 nickysn
|
90b69184f1
* use a {$if defined()} sequence, instead of nested ifdefs for selecting the
|
11 rokov pred |
 sergei
|
65d24e000f
* SPARC: generate PIC prologue as recommended by ABI, it does not require FPC_GETGOT helper.
|
11 rokov pred |
|
sergei
|
6469d41e2a
* SPARC: removed 32 bit shift code, and adjusted 64-bit shifts to take advantage of 3-address instructions (a port of r26142 for MIPS)
|
11 rokov pred |
|
sergei
|
58cc531dd9
* SPARC: convert from int64/qword to float using genmath helpers. Removes dependency on softfloat code.
|
11 rokov pred |
|
sergei
|
5ccdfcf106
* SPARC: optimized a_op_const_reg and a_op_const_reg_reg, take advantage of optimize_op_const functionality.
|
11 rokov pred |
|
sergei
|
16156cbcd7
- SPARC RTL: removed assembler PIC/threadvar hacks.
|
11 rokov pred |
|
sergei
|
9a486d73ba
+ SPARC: support FNEGd/FNEGq internal instructions, and use them to implement floating-point negation more efficiently.
|
11 rokov pred |
|
sergei
|
8ecad469a6
* Reverted r26256 since it produced wrong results if argument of inc/dec is actually unaligned. Added description of the issue and slightly increased node complexity at which address is cached, so operations on record/class fields do not take address.
|
11 rokov pred |
|
sergei
|
c3fe0c7379
* On SPARC, using shifts for zero-extending to OS_16 is more efficient than ANDing with 65535 (two instructions and no temp registers vs. three instructions and a temp register).
|
11 rokov pred |
|
sergei
|
b376f56bf3
* Use typed pointer to store address of x when processing inc(x)/dec(x) with range checking enabled. This preserves alignment of x and prevents generating unaligned code on alignment-sensitive targets.
|
11 rokov pred |
|
sergei
|
5cd0684d45
* tcg.a_load_ref_reg_unaligned: if loading signed 16-bit value, treat upper half as signed 8-bit to ensure that result is properly extended. This routine is used only by SPARC target (other CPUs use specific implementations), and it generates pretty inefficient code, therefore it's probably worth to drop it altogether and implement SPARC-specific version. Mantis #25440.
|
11 rokov pred |
 reiniero
|
48e4fc4bb2
* fcl-db: dbf,bufdataset (and descendents): allow filter using double delimiters inside string. Fixes issue #25432
|
11 rokov pred |
|
reiniero
|
4229c6b40f
* fcl-db: additional stringfilter tests
|
11 rokov pred |
|
sergei
|
f2096de53a
* tcg386.g_proc_exit: instead of recalculating stack size, use current_procinfo.final_localsize which was calculated in g_proc_entry.
|
11 rokov pred |
|
sergei
|
258e8c58a4
+ Added trecorddef.buildderefimpl, necessary since advanced records have been implemented. Fixes compiler crash on webtbs/tw18688.pp on x86_64-win64.
|
11 rokov pred |
|
reiniero
|
f27d3baccc
* fcl-db: test filter using double delimiters inside string to test syntax as presented in issue #25432
|
11 rokov pred |
|
reiniero
|
8c6fb52940
* fcl-db: cosmetic
|
11 rokov pred |
|
reiniero
|
e2d37b5561
* fcl-db: examples: dbftool: insert field data with ' useful for testing filters etc
|
11 rokov pred |
|
reiniero
|
50a349eb1c
* fcl-db: examples: slight fixes dbftool
|
11 rokov pred |
|
nickysn
|
be5f49b3e5
* tx86addnode.second_cmpfloat: on i8086 only load the high byte of the FPU
|
11 rokov pred |
|
nickysn
|
ada9ccf6a6
* tx86addnode.second_cmpfloat: do not emit a fwait after fstsw as my own testing
|
11 rokov pred |
|
nickysn
|
f23c521c50
- disabled the i8086 math.arctan2 asm code, because it doesn't work on the 8087/80287
|
11 rokov pred |
|
nickysn
|
2a0a8cfc00
* fixed CGA 320x200 hline bug when CurrentColor=1
|
11 rokov pred |
|
sergei
|
b82b6da493
+ Win32 SEH: to avoid memory leaks, it is necessary to free exception object for non-FPC exceptions on re-raise (it will be re-created by next exception handler).
|
11 rokov pred |
|
reiniero
|
ee152cde6a
* fcl-db TMemDataSet show Filter property is unimplemented; mantis issue #22051
|
11 rokov pred |
|
sergei
|
695f271b8c
+ Win32 SEH: actually install the exception handler around main procedure (write to fs:(0) was missing).
|
11 rokov pred |
|
sergei
|
74f9d719f1
* i386-win32 setjmp and longjmp: if SEH support is enabled, save/restore head of exception registration chain. Without this, a longjmp out of a function with exception frame(s) will leave these exception frames below ESP, which will fail integrity checks in OS and cause process termination if an exception occurs later on (or they may be overwritten by subsequent operations on stack and, again, cause undesired results).
|
11 rokov pred |
|
nickysn
|
779b92459e
* fixed ldexp() on 16/8-bit CPUs. Among other things, this fixes sin() and cos()
|
11 rokov pred |
|
reiniero
|
a3dab9b686
* fcl-db: cosmetic
|
11 rokov pred |
|
reiniero
|
586bca0c7d
* fcl-db TMemDataset.CopyFromDataset: return to original record position. Fixes issue #25426
|
11 rokov pred |
