE_LOGIC instead of GGI_BM_TYPE_LOGIC)
2000-05-02 10:27:39 +00:00
pierre
c051e40994
* fix a small problem when deciding to call DosExecute directly
2000-05-02 10:20:40 +00:00
pierre
2126a0993e
* new set of Gabor changes: see fixes.txt
2000-05-02 08:42:26 +00:00
Jonas Maebe
d82a0e5958
* changed NOT to A_NOP
2000-05-01 11:04:49 +00:00
Jonas Maebe
44845ec919
* some fixes, does not yet compile
2000-05-01 11:03:32 +00:00
Jonas Maebe
6e996c65e5
+ testcse2
2000-04-29 16:58:43 +00:00
Jonas Maebe
6e83066171
* fixed incompatibility with range chcking code, -O2 and higher
...
now work correctly together with -Cr
2000-04-29 16:57:44 +00:00
Jonas Maebe
2eb73a055e
* destroyreg overwrote some memory if the reg was an FPU register
2000-04-29 16:56:45 +00:00
peter
f40500579d
* fixed offset support intel reader, the gotoffset variable was not
...
always reset
* moved check for local/para to be only used for varsym
2000-04-29 12:51:33 +00:00
peter
ad5c964217
* support asm block without assembler directive for -Sd
2000-04-29 12:50:14 +00:00
peter
10c46e7f4d
* rebuild
2000-04-29 12:49:46 +00:00
peter
d371c67332
* fixed long line for tp7
2000-04-29 12:49:30 +00:00
Jonas Maebe
1f8972290d
* fixed arctan2 (tbug788 now works correctly)
2000-04-29 10:10:51 +00:00
Jonas Maebe
3e11694e49
* nmem compiles again (at least for powerpc)
2000-04-29 09:01:06 +00:00
pierre
c4f4ac6ae2
* fix my last fix for other targets then win32
2000-04-28 08:53:47 +00:00
pierre
695f24d2bd
new
2000-04-27 23:08:59 +00:00
pierre
e412711fa4
* expected failure
2000-04-27 23:08:33 +00:00
pierre
35f1a5def1
* expectedfailures
2000-04-27 23:08:09 +00:00
pierre
a3434cc825
* fix for bug report 929
2000-04-27 15:35:50 +00:00
pierre
71cbe13810
* power to ** operator fixed
2000-04-27 11:35:03 +00:00
pierre
22a6677394
* fix for snapshot failue
...
* order_overloaded reintrocduced and adapted to operators
2000-04-27 10:06:04 +00:00
pierre
e4463fc278
* dpmiexcp removed
2000-04-26 10:21:05 +00:00
pierre
477eacdff2
* correct bugs for ts010026 and ts010029 in win32 mode
...
in copyvaluparas
+ use SHL instead of IMUL if constant is a power of 2 in copyvalueparas
2000-04-26 10:03:45 +00:00
pierre
3174b39bcc
* More changes for operator bug
...
Order_overloaded method removed because it conflicted with
new implementation where the defs are ordered
according to the unit loading order !
2000-04-26 08:54:18 +00:00
pierre
25d48db77a
* test for operators in different units improoved
2000-04-26 08:50:23 +00:00
pierre
9560f264a3
+ sym_n_unit_not_used
2000-04-26 07:32:02 +00:00
pierre
dbdb7ee667
+ Hint about unused unit
...
* Testop bug fixed !!
Now the operators are only applied if the unit is explicitly loaded
2000-04-25 23:55:29 +00:00
pierre
4ab63fea0a
* avoid SIGSEGV
2000-04-25 21:41:35 +00:00
Jonas Maebe
e96d0a3d6e
- disabled "string_var := string_var + ... " and "string_var + char_var"
...
optimizations (were only active with -dnewoptimizations) because of
several internal issues
2000-04-25 14:43:36 +00:00
pierre
14fbf4bf69
* renamed lowercased
2000-04-25 10:59:02 +00:00
pierre
bf29b96e0e
* renamed lowercased
2000-04-25 10:28:52 +00:00
pierre
bbe65b02c0
* New Gabor changes : see fixes.txt
2000-04-25 08:42:32 +00:00
peter
8818701433
* added enhancedraise define
2000-04-24 13:34:29 +00:00
peter
47ffdafd5e
* added some more fpuint instructions
2000-04-24 13:33:35 +00:00
peter
a9cff3c8c5
* removed unused vars
2000-04-24 12:48:37 +00:00
peter
015cca99cd
* fixed check for exporting var or proc
2000-04-24 12:47:27 +00:00
peter
68e75531e5
* defines added
2000-04-24 12:46:34 +00:00
peter
013cfb9bc4
* made overloaded_operators local per unit, but it still doesn't work
...
correct
2000-04-24 12:45:44 +00:00
peter
ebbf2e578f
* backtraces for exceptions are now only generated from the place of the
...
exception
* frame is also pushed for exceptions
* raise statement enhanced with [,<frame>]
2000-04-24 11:11:50 +00:00
peter
70d3d42a5c
* removed eoln
2000-04-24 11:09:05 +00:00
Jonas Maebe
613ceaa8fb
* only enable string_concat optimization with uncertain optimizations,
...
because it gives wrong results with "s := s + f()" where s is a
string and f() is a call to a function that modifies s
2000-04-23 21:04:09 +00:00
Jonas Maebe
a48a8031b8
* changed "mov reg1, reg2; mov reg2, y" optimization that caused
...
regalloc info to become invalid (it's still performed, but the
regalloc info stays valid now)
2000-04-23 14:56:36 +00:00
Jonas Maebe
43e023d0af
* use FPC_SHPRTSTR_CONCAT_LEN for -dnewoptimizations (temp)
...
* more precise reg deallocation when calling the above)
2000-04-23 09:28:19 +00:00
Jonas Maebe
ff9782d70c
+ FPC_SHORTSTR_CONCAT_LEN (temporary, for -dnewoptimizations)
2000-04-23 09:26:51 +00:00
marco
1ea5a1824e
* Forgot to remove the tpoint stuff.
2000-04-22 19:51:08 +00:00
marco
796dd36a1f
* some redef fixes (for lazarus mwedit component)
2000-04-22 17:46:05 +00:00
Jonas Maebe
941f20c29a
* cleaner register (de)allocation in secondfor (for optimizer)
2000-04-22 15:29:26 +00:00
Jonas Maebe
57e41f9053
* aasm.pas: pai_align instead of pai_align_abstract if cpu <> i386
...
+ systems.pas: info for macos/ppc
* new/cgobj.pas: compiles again without newst define
* new/powerpc/cgcpu: generate different entry/exit code depending on
whether target_os is MacOs or Linux
2000-04-22 14:25:03 +00:00
