 Jonas Maebe
|
eb71a30933
* fix false possitive tconstexprint range errors after r41441 (mantis #35144)
|
6 years ago |
|
Jonas Maebe
|
07bd4ba517
* let all the case code generation work with tconstexprint instead of aint,
|
6 years ago |
|
Jonas Maebe
|
04d295f186
* first step towards supporting 32 bit targets with the LLVM code generator:
|
6 years ago |
 florian
|
644c47e121
* when deciding about the case code generation in case of size optimization, the number of labels is relevant, not the covered range, resolves #34818
|
6 years ago |
|
florian
|
a2cee252e8
* when optimizing range checks in case nodes, unsigned comparisions must be used, resolves #34782
|
6 years ago |
|
florian
|
bd4b7a6bc4
* more case fixes
|
6 years ago |
|
florian
|
f6e32ae90e
* (modified) path by Gareth Moreton to fix case handling
|
6 years ago |
|
florian
|
7f5c2fa3aa
* (modified) patch by Gareth Moreton: Speed improvement in case blocks, resolves #0034762
|
6 years ago |
 pierre
|
b64e9700aa
Avoid range errors or overflows on for AVR cpu, when computing address offsets
|
6 years ago |
|
florian
|
68eb921d46
* broken compilation after r39347 fixed
|
7 years ago |
|
florian
|
f08d3fdf8f
* moved execution weight calculation into a separate pass, so the info is available already available before the code generation pass if needed
|
7 years ago |
|
florian
|
29230ebdca
+ added comment why the code path is not yet enabled for 16 bit
|
7 years ago |
|
florian
|
f37bad774c
* do not use 32 bit operations on 8 bit cpu in tcginnode.pass_generate_code if not needed, resolves #32071
|
7 years ago |
|
florian
|
cc44328109
* correctly calc case label distance after r36362, resolves #32115 and #32311
|
7 years ago |
 nickysn
|
ddba821561
* GetNextReg(), used by 16-bit and 8-bit code generators (i8086 and avr) moved
|
7 years ago |
|
florian
|
991d962ae2
* more range check errors on 8 bit CPUs fixed
|
8 years ago |
|
florian
|
b1dff29cbf
* removed unused units
|
8 years ago |
 yury
|
95094e9a8f
* Removed unused vars.
|
8 years ago |
|
florian
|
50dba9ad66
* if both labels of a case label are equal, we still cannot skip the comparison in a jump tree, resolves #31589
|
8 years ago |
|
florian
|
74423b88e0
* comment on border between linear list and jmp tree for case statements extended
|
8 years ago |
 svenbarth
|
abc4cf21d1
* fix compiling the 64-bit compiler with a 32-bit one
|
8 years ago |
|
florian
|
1ae90bce53
* commented how the limit for a case jmp tree had been determined
|
8 years ago |
|
florian
|
d5754cf47c
* create jmp trees for really big case statements
|
8 years ago |
|
Jonas Maebe
|
b156744b62
* left has been forced into a register of size uopdef a few statements
|
9 years ago |
|
Jonas Maebe
|
da696057ab
* converted register_maybe_adjust_setbase() to the high level code generator
|
9 years ago |
|
Jonas Maebe
|
2ad8bc3b27
* fixed operation size of comparison with maximum legal value in
|
9 years ago |
|
Jonas Maebe
|
c8a5994e2e
* fixed size used for left in in_smallset (it's been forced in a register
|
9 years ago |
|
Jonas Maebe
|
09c8c5eba0
* adjusted comment after previous commit
|
9 years ago |
|
Jonas Maebe
|
a32605bd38
* don't force setelementn's into a 32 bit modifiable register if they were
|
9 years ago |
|
Jonas Maebe
|
175f4376bc
* give an internalerror if a setelementn is range in pass_generate_code,
|
9 years ago |
