 Jonas Maebe
|
851cb65021
* activate code to remove unnecessary widenings to 64 bit in case
|
14 years ago |
|
Jonas Maebe
|
d48fa1d7e2
* for all managed VM targets, reject ord->ptr typecasts (except for 0),
|
14 years ago |
|
Jonas Maebe
|
979f55e1db
+ support for procedural variables for the JVM target
|
14 years ago |
|
Jonas Maebe
|
569228447d
* converted all enum handling for the JVM target so that it uses the
|
14 years ago |
|
Jonas Maebe
|
223057f1a7
+ support for is/as and Java interfaces
|
14 years ago |
|
Jonas Maebe
|
0706cb5eb6
+ support for pointers to types that are implicit pointer types in the JVM
|
14 years ago |
|
Jonas Maebe
|
6857dde33e
+ shortstring support for the JVM target (including accessing character 0 as
|
14 years ago |
|
Jonas Maebe
|
446d91eaab
+ ansistring support. Items of note:
|
14 years ago |
|
Jonas Maebe
|
aa1f299a17
+ added hook for "individual value to open array of one element"-conversion,
|
14 years ago |
|
Jonas Maebe
|
60f8254d63
* don't internalerror when converting to an open array and passing as
|
14 years ago |
|
Jonas Maebe
|
10a7532968
* differentiate between typecasts on the assignment side or not, because we
|
14 years ago |
|
Jonas Maebe
|
c75246706d
+ stubbed ansistring support (using ansistrings compiles, but does not
|
14 years ago |
|
Jonas Maebe
|
be92f47cc7
* don't simply replace the resultdef of typeconverted nodes if the new
|
14 years ago |
|
Jonas Maebe
|
d79160a6af
* correctly handle explicit typecasts between records and jlobject/
|
14 years ago |
|
Jonas Maebe
|
b4ffcaf88e
* redirect the ttypeconvnode.typecheck* methods just like the firstpass
|
14 years ago |
|
Jonas Maebe
|
fde622e050
* generate a loadvmtaddrnode wrapper around Java objectdef typenode arguments
|
14 years ago |
|
Jonas Maebe
|
8643ab7905
+ hooks for target-specific handling of explicit type conversions and
|
14 years ago |
|
Jonas Maebe
|
48695a7705
* when removing unnecessary 64 bit widenings, modify both the resultdef
|
14 years ago |
 florian
|
d35d1ed357
+ initial support for pascal booleans with sizes 2, 4 and 8
|
14 years ago |
 marco
|
90c47c5c86
* fix for renaming of type_interface_noguid (or so) to type_E_interface...
|
14 years ago |
 tom_at_work
|
1f023f4b36
* Use new qword_to_float32/64 compilerproc functions in first_int_to_real
|
14 years ago |
 paul
|
901a9a974e
compiler: don't create set constants from arrays with high bound > 255 if elements of array are lower than 255
|
14 years ago |
|
florian
|
58fcf9dc52
* probably mark an overloaded := operator used as type conversion as used, resolves #18909
|
14 years ago |
|
Jonas Maebe
|
eab29db98a
* merged tasnode.pass_typecheck() and tisnode.pass_typecheck() into a single
|
14 years ago |
|
Jonas Maebe
|
d76ddcabe6
* don't give an internalerror when trying to explicitly apply an invalid
|
14 years ago |
 sergei
|
331a72c8d6
* ncnv.pas, insert_varargstypeconv(): Do not cast a string literal to AnsiString if it has been already casted to Wide/UnicodeString, resolves #18266.
|
14 years ago |
|
florian
|
eb3644140a
* patches from Aleksa Todorovic for #15480 to improve error reporting, does not resolve #15480
|
14 years ago |
|
Jonas Maebe
|
14b95b3b9b
* always force range checking for the upper and lower bounds of for-loops if
|
15 years ago |
|
Jonas Maebe
|
769eb3f604
* moved the initialisation of the nodetree and tai class reference constants
|
15 years ago |
|
Jonas Maebe
|
94d976bc87
* when simplifying ordinal expressions during inlining, keep the resultdef
|
15 years ago |
