paweld/fpc
1
0
Fork 0
mirror of https://gitlab.com/freepascal.org/fpc/source.git synced 2025-11-25 06:57:20 +01:00
Code Issues Packages Projects Releases Wiki Activity
39,157 Commits 102 Branches 76 Tags 289 MiB
fffa2e163a
Commit Graph

14746 Commits

Author SHA1 Message Date
Jonas Maebe
fffa2e163a - removed unused local type in internal jvm record clone implementation 
git-svn-id: trunk@23526 -
 2013-01-26 21:28:26 +00:00
Jonas Maebe
dc7760bda0 * let type conversion nodes handle type checking for if/while/repeat 
conditions, so that used-defined implicit boolean type conversions are
    supported (patch by Peter-Jan Roes, mantis #23568)

git-svn-id: trunk@23525 -
 2013-01-26 21:28:22 +00:00
Jonas Maebe
2ed4b295fe * only change the visibility of an existing getter/setter to that of 
its property in case the visibility of the property is higher than
    that of the getter/setter (on the JVM target when automatically
    generating getters/setters)

git-svn-id: trunk@23523 -
 2013-01-26 16:52:32 +00:00
Jonas Maebe
e7315d035c * inherit po_auto_raised_visibility flag when the visibility of a method is 
raised because it was automatically raised in the parent class (can happen
    for the JVM target when letting the compiler generate getters/setters for
    properties)

git-svn-id: trunk@23522 -
 2013-01-26 16:52:28 +00:00
sergei
d762ef9e78 + PECOFF targets: support more than 65535 relocations per object section, resolves #23752. DJCOFF targets: give a fatal error, since this object file format does not support extended relocations. 
git-svn-id: trunk@23521 -
 2013-01-26 16:15:59 +00:00
svenbarth
7bad1763b8 Decrease amount of code duplication regarding handling of safecall and record constructors. 
paramgr.pas, tparamanager:
    + add new method "handle_common_ret_in_param"
    * extract common code (safecall + record constructor handling) from "ret_in_param" to "handle_common_ret_in_param" and call the latter in "ret_in_param"
[aarch64,arm,avr,i386,x86_64]/cpupara.pas, tCPUparamanager.ret_in_param:
    * call "handle_common_ret_in_param" instead of implementing the same check again and again
ncgcal.pas, tcgcallnode.handle_return_value:
    * move the check for (record) constructors to "tparamanager.handle_common_ret_in_param"

git-svn-id: trunk@23520 -
 2013-01-25 20:39:28 +00:00
sergei
cb1098c26c * Do not create SEH finalization procedures in generic methods, their similarity with nested procedures causes compilation errors because actual nested procedures are not allowed for generics. Not creating them doesn't hurt because generic node tree is not used for code generation. Resolves #23724. 
git-svn-id: trunk@23519 -
 2013-01-25 18:50:04 +00:00
yury
d8902af3d6 * Do not replace ADD by SUB and vice-versa when overflow checking is enabled. Fixes tw15304 for ARM. 
git-svn-id: trunk@23517 -
 2013-01-24 21:49:21 +00:00
Jonas Maebe
9cbff6f5a6 + support for record constructors for the JVM target 
git-svn-id: trunk@23510 -
 2013-01-24 09:45:12 +00:00
Jonas Maebe
3abc7b0d9c * removed superfluous check (a procvardef can never be a constructor) and 
typecasts

git-svn-id: trunk@23509 -
 2013-01-24 09:45:06 +00:00
Jonas Maebe
10e3cab1b1 * fixed formatting 
git-svn-id: trunk@23508 -
 2013-01-24 09:44:59 +00:00
Jonas Maebe
9145f1d67d * integrated potype_constructor check in maybe_create_funcret_node 
git-svn-id: trunk@23507 -
 2013-01-24 09:44:52 +00:00
sergei
40c510403f * Don't crash on a PLT-generating relocation to a local symbol (which should not normally appear, but nevertheless is possible). 
git-svn-id: trunk@23481 -
 2013-01-21 16:49:10 +00:00
paul
5d74e0578a compiler: remove unused variables and unneeded assignments 
git-svn-id: trunk@23467 -
 2013-01-21 01:24:36 +00:00
svenbarth
78f0d6f70e Revert the changes from r23383 for m68k/n68kadd.pas. Memo to self: Don't change code if you don't really understand why it does things the way it does. 
With this change and the change from r23465 we are down from ~950 to ~650 failures :D

git-svn-id: trunk@23466 -
 2013-01-20 18:00:51 +00:00
svenbarth
b455ae9534 m68k/cpupara.pas, tm68kparamanager.getintparaloc: 
* initialize paraloc^.size to OS_INT instead of leaving it at 0

This fixes tests/test/cg/ttryexc1.pp. The problem was that the raise node generated code which resulted in a word(!) move of the raised object's address to the stack location for fpc_raiseexception. This then resulted in an error when freeing the exception object.

git-svn-id: trunk@23465 -
 2013-01-20 16:33:10 +00:00
florian
9d76168b57 * refactored Bl2B condition code to make it more readable 
git-svn-id: trunk@23464 -
 2013-01-20 15:38:38 +00:00
florian
7184306a4c + option -Wx to generate thumb interworking safe code on arm 
git-svn-id: trunk@23463 -
 2013-01-20 15:26:59 +00:00
florian
3fff969ced * take care of instruction in between when doing ShiftShiftShift2ShiftShift optimizations 
git-svn-id: trunk@23462 -
 2013-01-20 14:58:03 +00:00
florian
f1b8fad6dc * make AndAnd2And working with other instructions in between 
git-svn-id: trunk@23461 -
 2013-01-20 14:57:59 +00:00
florian
fcef2dc3df * refactored some code and introduced usage of new method RegEndOfLife 
git-svn-id: trunk@23460 -
 2013-01-20 14:57:55 +00:00
florian
82f3ec7922 * if cs_opt_level3 is not set, limit lookahead of GetNextInstructionUsingReg to one instruction: 
only -O3 means do the really slow optimizations

git-svn-id: trunk@23459 -
 2013-01-20 14:57:51 +00:00
florian
a78af5b8fe + AndLslXsr2And and AndLsl2Lsl optimization 
git-svn-id: trunk@23458 -
 2013-01-20 14:57:46 +00:00
florian
ff522d7e18 * improve ShiftShiftShift2ShiftShift to look further ahead 
* check register usage so the destination register can be different

git-svn-id: trunk@23457 -
 2013-01-20 14:57:43 +00:00
florian
12d0c05ede * remove bic instructions after lsr if possible 
git-svn-id: trunk@23456 -
 2013-01-20 14:57:38 +00:00
svenbarth
632db2229f Fix for Mantis #23700 . 
scanner.pas, parse_compiler_expr.read_factor:
  * initialize result value

git-svn-id: trunk@23454 -
 2013-01-20 11:21:20 +00:00
florian
81e0e3114c * throw an error if threadvars are used but not supported 
* fix building for arm-embedded

git-svn-id: trunk@23452 -
 2013-01-19 15:17:47 +00:00
pierre
9e6a2b868c Fix -Cp option failure due to cputypestr array that must be uppercased 
git-svn-id: trunk@23446 -
 2013-01-18 17:09:28 +00:00
paul
7c663af588 compiler: also don't allow record constructors with only default arguments 
git-svn-id: trunk@23438 -
 2013-01-18 11:59:25 +00:00
paul
fe338013a5 compiler: calculate offset of record fields in taddrnode also for bitpacked structures (Jonas comment in bug #0019357) 
git-svn-id: trunk@23437 -
 2013-01-18 01:56:55 +00:00
paul
1b8369dd99 compiler: don't allow parameterless constructors 
git-svn-id: trunk@23436 -
 2013-01-18 01:24:04 +00:00
Jonas Maebe
583eb3f9a4 * re-enabled loadvmtaddrnode for records on JVM targets, fixes 
tests/test/jvm/tpvar.pp after r23417

git-svn-id: trunk@23430 -
 2013-01-17 19:45:09 +00:00
paul
65cbb4e9ba compiler: calculate offset of record fields in taddrnode (based on patch of Jeppe Johansen, fixes bug #0019357) 
git-svn-id: trunk@23428 -
 2013-01-17 14:30:55 +00:00
paul
3a3c10a474 compiler: don't allow constants, methods, class members and properties for local or anonymous records (fixes bug #0023000) 
git-svn-id: trunk@23421 -
 2013-01-17 09:05:59 +00:00
paul
4d79a44e4c compiler: don't create loadvmtaddrnode for record references, they have no VMT (fixes bug #23130) 
git-svn-id: trunk@23417 -
 2013-01-17 07:30:00 +00:00
florian
cad1309935 * don't write directly to the function result destination if we are inlining and if the destination is too complex to evaluate (threadvar etc.) 
git-svn-id: trunk@23415 -
 2013-01-16 20:25:49 +00:00
florian
247d6dd394 + implemented a tcasenode.printnodetree printing all cases 
git-svn-id: trunk@23414 -
 2013-01-16 20:24:56 +00:00
florian
abfa6c1b43 * redo LsrAnd2Lsr optimization 
git-svn-id: trunk@23413 -
 2013-01-16 20:24:07 +00:00
masta
fe520c215b New ARM Peephole optimizer FoldShiftLdrStr 
This one folds
      mov r1, r2, lsl #2
      ldr/ldrb r0, [r0, r1]
into
      ldr/ldrb r0, [r0, r2, lsl #2]

There is still some room for improvement, maybe it would be better to do this before
the register allocator runs, as we'll currently waste a register (r1 in the above example)
in many cases. That would also allow to to fold more operations, because currently if r2
gets reused between the mov and ldr we'll not be able to do the optimization.

git-svn-id: trunk@23408 -
 2013-01-16 14:37:28 +00:00
paul
e9615716c1 compiler: allow constructors in helpers for records 
git-svn-id: trunk@23407 -
 2013-01-16 13:47:22 +00:00
pierre
2b300ec2b4 Add breakpoint at Verbose.GenerateError procedure 
git-svn-id: trunk@23399 -
 2013-01-16 09:46:34 +00:00
pierre
6d8b1e03d7 Increase status.errorcount only inside GenerateError procedure, to simplify debugging 
git-svn-id: trunk@23398 -
 2013-01-16 09:45:57 +00:00
paul
b2a613c17f compiler: implement record constructors + tests 
git-svn-id: trunk@23395 -
 2013-01-16 02:07:42 +00:00
paul
51825b6f2e compiler: change ret_in_param to accept tabstractprocdef instead of tproccalloption to allow check more options (required for record constructor implementation) 
git-svn-id: trunk@23394 -
 2013-01-16 01:14:23 +00:00
sergei
b35d04ffa8 * MIPS: methodpointers must be compatible with records; since records are never passed by address in this ABI, so must be methodpointers. Fixes webtbs/tw9141.pp. 
git-svn-id: trunk@23392 -
 2013-01-15 19:09:19 +00:00
sergei
d2758265e3 * MIPS int->real conversion: When converting unsigned to single, load it into 64-bit FP register, otherwise further subtracting a 64-bit offset leads to invalid result. Fixes tw17714.pp. 
* The addend is endian-dependent, fixes tcnvint4.pp on big-endian targets.

git-svn-id: trunk@23388 -
 2013-01-15 10:49:37 +00:00
sergei
b5eecab582 * Procvars (methodpointers) must be treated as records, too. Fixes tw12318.pp and tw14155.pp after r23377. 
* Factored related code into a separate method, as it will probably need further refinement.

git-svn-id: trunk@23387 -
 2013-01-15 10:39:14 +00:00
svenbarth
c7a9e17bc5 Patch by Vasiliy Kevroletin. Fixes Mantis #23655. 
compiler/node.pas, tnode.printnodeinfo:
  * write flags to file "t" instead of stdout

git-svn-id: trunk@23384 -
 2013-01-15 06:04:10 +00:00
svenbarth
ccecf2c13c Fix comparisons (aka usage of flag/CCR register) 
m68k/aasmcpu.pas, taicpu.spilling_get_operation_type:
  * add all Sxx instructions as "operand_write" instructions

m68k/n68kadd.pas, t68kaddnode.getresflags:
  * use the correct operation in case of swapped nodes

m68k/cgcpu.pas, tcg68k.g_flags2reg:
  - don't move a 0 to the register, because this will CLR it and thus the flags won't be valid anymore...
  - NEG would have been the wrong operation (NOT would have been correct), but it isn't needed anyway...
  * simplify the method by handling the address register case only when necessary

git-svn-id: trunk@23383 -
 2013-01-14 20:31:15 +00:00
svenbarth
741992bae4 m68k/n68kmat.pas, tm68kshlshrnode.first_shlshr64bitint: 
use RTL helper functions (through the inherited method) if we're not shifting by a constant

This fixes test/cg/tshlshr

git-svn-id: trunk@23378 -
 2013-01-13 19:33:23 +00:00