Commit Graph

521 Commits

Author SHA1 Message Date
yury
d306e25dc8 * Fixed warnings.
git-svn-id: trunk@9514 -
2007-12-22 12:27:03 +00:00
florian
d08abbca46 * forgotten commit
git-svn-id: trunk@9334 -
2007-11-25 20:49:57 +00:00
yury
c4ed91b9ff * Fixed warnings about EBP based access.
* Improved SetJmp and longJmp by not using stack frame.

git-svn-id: trunk@9112 -
2007-11-03 18:05:12 +00:00
yury
7230661978 * Fixed warnings about EBP based access.
git-svn-id: trunk@9111 -
2007-11-03 17:09:39 +00:00
yury
4662eb98fc * Fixed warnings.
git-svn-id: trunk@9102 -
2007-11-03 12:22:54 +00:00
florian
b2b0e749bb * win32 system unit doesn't depend anymore directly on the main program, this is necessary to be able to keep it in a dll
git-svn-id: trunk@9052 -
2007-11-01 21:59:43 +00:00
florian
7da7364ee7 * refactored SysResetFPU into SysInitFPU and SysResetFPU
git-svn-id: trunk@8966 -
2007-10-28 12:06:49 +00:00
florian
5a3f71754a - removed odd assembler implementation, using the inline variant usually creates better code
git-svn-id: trunk@8901 -
2007-10-21 19:53:11 +00:00
florian
7878f0feb3 * made all fpc_*_to_shortstr helpers a procedure, resolves #8580
git-svn-id: trunk@8898 -
2007-10-21 17:33:18 +00:00
florian
3752cfb649 * fix cotan, resolves #9944
git-svn-id: trunk@8819 -
2007-10-15 17:25:52 +00:00
Jonas Maebe
cc843f5f62 * fixed comparedword in case len=0
git-svn-id: trunk@8815 -
2007-10-15 13:44:57 +00:00
peter
e8322a83e4 * optimized generic implementations to use pointers instead of
array accesses. This also allows better register variable
    optimizations

git-svn-id: trunk@8742 -
2007-10-07 18:36:18 +00:00
Jonas Maebe
f51bac256d * made pic-safe (by simply not accessing global variables from
assembler code anymore, as I don't see how to easily support
    PIC access to global variables for Darwin/i386)

git-svn-id: trunk@8663 -
2007-09-27 15:08:09 +00:00
micha
d95b9082bb * add memory barrier prototypes + implementation for i386 and sparc + use lwsync for powerpc64
git-svn-id: trunk@8240 -
2007-08-07 20:37:08 +00:00
florian
6f6e60b723 * improved SSE check by Martin Schreiber
git-svn-id: trunk@8040 -
2007-07-13 21:16:48 +00:00
daniel
b293a9bdd4 - Revert patch by Alex Raine.
git-svn-id: trunk@7669 -
2007-06-15 14:13:26 +00:00
daniel
14f2edc56e + Assembler implementation of min/max by Alex Rayne.
git-svn-id: trunk@7668 -
2007-06-15 13:59:20 +00:00
florian
80d55022e9 * tried to fix SSE OS support detection
git-svn-id: trunk@7553 -
2007-06-02 19:26:06 +00:00
florian
a8062fafe7 * valgrind safe primitive move implementation
git-svn-id: trunk@7503 -
2007-05-28 18:21:44 +00:00
Jonas Maebe
5cf2511283 - removed/disabled all normalset code (except for bootstrapping), since
it's no longer needed

git-svn-id: trunk@7403 -
2007-05-20 11:19:36 +00:00
florian
1cd41b68f7 * compiler handled round/trunc for x86
git-svn-id: trunk@6827 -
2007-03-13 20:23:24 +00:00
ivost
2c22173534 * Added InterlockedCompareExchange64 for i386 (supported over cmpxchg8b instruction)
git-svn-id: trunk@6753 -
2007-03-09 09:29:17 +00:00
marco
57cfd89fbd * Linux 64-bit FS patch from neli for FPC_USE_LIBC
git-svn-id: trunk@6517 -
2007-02-16 22:16:54 +00:00
peter
cd767cf5c8 * i386 uses generic fpc_compare_shortstr_equal
git-svn-id: trunk@6335 -
2007-02-04 16:24:02 +00:00
Jonas Maebe
5ea26c5f5f * fixed darwin compilation
git-svn-id: trunk@6310 -
2007-02-02 22:48:02 +00:00
florian
bcf65051e3 * detect sse properly on linux, fixes tmask*.pp in the testsuite with -Cfsse2
git-svn-id: trunk@6277 -
2007-01-31 21:51:47 +00:00
daniel
e8050233af + Patch by Thorsten Engler to optimize string comparisons for = and <> cases. (Mantis #8231)
git-svn-id: trunk@6241 -
2007-01-28 10:58:52 +00:00
florian
685844b61c * set SSE CSR if sse is supported by the CPU
git-svn-id: trunk@6021 -
2007-01-17 08:27:48 +00:00
florian
b3a1868ff0 * correct masking of exceptions in genmath code
git-svn-id: trunk@5965 -
2007-01-14 10:44:10 +00:00
Jonas Maebe
3381fd2ea1 * ismultithread is a longbool nowadays (cmpb -> cmpl)
git-svn-id: trunk@5893 -
2007-01-11 17:32:15 +00:00
florian
4b88079c41 + i386 assembler implementations of tan, cotan, sincos
* default tan, cotan use now sincos

git-svn-id: trunk@5809 -
2007-01-04 12:58:40 +00:00
florian
90787bb436 * disable sse/mmx fastmode code in case of valgrind usage
git-svn-id: trunk@5342 -
2006-11-12 16:54:44 +00:00
florian
4ee9f1b394 * use DefaultCW to reload fpu settings
git-svn-id: trunk@5169 -
2006-11-02 08:58:01 +00:00
Jonas Maebe
5acc8b44a8 * activated internal get_frame for x86
* turn off stackframe optimizations on x86 if get_frame is called
    in the current routine, or if the address of a nested function
    is taken in the current routine
  + test for the above
  * this fixes the IDE when compiled with stackframe optimizations
    on x86

git-svn-id: trunk@5146 -
2006-11-01 12:48:53 +00:00
peter
2f0ce31751 * string concat changed from function to procedure to
allow runtime optimization if the destination is the
    same as a source parameter
  * tassignmentnode now sets aktassignmentnode global that can be used
    to use the left node as a destination parameter and
    skip the assignment
  * disabled all cpu specific shortstr concat/append

git-svn-id: trunk@4770 -
2006-10-02 20:00:06 +00:00
florian
0ade2e487d * cleaned up and fixed (from mftq75) sse unit initialization, fixes #7268
git-svn-id: trunk@4501 -
2006-08-24 20:00:06 +00:00
Jonas Maebe
982c1a39c4 * and yet more ebx saving/restoring
git-svn-id: trunk@4176 -
2006-07-14 11:57:07 +00:00
florian
b24388da8d * fix comparedword, fixes #7143
git-svn-id: trunk@4153 -
2006-07-12 20:50:48 +00:00
florian
81b98c9d82 + added more ebx saving to cpuid invocations
git-svn-id: trunk@4147 -
2006-07-12 07:23:03 +00:00
Jonas Maebe
bddcc8f902 * save/restore ebx around cpuid invocations
git-svn-id: trunk@4145 -
2006-07-11 20:50:44 +00:00
florian
6185a32923 * amd specific flags are checked against cpu vendor
git-svn-id: trunk@4138 -
2006-07-10 18:03:30 +00:00
Jonas Maebe
b9125f97d0 * fixed InterlockedCompareExchange
git-svn-id: trunk@3947 -
2006-06-25 18:52:12 +00:00
peter
4c065bce45 * move InterLocked functions to system unit
git-svn-id: trunk@3933 -
2006-06-25 09:26:23 +00:00
daniel
6f1e0eff45 * Add ifdef internal_backtrace
git-svn-id: trunk@3518 -
2006-05-14 08:49:34 +00:00
florian
47fac4ee6e * moved *SSECSR to system unit
* exposed cpu feature detection on i386 from system unit
+ SafeLoadLibrary

git-svn-id: trunk@3481 -
2006-05-11 19:13:54 +00:00
peter
2d3c01d83a * basic work to make Move() and FillChar() internal under ifdef INTERNALMOVEFILLCHAR
git-svn-id: trunk@3021 -
2006-03-24 17:32:39 +00:00
Jonas Maebe
83b1c3736a * fixed assembler version of fpc_AnsiStr_Decr_Ref (based on remark
from Martin Schreiber)

git-svn-id: trunk@2856 -
2006-03-11 14:16:11 +00:00
Jonas Maebe
15270b67be - disabled assembler ansistring helpers for darwin/x86 because they don't
keep 16 byte stack alignment

git-svn-id: trunk@2852 -
2006-03-11 13:18:28 +00:00
peter
f411f615c6 * nostackframe directives
git-svn-id: trunk@2648 -
2006-02-20 10:37:42 +00:00
peter
340cf721b5 * faster int_str
* removed i386 int_str since the generic implementation is faster

git-svn-id: trunk@2646 -
2006-02-20 08:31:20 +00:00
florian
0d61229d63 * fpc_ansistring_decr improved
git-svn-id: trunk@2629 -
2006-02-19 14:38:06 +00:00
florian
af9d6344b8 * activated fastmove on win32
git-svn-id: trunk@2218 -
2006-01-08 10:10:28 +00:00
florian
fb4557d71e * first implementation of pic for i386
git-svn-id: trunk@2107 -
2006-01-01 20:14:48 +00:00
Jonas Maebe
5b9f58ef73 * reverted r2037 because of braindead "out" semantics for refcounted
types

git-svn-id: trunk@2045 -
2005-12-24 01:06:05 +00:00
Jonas Maebe
84aab46550 * changed formal "var" paramters of move, movechar0 and fill* into
"out" parameters to avoid wrong "uninitialized" warnings

git-svn-id: trunk@2037 -
2005-12-23 20:59:08 +00:00
florian
e2a4dac215 + i386 assembler versions of decr_ansistring and unique_ansistring
git-svn-id: trunk@1963 -
2005-12-15 19:51:57 +00:00
marco
8f50bb556e * small fix to inclocked inlining changes
git-svn-id: trunk@1340 -
2005-10-09 11:06:47 +00:00
marco
5c92d88da8 * inlined a few funcs, inlined the non threaded part of inc/declocked. The lock prefixed one
remains a separate proc, as it is slow anyway.

git-svn-id: trunk@1339 -
2005-10-09 10:57:29 +00:00
michael
9beb37574e + Removed SHORTSTRCOMPAREINREG define
git-svn-id: trunk@309 -
2005-06-09 17:59:46 +00:00
michael
addadd41bf + removed HAS_GENERICCONSTRUCTOR define
git-svn-id: trunk@297 -
2005-06-09 17:43:58 +00:00
michael
a775d3d487 + Removed HASCURRENCY VALUEGETMEM VALUEFREEMEM defines
git-svn-id: trunk@293 -
2005-06-09 17:33:22 +00:00
michael
3a2eaa94b1 + Removed INTERNCONSTINTF define
git-svn-id: trunk@267 -
2005-06-07 22:04:18 +00:00
michael
93ba0409be + Removed HASCOMPILERPROC define
git-svn-id: trunk@265 -
2005-06-07 21:41:02 +00:00
michael
0fe5266f4d + Removed NOBOUNDCHECK define
git-svn-id: trunk@263 -
2005-06-07 21:17:22 +00:00
michael
db5d9d75aa + Removed INTERNCOPY define
git-svn-id: trunk@259 -
2005-06-07 21:11:49 +00:00
peter
4ace790492 * remove $Log
git-svn-id: trunk@231 -
2005-06-07 09:47:55 +00:00
fpc
790a4fe2d3 * log and id tags removed
git-svn-id: trunk@42 -
2005-05-21 09:42:41 +00:00
michael
ffb55c4254 + Removed VER1_0 defines
git-svn-id: trunk@33 -
2005-05-19 21:14:45 +00:00
fpc
50778076c3 initial import
git-svn-id: trunk@1 -
2005-05-16 18:37:41 +00:00
peter
16b121f8c4 * unblock signal in signalhandler, remove the sigprocmask call
from setjmp
2005-04-24 21:19:22 +00:00
florian
9445db95f3 * widestring manager can handle now ansi<->wide string conversions even if the lens don't match 2005-03-14 21:09:04 +00:00
florian
f94047b703 * some floating point routines improved 2005-03-13 17:14:46 +00:00
florian
3e5f68d360 * C. Western: utf-8 reading from resource files 2005-03-09 20:50:11 +00:00
peter
e417e34496 * truncate log 2005-02-14 17:13:06 +00:00
peter
ae8b7b153b * include file cleanup 2005-02-13 20:01:37 +00:00
florian
d7dbd71881 * fastmove disabled, has too much drawbacks 2005-01-26 20:58:10 +00:00
peter
8be342c36b * pinteger in non-delphi/objfpc mode is psmallint 2005-01-24 18:03:19 +00:00
peter
445fe79686 * no sigprocmask for 1.0.x bootstrapping 2005-01-24 16:54:16 +00:00
florian
8a2bc0e507 * 1.0.x fix 2005-01-24 08:17:00 +00:00
florian
6780d226ce * made IA32 the default 2005-01-23 20:10:05 +00:00
florian
3600b51d32 + fastmove from John O'Harrow integrated 2005-01-23 20:03:23 +00:00
peter
7395e7e546 * restore sigprocmask for linux 2005-01-20 16:38:28 +00:00
florian
43e425072b * fixed div by zero check for qword mod qword 2004-11-21 16:41:16 +00:00
peter
2633b4883f * external fixes 2004-11-21 16:33:55 +00:00
peter
4d8460ec2f * float routines all use internproc and compilerproc helpers 2004-11-21 15:35:23 +00:00
peter
14eb8f59b4 internconst, internproc and some external declarations moved to interface 2004-11-17 22:19:04 +00:00
peter
e85d8c970c fixed result of strlcopy 2004-11-04 20:57:33 +00:00
florian
07c28067aa * fixed sse exception handling 2004-11-02 15:26:21 +00:00
peter
f80f3fc0fa * shortstr compare with empty string fixed
* removed special i386 code
2004-11-01 12:43:28 +00:00
peter
fd759e521f * saveregisters calling convention is obsolete 2004-10-24 20:01:41 +00:00
florian
9205c6e800 * fixed overflow checking in qword multiplication 2004-09-26 08:52:51 +00:00
florian
051126d079 * assembler version of mod_qword_word reactivated 2004-07-25 11:50:39 +00:00
florian
7a7c34e1b7 + several unicode (to/from utf-8 conversion) stuff added
* some longint -> SizeInt changes
2004-07-18 20:21:44 +00:00
Jonas Maebe
d43436e1cd * fixed indexbyte/word/dword when length is 0 (return -1 instead of 0) 2004-07-18 16:40:08 +00:00
peter
783f8068a5 * add fclex for fpu exceptions to round/trunc 2004-07-09 23:06:11 +00:00
daniel
509870553d * Aligment code in fillchar proved to slow down stuff seriously instead of
speeding it up. This is logical, the compiler aligns everything very well,
    it is possible that fillchar gets called on misaligned data, but it seems
    this never happens.
2004-07-07 17:38:58 +00:00
florian
6427a54c3d * fixed some bootstrapping problems as well as some 64 bit stuff 2004-06-13 10:49:50 +00:00
florian
7f2606ed8a * fpc_mul_qword uses always the assembler implementation 2004-05-10 20:58:20 +00:00
peter
61036b2149 * fixed wrong typecasts 2004-05-09 15:47:56 +00:00
Jonas Maebe
dedf697f3a * use some more string routines from libc if FPC_USE_LIBC is used 2004-05-01 15:26:33 +00:00
peter
269d24955a * inclocked_longint ifdef fixed 2004-04-29 20:00:47 +00:00
peter
b4474438d9 * FPC_MOVE alias 2004-04-26 15:55:01 +00:00
marco
e546db7a23 * interlocked* changed to longints, including winapi. (which was a bug) 2004-03-05 12:17:50 +00:00
florian
c9122a4719 + completed x86-64/linux system unit 2004-02-05 01:16:12 +00:00
Jonas Maebe
5439babeb5 + cgeneric.inc: implementations of rtl routines based on libc
* system.inc: include cgeneric.inc before powerpc.inc/i386.inc/... if
    FPC_USE_LIBC is defined
  * powerpc.inc, i386.inc: check whether the routines they implement aren't
    implemented yet in another include file (cgeneric.inc)
2004-01-11 11:10:07 +00:00
Jonas Maebe
6bd3eccdac + fpc_cpuinit procedure to allow cpu/fpu initialisation before any unit
initialises
  + fpu exceptions for invalid operations and division by zero enabled for
    ppc
2004-01-02 17:21:50 +00:00
peter
31a33a1a7b * fixed indexbyte for regcall 2003-12-24 23:07:28 +00:00
peter
bcdf065947 * fix set_set_range 2003-12-24 22:38:24 +00:00
peter
d11cecb354 * removed assembler
* cleanup
2003-12-24 22:27:13 +00:00
peter
0105b62700 * fix call to handleerror for regcall 2003-12-23 23:09:43 +00:00
peter
c1a25f53b6 * increment should use ofcourse use 1 instead of -1 2003-12-23 22:23:34 +00:00
michael
71da578b33 + Patch from Peter to fix strlcomp 2003-12-19 09:28:00 +00:00
peter
6b6741b41c * win32 regcall patches 2003-12-17 21:56:33 +00:00
peter
48b84272bf * fix warning in gas 2003-12-04 21:44:39 +00:00
peter
0d3948fc5f * register calling updates 2003-12-04 21:42:07 +00:00
peter
07f38cd53a * more fixes 2003-12-03 23:06:35 +00:00
peter
e0bf2ea8d1 * typo in sub_set 2003-12-03 22:46:54 +00:00
Jonas Maebe
89d4ca293b * fix power() for negative base 2003-11-29 16:40:12 +00:00
michael
8373e40e10 + Patch from Johannes Berg for bug #2759 2003-11-24 21:57:43 +00:00
peter
23ae4449ec * register calling updates 2003-11-23 16:50:49 +00:00
peter
5e5462edd1 * make strpas assembler function 2003-11-19 16:58:44 +00:00
peter
8f8b47194a * REGCALL define added 2003-11-11 21:08:17 +00:00
marco
b15ba0a5ae * Peter's Cardinal<->Longint fixes patch 2003-11-03 09:42:27 +00:00
mazen
84e8f0e96b + assembler mode forced to ATT style for x86 cpu 2003-10-31 09:20:11 +00:00
marco
41e3249c4b * fix from peter i386.inc to circumvent ebx destroying 2003-10-27 09:16:57 +00:00
florian
4dbc76b90d * fixed arctan2 to handle x=0 correctly as well 2003-10-26 15:58:05 +00:00
peter
1b2d0e4de9 * save edi,ebx,esi in int_str 2003-10-23 17:01:27 +00:00
peter
26434b49f1 * use __HIGH() 2003-10-16 21:28:40 +00:00
florian
a8374188bc + some code for PIC support added 2003-10-14 00:57:48 +00:00
peter
abbc3fa755 * moved int64 asm code to int64p.inc
* save ebx,esi
2003-09-14 11:34:13 +00:00
peter
cba9b5206d * save edi,esi,ebx 2003-09-08 18:21:37 +00:00
peter
a980741757 * generic string routines added 2003-07-07 20:22:05 +00:00
Jonas Maebe
f0227e6a17 * fpc_shortstr_append_shortstr has to use high(s1) instead of 255 as
maxlen
  + ppc version of fpc_shortstr_append_shortstr
2003-06-01 14:50:17 +00:00
peter
84f2b74142 * FPC_SHORTSTR_APPEND_SHORTSTR public added 2003-05-26 21:18:13 +00:00
peter
60407e302b * fpc_shortstr_concat is now the same for all targets
* fpc_shortstr_append_shortstr added for optimized code generation
2003-05-26 19:36:46 +00:00
florian
35bb4d682f * fixed generic shortstr_compare 2003-05-16 22:40:11 +00:00
florian
fa4496bddf + support for generic pchar routines added
+ some basic rtl stuff for x86-64 added
2003-04-30 16:36:39 +00:00
florian
345117b9ba * initial implementation with code from math.pp 2003-04-24 09:16:31 +00:00
peter
a7691d8bb8 * fpc_round added, needed for int64 currency 2003-04-23 21:28:21 +00:00
michael
27062c2e01 + Patch from Mattias Gaertner for single typeinfo 2003-03-29 16:55:56 +00:00
peter
7a2f4e9e26 * ifdef HAS_GENERICCONSTRUCTOR 2003-03-26 00:19:10 +00:00
peter
1e86b585f9 * changed address parameter/return values to pointer instead
of longint
2003-03-17 14:30:11 +00:00
Jonas Maebe
15419c49a6 - removed buggy i386-specific FPC_CHARARRAY_TO_SHORTSTR
* fixed generic FPC_CHARARRAY_TO_SHORTSTR (web bug #2382)
  * fixed some potential range errors in indexchar/word/dword
2003-02-18 17:56:06 +00:00
carl
c430df4732 * round bugfix with -Or switch 2003-02-05 19:53:17 +00:00
peter
dd08014a49 * use generic int64 power 2003-01-15 00:45:17 +00:00
peter
166fe37203 * power returns int64 2003-01-15 00:40:18 +00:00
mazen
675f2770da + defining FPC_SYSTEM_HAS_DECLOCKED and FPC_SYSTEM_HAS_INCLOCKED to avoid
compilation error on generic.inc
2003-01-06 23:03:13 +00:00
peter
892e9c864a * i386 fpu controlword functions added 2003-01-03 20:34:02 +00:00
peter
69ec4efe4d * fix possible overflow when array len > 255 when converting to
shortstring
2003-01-03 17:14:54 +00:00
peter
8bb7e03d19 * fixed return value when len=0 for indexchar,indexword 2002-12-15 22:32:25 +00:00
florian
f3f3ed17b7 * AMD CPUs could have SSE as well 2002-11-29 00:06:24 +00:00
carl
dec8b274f3 * avoid crashes with negative len counts on fills/moves 2002-10-20 11:50:57 +00:00
carl
9afe6d5835 * Replace 220 -> 219 2002-10-15 19:23:43 +00:00
peter
5fff238567 * threads unit added for thread support 2002-10-14 19:39:16 +00:00
peter
ecc3b24fd4 * round returns int64 2002-10-06 21:26:17 +00:00
peter
ea8118b37c * fpc_pchar_length compilerproc and strlen alias 2002-10-05 14:20:16 +00:00
peter
603b3bc48b * Copy() changed to internal function calling compilerprocs
* FPC_SHORTSTR_COPY renamed to FPC_SHORTSTR_ASSIGN because of the
    new copy functions
2002-10-02 18:21:51 +00:00
Tomas Hajny
d7620aec7b * Opcode syntax corrected 2002-09-23 18:55:19 +00:00
Tomas Hajny
67d1c93691 * femms replaced with opcode to allow older as 2002-09-23 18:25:55 +00:00
carl
876fca9e8f * make the record packed (just in case) 2002-09-15 10:03:04 +00:00
carl
5ed7522ade - removed unused defines 2002-09-07 21:32:08 +00:00
peter
9f31783a0a * old logs removed and tabs fixed 2002-09-07 16:01:16 +00:00
peter
685caf03b6 * add alias for fpc_dispose_class so it can be called from
fpc_help_fail_class
2002-09-03 15:43:36 +00:00
peter
24384c648b * moved genrtti.inc code to rtti
* removed rttip.inc, the generic code is almost as fast and
    much easier to maintain and has less risks on bugs
2002-09-02 18:42:41 +00:00
peter
c3af3cda9b * SYSTEMINLINE define that will add inline directives for small
functions and wrappers. This will be defined automaticly when
    the compiler defines the HASINLINE directive
2002-08-19 19:34:02 +00:00
Jonas Maebe
0ab6c37c31 * fixed compilation errors with 1.0 2002-07-31 11:52:57 +00:00
florian
3f2a82784e * changed multi threading define: it's MT instead of MTRTL 2002-07-26 15:45:33 +00:00
carl
dfa44b9f00 + added TEST_GENERIC to test generic version 2002-07-06 20:31:59 +00:00
carl
2604402b16 + TEST_GENERIC to test generic versions of code 2002-06-16 08:21:26 +00:00
Jonas Maebe
0d606ac980 * fixed memory corruption bug in fpc_help_constructor 2002-06-09 12:54:37 +00:00
peter
90ff7a6db8 * fpc_freemem and fpc_getmem compilerproc 2002-04-21 18:56:59 +00:00
carl
fe48bc8f27 - no need for runerror 203, already fixed! 2002-04-01 14:23:17 +00:00
carl
16b2fe5095 * cause runtime error 203 on failed class creation 2002-03-30 14:52:04 +00:00
peter
e6841b9f3c * fixed load_smallset 2002-03-29 20:15:44 +00:00
peter
3d1d485d7c * femms added 2002-03-16 12:01:24 +00:00
peter
97f28dbdfb * sse and 3dnow extensions from Michail added 2002-03-16 11:51:50 +00:00
peter
21a8c2cc8e * merged fixes from 1.0.x 2001-12-26 21:03:56 +00:00
peter
7484da248d * freemem(var) -> freemem(value) 2001-12-03 21:39:19 +00:00
florian
5050f1ff66 * init and final code in genrtti.inc updated 2001-11-17 16:56:08 +00:00
michael
2a27de0fb0 + Initial variant support 2001-11-14 22:59:11 +00:00
peter
9aeb958a8e * InterLocked routines added 2001-10-22 21:40:55 +00:00
peter
df8692dd2e * reset %edx in getordprop so functions returning only %eax don't
generate rangecheck errors
2001-10-20 17:25:22 +00:00
carl
901072c39e * bugfix #1639 (IsMultiThread varialbe setting) 2001-10-09 02:42:05 +00:00
Jonas Maebe
ed449defca + searchsystype() and searchsystype() functions in symtable
* changed ninl and nadd to use these functions
  * i386 set comparison functions now return their results in al instead
    of in the flags so that they can be sued as compilerprocs
  - removed all processor specific code from n386add.pas that has to do
    with set handling, it's now all done in nadd.pas
  * fixed fpc_set_contains_sets in genset.inc
  * fpc_set_in_byte is now coded inline in n386set.pas and doesn't use a
    helper anymore
  * some small fixes in compproc.inc/set.inc regarding the declaration of
    internal helper types (fpc_small_set and fpc_normal_set)
2001-09-04 11:38:54 +00:00
Jonas Maebe
f256a47f04 * compilerproc implementation of set addition/substraction/...
* changed the declaration of some set helpers somewhat to accomodate the
    above change
  * i386 still uses the old code for comparisons of sets, because its
    helpers return the results in the flags
  * dummy tc_normal_2_small_set type conversion because I need the original
    resulttype of the set add nodes
  NOTE: you have to start a cycle with 1.0.5!
2001-09-03 13:27:41 +00:00
Jonas Maebe
41a57028b9 * converted adding/comparing of strings to compileproc. Note that due
to the way the shortstring helpers for i386 are written, they are
    still handled by the old code (reason: fpc_shortstr_compare returns
    results in the flags instead of in eax and fpc_shortstr_concat
    has wierd parameter conventions). The compilerproc stuff should work
    fine with the generic implementations though.
  * removed some nested comments warnings
2001-08-30 15:43:14 +00:00
Jonas Maebe
06f9cd4c96 * some fixes in compilerprocs for chararray to string conversions
* conversion from string to chararray is now also done via compilerprocs
2001-08-29 19:49:03 +00:00
Jonas Maebe
fc92c3b336 + compilerproc implementation of most string-related type conversions
- removed all code from the compiler which has been replaced by
    compilerproc implementations (using {$ifdef hascompilerproc} is not
    necessary in the compiler)
2001-08-28 13:24:46 +00:00
peter
75676a4866 * moved i386 specific code to include file 2001-08-04 11:03:42 +00:00
Jonas Maebe
b9f6efc85b + "compproc" helpers
* renamed several helpers so that their name is the same as their
    "public alias", which should facilitate the conversion of processor
    specific code in the code generator to processor independent code
  * some small fixes to the val_ansistring and val_widestring helpers
    (always immediately exit if the source string is longer than 255
     chars)
  * fixed fpc_dynarray_high and fpc_dynarray_length if the dynarray is
    still nil (used to crash, now return resp -1 and 0)
2001-08-01 15:00:09 +00:00
peter
e7c86006db * m68k updates merged 2001-07-30 21:38:54 +00:00
peter
ca8d604964 * various widestring updates, it works now mostly without charset
mapping supported
2001-07-08 21:00:14 +00:00
florian
a4755024cc * some fixes for widestrings and variants 2001-05-31 22:42:56 +00:00
peter
8bf13fd185 *** empty log message *** 2001-05-09 19:57:07 +00:00
peter
802acf6940 * m68k updates 2001-04-23 18:25:44 +00:00
peter
64b57cd77d * add nop after popa (merged) 2001-04-21 12:18:09 +00:00
peter
73a59b9ca3 * merged popal fix 2001-04-16 11:22:53 +00:00
peter
e150c46c78 * remove warnings 2001-04-13 22:26:32 +00:00
Jonas Maebe
547bc5f275 * optimized FPC_HELP_CONSTRUCTOR a bit 2001-04-08 13:19:28 +00:00
Jonas Maebe
ab2ea0a472 * moved implementations of strlen and strpas to separate include files
(they were duplicated in i386.inc and strings.inc/stringss.inc)
  * strpas supports 'nil' pchars again (returns an empty string)
    (both merged)
2001-03-05 17:10:04 +00:00