 masta
|
80d2ee8ad0
Don't reuse registers on ARM 64 bits shift
|
11 years ago |
 lacak
|
cdc5e9a73a
fcl-db: tests: Improve TestOldValue; tests also Blobs (which has special storage) and state of TField.OldValue after MergeChangeLog
|
11 years ago |
 sergei
|
9c1f917e3a
* a_call_ref functionality cannot be implemented efficiently at code generator level, because references need specific preparations at earlier points. Moved this support to tcgcallnode and its x86 descendants, and got rid of all ifdef's around.
|
11 years ago |
|
sergei
|
0d3f36eebf
- Remove references to global variable 'cg' from methods of tcg and some of its descendants.
|
11 years ago |
|
sergei
|
7104157a57
* Changed tabstractprocdef.address_size to return the def, which can be used to get other properties like size and alignment. Removes one of i8086 ifdef's.
|
11 years ago |
|
sergei
|
53556f529d
* gen_alloc_symtable: don't use current_procinfo.procdef, the necessary procdef is passed in 'pd' parameter.
|
11 years ago |
 marco
|
b7d94992eb
* moved unxsockh for solaris, mantis #25657
|
11 years ago |
 michael
|
7afd5f6afe
* Fixed bug ID #25091
|
11 years ago |
|
sergei
|
949907f1f3
* Don't emit "Temps allocated between..." comment in assembler listing if procedure has no temps.
|
11 years ago |
|
sergei
|
1626667374
* Object helper routines review/improvement for alignment-sensitive targets:
|
11 years ago |
 nickysn
|
f4f35ad47b
* changed ExceptTryLevel from longint to nativeint (because i8086)
|
11 years ago |
|
sergei
|
e0082262cc
* Don't generate VMT checking code twice (gen_load_vmt_register generates it already).
|
11 years ago |
|
sergei
|
a5d5ce8eb2
* TP-style object destructors do not need actual VMT in hidden parameter, a constant flag indicating need to free memory is sufficient.
|
11 years ago |
 florian
|
be17962748
* delete test file immediately after they were used, external tools might get into trouble trying deleting them due to the uncommon names
|
11 years ago |
|
marco
|
e7639e582c
* try fix #25657
|
11 years ago |
 Jonas Maebe
|
868b4f50e9
* use spilling_operation_type() instead of a big table to determine whether
|
11 years ago |
|
Jonas Maebe
|
de0ced1522
* A_INSLWI, A_INSLWI_, A_INSRWI, A_INSRWI_ also modify operand 0
|
11 years ago |
|
Jonas Maebe
|
6a4afd3a33
* changed uses of "exit" in constructors to if/then constructs, to
|
11 years ago |
|
Jonas Maebe
|
4e9c54278e
* the function result of constructors always is self, always is valid and is
|
11 years ago |
|
florian
|
e6b68d19c8
* reordered decision how arguments for inlining are handled
|
11 years ago |
 Jeppe Johansen
|
07b2982e77
Don't do ARM FoldShiftLdrStr peephole optimization if there's an offset in the reference.
|
11 years ago |
|
florian
|
2a7dcf6c7b
* formatting
|
11 years ago |
|
florian
|
e3f9982ee1
* make_not_regable is a matter of code generation so do it in pass_1
|
11 years ago |
|
florian
|
fbcfb63826
+ implement tcasenode.pass_typecheck
|
11 years ago |
|
florian
|
d71bfe4c6f
* var parameters can be put in registers if the callee is inlined
|
11 years ago |
|
florian
|
cd6f4d4469
* fix tsubscriptnode to handle left.location being an fpu or mm register as well.
|
11 years ago |
 Tomas Hajny
|
77967432ef
* always use forward slashes in path to fpmake similarly to paths to other tools in Makefiles
|
11 years ago |
|
masta
|
81517fdf70
Rerun peephole optimizers on the whole block
|
11 years ago |
|
marco
|
4fa8998541
* add SMsgDlgClose, Patch by stocki Mantis #25616
|
11 years ago |
|
masta
|
3f8549365a
Don't schedule LDRD on ARM
|
11 years ago |
