 florian
|
12adb47d65
* clean up
|
4 years ago |
 michael
|
7ee4954b0b
* Refactor HTML engine so it can more easily be extended
|
4 years ago |
|
michael
|
7be3d2f80c
* get rid of some warnings
|
4 years ago |
 svenbarth
|
b8578f804f
* fix compilation on targets that use Stabs debug information: ignore generic dummy symbols with no procdefs for generating stabs data
|
4 years ago |
|
svenbarth
|
4a274cd63a
+ add a copyright/license header for the nullable unit
|
4 years ago |
|
michael
|
1afca037c6
* Help for markdown
|
4 years ago |
 Mattias Gaertner
|
9825d3b552
pastojs: filer export nameexpr
|
4 years ago |
|
Mattias Gaertner
|
e911431ed4
fcl-passrc: parser: export unit.symbol, resolver: started library export
|
4 years ago |
|
svenbarth
|
c96029ebd5
* Delphi does not allow a generic method to be overloaded by a non generic type of the same name (unlike for generic types and non generic routines); this is probably done to simplify the implementation of implicit specializations of generic methods so we do this as well. For this we change the dummy symbol for generic routines from a typesym to a procsym
|
4 years ago |
|
svenbarth
|
7343e9c4a2
* correctly handle the case should the generic dummy symbol be a procsym instead of a typesym
|
4 years ago |
|
svenbarth
|
e4eed4e259
* when only a symbol name is provided to generate_specialization_phase1 allow a symbol table to be provided as well
|
4 years ago |
|
svenbarth
|
a65d778f73
* also handle call nodes when determining the generic symbol for inline specializations
|
4 years ago |
|
svenbarth
|
f42b39a7b5
* use a case statement instead of nested if-then statements
|
4 years ago |
|
svenbarth
|
43ba5b69d2
* ensure that the dummy symbol is registered if it's added due to a routine
|
4 years ago |
|
svenbarth
|
9179f9d43b
* always add a generic dummy if it is a procsym
|
4 years ago |
|
svenbarth
|
d854d18bd9
* instead of blindly consuming whatever comes next trigger an explicit error if the parsed expression does not match for "INDEX ordexpr" or "NAME strexpr"
|
4 years ago |
|
florian
|
3109005791
+ fold also (string const+(string const+string var))
|
4 years ago |
|
michael
|
b387dd8a94
* Fix AV
|
4 years ago |
 marco
|
bb8fd96eaa
* fix destructor
|
4 years ago |
|
florian
|
9430436a24
+ include package/hash in regression tests
|
4 years ago |
|
michael
|
33d7935b93
* Fix class tree & lists
|
4 years ago |
|
florian
|
7de94e7a1d
* properly fold string+string const+string const., resolves #38267
|
4 years ago |
|
florian
|
8b08079224
* fpc_AnsiStr_Concat: if one of the strings is empty and CP_NONE is passed, the code page
|
4 years ago |
|
michael
|
390be00327
* Markdown support for fpdoc
|
4 years ago |
|
Mattias Gaertner
|
f0122bd4a0
fcl-passrc: resolver: parse library
|
4 years ago |
|
florian
|
e6037961e3
* refactored constant folding code
|
4 years ago |
|
florian
|
97abf6b495
* first part of fixing #38267: do not bail out early during constant folding
|
4 years ago |
|
marco
|
8311837468
--no-offset option for chmls, makes it easier to compare listings.
|
4 years ago |
|
florian
|
b4a6c22234
* do not check inlined exit nodes for unset results, resolves #38259
|
4 years ago |
|
florian
|
3482121831
* patch by Christo Crause: Fix missed optimization opportunities, resolves #38285
|
4 years ago |
