Commit Graph

41059 Commits

Author SHA1 Message Date
michael
3ffae89e12 * Null stream implementation
git-svn-id: trunk@25495 -
2013-09-16 08:24:46 +00:00
paul
99dadf2998 compiler: allow comparative operators to have result other than Boolean. Reasons:
1. It allows to use comparative operators in some unusual cases (issue #25004).
2. Regular type checking does not allow to use other than Boolean types in IF expressions anyway.
3. Delphi compatibility (although Delphi documentation states otherwise).

git-svn-id: trunk@25494 -
2013-09-16 05:10:31 +00:00
paul
93f1ba4493 compiler: don't allow to access an enum members through the enum member. It should be only possible throught the enum type name. Fixes issue #0025029
git-svn-id: trunk@25493 -
2013-09-16 02:48:52 +00:00
paul
9577aa1d0b packages: fixed syntax error in fpmake.pp
git-svn-id: trunk@25492 -
2013-09-16 02:18:24 +00:00
nickysn
408f53fbda + detect whether CGA snow checking is required and initialize the CheckSnow variable accordingly
git-svn-id: trunk@25491 -
2013-09-16 00:01:36 +00:00
nickysn
d2c2193505 + 83/84-key keyboard support added to unit keyboard for i8086-msdos
git-svn-id: trunk@25490 -
2013-09-15 19:53:24 +00:00
nickysn
034f9c6b71 + added and implemented unit keyboard for i8086-msdos, based on the go32v2 version of the unit
git-svn-id: trunk@25489 -
2013-09-15 17:48:35 +00:00
sergei
1bd43f0ea1 * Outgoing parameters on stack do not need to inhibit stack frame optimization on targets with fixed stack, because there they do not modify stack pointer in the middle of routine.
git-svn-id: trunk@25488 -
2013-09-15 12:19:11 +00:00
nickysn
5c1fb7b009 * Crt.CursorOff fixed after r25483; the correct value of CX is 2000h, not 200h
git-svn-id: trunk@25487 -
2013-09-14 15:39:07 +00:00
nickysn
e54c87dab1 * changed a method address typecast to CodePointer instead of Pointer in TStringList.QuickSort for compatibility with the i8086 medium and compact memory models
git-svn-id: trunk@25486 -
2013-09-14 13:52:41 +00:00
nickysn
4d4ad5734b * Crt.CursorOn and CursorBig: use CX values that are CGA compatible, which makes them CGA/EGA/VGA compatible, because EGA/VGA bioses emulate CGA cursor values; as a side effect this also makes these functions work properly under EGA/VGA's 43- and 50-line text modes
git-svn-id: trunk@25485 -
2013-09-14 13:28:48 +00:00
nickysn
19d6d202cd * Crt.CursorOn and CursorBig: use CX values that are CGA compatible, which makes them CGA/EGA/VGA compatible, because EGA/VGA bioses emulate CGA cursor values; as a side effect this also makes these functions work properly under EGA/VGA's 43- and 50-line text modes
git-svn-id: trunk@25484 -
2013-09-14 13:28:40 +00:00
nickysn
e60c3ee740 * Crt.CursorOff: use CX=bash200, instead of CX= for making the cursor invisible; this is the correct value, according to the BIOS documentation I've found and it is also the value used by the go32v2 video unit
git-svn-id: trunk@25483 -
2013-09-14 12:58:38 +00:00
florian
02e054a09e + fpc -i now writes also the available inline assembler input modes
git-svn-id: trunk@25482 -
2013-09-13 19:10:59 +00:00
sergei
38c2fb755a * RTL, fpc_val_real_shortstr(): somewhat simplified, functionality is not changed.
git-svn-id: trunk@25481 -
2013-09-13 16:17:58 +00:00
svenbarth
08543ddeba Fix for Mantis #23899 . Allow to overwrite TStringList.ExchangeItems if necessary.
rtl/objpas/classes/classesh.inc, TStringList:
  * rename the private ExchangeItems to ExchangeItemsInt and allow inlining
  + add a protected virtual ExchangeItems
rtl/objpas/classes/stringl.inc, TStringList:
  * Exchange: call ExchangeItemsInt
  + let default implementation of ExchangeItems simply call ExchangeItemsInt
  * QuickSort: call ExchangeItemsInt directly if there is no override otherwise call ExchangeItems

+ added test

git-svn-id: trunk@25480 -
2013-09-13 14:17:01 +00:00
sergei
cf20bbc886 * RTL, Pos() with Variant: added 'const' modifiers to avoid unnecessary local copies of managed parameters.
git-svn-id: trunk@25479 -
2013-09-13 11:21:55 +00:00
sergei
b99cf8d680 * RTL: Changed 'shift' parameter of 64-bit shift helpers to CPU-native size, it doesn't have to be 64-bit because only its lower 6 bits are used. Not using 64 bit parameter improves code quality a bit.
git-svn-id: trunk@25478 -
2013-09-13 08:32:45 +00:00
nickysn
7e6092c4f2 * ScreenWidth and ScreenHeight changed to word; several other types also changed to 16-bit and added typecasts to avoid 32-bit calculations due to mixing signed and unsigned 16-bit ints
git-svn-id: trunk@25477 -
2013-09-13 08:31:53 +00:00
paul
60c0de6c09 compiler: formatting
git-svn-id: trunk@25476 -
2013-09-13 02:24:47 +00:00
paul
73dab82a7f compiler: refactor preproc_substitutedtoken - pass the searchstr into it
git-svn-id: trunk@25475 -
2013-09-13 01:19:19 +00:00
nickysn
70937d93c1 * WinMinX, WinMinY, WinMaxX and WinMaxY helper functions replaced with absolute packed records overlaid on top of WindMin and WindMax for faster access
git-svn-id: trunk@25474 -
2013-09-12 23:39:56 +00:00
paul
e42e9bb677 compiler: support ORD() intrinsic in preprocessor expressions
git-svn-id: trunk@25471 -
2013-09-12 12:52:07 +00:00
paul
974aeb5573 compiler: support MOD, DIV, SHR, SHL in preprocessor expressions
git-svn-id: trunk@25469 -
2013-09-12 11:22:59 +00:00
paul
cc37b19bb2 compiler: allow 1 and 0 to be used as Booleans in preprocessor expressions
git-svn-id: trunk@25467 -
2013-09-12 09:30:48 +00:00
paul
06e7808f64 compiler: return short evaluation of preprocessor expressions with AND or OR. This fixes univint package compilation.
git-svn-id: trunk@25466 -
2013-09-12 09:06:13 +00:00
paul
038b7746fb compiler: implement preprocessor expressions (fixes mantis #0010671)
- move operator_levels to topens.pas - it is used from 2 units now
  - implement pexpr like sub_expr for preprocessor expressions
  - implement +,-,*,/ expressions for the moment
  * move OR, AND, IN implemenetation to the new logic

git-svn-id: trunk@25465 -
2013-09-12 08:35:24 +00:00
paul
6eba4226b6 compiler: move OR and AND preprocessor evaluation to texprvalue.evaluate
git-svn-id: trunk@25464 -
2013-09-12 06:59:12 +00:00
paul
7ecaaf0ef8 compiler: further improve preprocessor evaluator:
* support TRUE,FALSE for all modes
  * refactor texprvalue.evaluate to support arithmetic expressions and implement them here
  * simplify read_expr and read_factor code

git-svn-id: trunk@25463 -
2013-09-12 06:01:49 +00:00
paul
cac09f9df1 compiler: better handling of '$','%','&' in macros
git-svn-id: trunk@25462 -
2013-09-12 02:33:02 +00:00
paul
0eb4244a67 compiler: implement compile-time expression with float numbers (issue #0010670), better handling of integer expressions too
git-svn-id: trunk@25461 -
2013-09-12 01:59:26 +00:00
paul
eff0403fc1 compiler: reimplement handling of compile time variables - they are stored like tconstsym now instead of string. Expressions are evaluated more accurate now.
git-svn-id: trunk@25460 -
2013-09-12 00:38:17 +00:00
svenbarth
6f5a648516 Improve the cpu type handling for M68k just in case we should branch 2.8.0 before I can start working on M68k again.
Therefor the cpu type (-Cp...) "coldfire" was split up into "isaa", "isaa+", "isab" and "isac". The Linux RTL can currently compiled for "68020", "isab" and "isac". For the other three Bcc.L must be handled differently (only Bcc.B/W supported) and for "68000" also EXT.L needs to be handled differently.

fpcdefs.inc:
  + define CPUCAPABILITIES if capabilities can be set for a certain CPU type (currently ARM, AVR and M68k)
options.pas:
  * check for CPUCAPABILITIES instead of specific CPUs
assemble.pas:
  - the handling of the CPU type is already done in m68k/ag68kgas.pas, Tm68kGNUAssembler.MakeCmdLine (and thereby already using the gascputypestr array!)
m68k/cpuinfo.pas:
  - tcputype: remove "cpu_coldfire"
  + tcputype: add "cpu_isa_a", "cpu_isa_a_p", "cpu_isa_b" and "cpu_isa_c"
  + add "cpu_coldfire" constant which contains all Coldfire specific cpu types
  * adjust "cputypestr" and  "gascputypestr"
  + add tcpuflags and cpu_capabilities (DBRA restriction was checked with CPUCOLDFIRE, CAS/TAS will be needed for atomic operations and BRAL restriction was discovered during testing of new cpu types)
m68k/cgcpu.pas:
  * adjust checks for "cpu_coldfire"
m68k/n68kadd.pas:
  * don't use a BRA.L if it is not supported, but (at least for now) a BRA.W
aggas.pas:
  * adjusted check for Coldfire

git-svn-id: trunk@25457 -
2013-09-11 17:07:32 +00:00
svenbarth
2e70575afc Allow compilation of compiler with 2.6.x RTL for simplyfied development.
cfileutl.pas:
  + define a TRawByteSearchRec if we're not compiling with at least a 2.7.1 compiler

git-svn-id: trunk@25456 -
2013-09-11 16:05:24 +00:00
nickysn
6ef912a7a0 * GetScreenCursor (and all places where it's called) var parameters changed to smallint, which is more optimal on i8086
git-svn-id: trunk@25455 -
2013-09-11 14:51:59 +00:00
nickysn
771301d7db * SetScreenCursor parameters changed to smallint, which is more optimal on i8086
git-svn-id: trunk@25454 -
2013-09-11 14:40:26 +00:00
nickysn
987e759e20 * WinMinX, WinMinY, WinMaxX and WinMaxY helper functions changed to return 0 based result, which results in slightly more optimal code in a lot of places
git-svn-id: trunk@25453 -
2013-09-11 14:18:30 +00:00
lacak
47f9c32f11 fcl-db: tests: fix TestStreamingNullFieldsXML (only fields which are not required and not readonly can be set to Null)
git-svn-id: trunk@25452 -
2013-09-11 12:05:20 +00:00
lacak
91783105cc fcl-db: tests: improve testing Blobs (use not only ascii characters, but also binary data etc.)
git-svn-id: trunk@25451 -
2013-09-11 10:48:44 +00:00
paul
385fc9cb37 compiler: remove strange nowhere used define
git-svn-id: trunk@25450 -
2013-09-11 03:06:47 +00:00
nickysn
8251031adf - rm redundant double initialization of regs.realecx from Crt.CursorOn
git-svn-id: trunk@25448 -
2013-09-10 15:28:39 +00:00
nickysn
d64cea05df - rm redundant double initialization of regs.realecx from Crt.CursorOn
git-svn-id: trunk@25447 -
2013-09-10 15:28:01 +00:00
nickysn
e7cbd41670 - rm redundant double initialization of regs.cx from Crt.CursorOn
git-svn-id: trunk@25446 -
2013-09-10 15:27:06 +00:00
nickysn
76cbda0983 + 83/84-key keyboard support added to unit crt
git-svn-id: trunk@25445 -
2013-09-09 20:55:08 +00:00
nickysn
766b7745bd * counter variable in CrtWrite changed to SmallInt in order to generate more efficient code on the 16-bit i8086
git-svn-id: trunk@25444 -
2013-09-09 12:29:13 +00:00
sergei
5de4bdac40 * tx86addnode.second_cmpfloat: use getresflags() to reduce code duplication.
git-svn-id: trunk@25443 -
2013-09-09 11:25:49 +00:00
nickysn
5ffca2e66f * CrtOpen, CrtClose, CrtRead, CrtWrite and CrtReturn changed to procedures, instead of functions
git-svn-id: trunk@25442 -
2013-09-08 21:22:40 +00:00
nickysn
4f96e8a20d * convert i8086 inline asm instruction 'call symbol' to 'call far symbol' in memory models with far code
git-svn-id: trunk@25438 -
2013-09-08 16:34:12 +00:00
nickysn
5a42a8d707 + added an i8086 test tfarcal1.pp, which tests calling a far procedure from within inline asm code
git-svn-id: trunk@25435 -
2013-09-08 13:43:52 +00:00
florian
b4481544ef * work with actualtargetnode when checking load nodes for cse
git-svn-id: trunk@25434 -
2013-09-07 17:44:44 +00:00