 sergei
|
274c2fc82f
* handle_ln_const: if/else branches were swapped, correct values are: ln(0)=-Inf and ln(-#)=NaN.
|
há 11 anos atrás |
 marco
|
aebe66d0cf
* some fixes and a lot of updating of Hungarian names. Mantis #26196
|
há 11 anos atrás |
|
marco
|
b7aff4eae0
* flush output on halt. Patch by Mattias mantis #26225
|
há 11 anos atrás |
|
marco
|
0a0739b916
* improved winexec comment, updated copyright year and add minor comment to gettickcount64
|
há 11 anos atrás |
|
sergei
|
a8bc2085e8
- Delete strpas.inc files, no longer used, their contents has been merged into corresponding $(CPU).inc.
|
há 11 anos atrás |
|
sergei
|
e5f6f9c3a1
- powerpc and powerpc64: cleaned out code corresponding to FPC_STRTOSHORTSTRINGPROC not defined (obsolete and removed from other targets long ago). Also merged strpas.inc files into powerpc*.inc, because by now they are included only once and therefore don't need to be separate files.
|
há 11 anos atrás |
|
sergei
|
a94187c79b
- ARM: cleaned out code corresponding to FPC_STRTOSHORTSTRINGPROC not defined (obsolete and removed from other targets long ago).
|
há 11 anos atrás |
 reiniero
|
e8a3715afd
* fcl-db: limit blob memory usage when using unidirectional datasets. Patch by Laco. Issue #24509
|
há 11 anos atrás |
 Jonas Maebe
|
19c310f5d3
* fixed unit name
|
há 11 anos atrás |
 Károly Balogh
|
f3bbad26c3
m68k: use MOVEA without size to load values to address regs. use a_op_const_reg in a_op_const_ref for smarter const loading when necessary
|
há 11 anos atrás |
 nickysn
|
47a7a9835a
* added hack to support stack exceeding 32kb in size in the i8086 far data
|
há 11 anos atrás |
 svenbarth
|
2c6d3f1428
Read the deref of the fieldvarsym into the correct variable. I wonder how this ever went past the test suite O.o
|
há 11 anos atrás |
|
nickysn
|
e9cb275f66
* put the i8086-msdos stack segment in a separate object file
|
há 11 anos atrás |
|
nickysn
|
d4565aeec2
* in i8086 far data memory models add reference to the beginning of stack from
|
há 11 anos atrás |
|
Jonas Maebe
|
03ef4b6477
+ test for mantis #26226 (already works with FPC 2.7.1)
|
há 11 anos atrás |
|
nickysn
|
daf71e6d88
* fixed multiplication by 64-bit constants on x86_64, mantis #26230
|
há 11 anos atrás |
|
Károly Balogh
|
2936335f68
removed silly ancient writeln debug from me...
|
há 11 anos atrás |
|
nickysn
|
3cc8ff11e3
+ generate the stack segment for i8086 far data memory models from within fpc
|
há 11 anos atrás |
|
sergei
|
3c1368903f
* Updated header: this file is in cp866 encoding.
|
há 11 anos atrás |
|
sergei
|
743326dc70
* abs(): when translated to generic fpc_abs_real() helper, typecast result back to original precision, similar to how it is done for sqr and sqrt. Fixes compilation on m68k after r27808 (m86k looks like the only target still needing this generic handling).
|
há 11 anos atrás |
|
Károly Balogh
|
28323135f8
m68k: minor optimizations to g_proc_exit() - use LEA for stackpointer math which is better than ADDing large constants, also only modify the SP reg once
|
há 11 anos atrás |
|
nickysn
|
02dff17ae0
* taddrnode: fixed getting the address of nested procs in tp mode in i8086 far
|
há 11 anos atrás |
|
sergei
|
37625211f5
* Fixed sqrt procedure selection on arm-wince after r27809.
|
há 11 anos atrás |
|
nickysn
|
cb05907e18
* fixed the global function to nested procvar conversion in the i8086 far data
|
há 11 anos atrás |
|
marco
|
46486a9e63
* add comments wrt TIFF magic numbers, mantis #26210
|
há 11 anos atrás |
|
nickysn
|
24fcac9f87
+ added support for nested procvars in the i8086 far data memory models
|
há 11 anos atrás |
 yury
|
f526f39f8f
* pas2jni: Always specify unit prefix for types in the generated bridge code.
|
há 11 anos atrás |
|
sergei
|
1c7a5d5adc
* x86: fix spilling of sqrtss and sqrtsd instructions, they cannot have memory as destination.
|
há 11 anos atrás |
|
sergei
|
22e099d000
* sqrt(real): for targets with emulated floating point, invoke float64_sqrt or float32_sqrt from softfpu code. Testing on ARM CPU shows that float64_sqrt executes about twice faster and offers better accuracy than fpc_sqrt_real from genmath.inc.
|
há 11 anos atrás |
|
sergei
|
fc5f45f65c
* sqr(real) and sqrt(real): remove typeconv node inserted by initial call processing (see explanation in comments), allowing these functions to be evaluated using precision of argument. In particular, sqrt(single) and sqrt(double) now emit 'sqrtss' and 'sqrtsd' instructions on x86 targets with -Cfsse3. Non-x86 targets already have the necessary support in code generators.
|
há 11 anos atrás |
