Jonas Maebe
|
04d295f186
* first step towards supporting 32 bit targets with the LLVM code generator:
|
6 years ago |
Jonas Maebe
|
ff8ccb06e5
* changed reverseparameters() from function into procedure: it does not make
|
6 years ago |
Jonas Maebe
|
3ccc3e329b
* prevent calling firstpass multiple times for a node (such as when
|
6 years ago |
Jonas Maebe
|
122d0d36d6
+ volatile() expression that marks an expression as volatile
|
6 years ago |
pierre
|
796db4cef5
Implement mark_write override for tinilinenode
|
6 years ago |
yury
|
4682ac269d
* Further improvement for r40180:
|
6 years ago |
yury
|
4357caaad8
* Removed unused local vars.
|
6 years ago |
yury
|
1b151df89f
* Make SetLength() to emit hint "Local variable "s" of a managed type does not seem to be initialized" instead of the warning. This change is needed because when calling a regular procedure with a ver parameter of a managed type the hint is generated. Now SetLength() behaves the same.
|
6 years ago |
pierre
|
d512d8cae1
Also fix first_abs_real like in rev 39988
|
6 years ago |
florian
|
3078cc4b0d
* same fix as in r39988 for fpc_first_sqrt
|
6 years ago |
pierre
|
7d7f966ccb
Fix wrong explicit typecast that lead to memory corruption in first_sqr_real
|
6 years ago |
Jonas Maebe
|
0b246f3dbd
* converted Boolean8 to an internal type, and mapped Boolean to the
|
6 years ago |
svenbarth
|
d5f0d8e306
* addendum for Mantis #34333: also complain about number of arguments with the maximum number of arguments for Copy ( = 3 ) is exceeded
|
6 years ago |
svenbarth
|
76c45b1a72
* fix for Mantis #34333: improve error output for incorrect calls to Copy()
|
6 years ago |
florian
|
a34d4e715c
Merged riscv_new branch
|
6 years ago |
svenbarth
|
bf91991bae
* fix for Mantis #34332: allow 2 parameter form of Copy also for ShortString variables
|
6 years ago |
florian
|
4f052e4f90
o fix several issues with floating point exceptions
|
7 years ago |
svenbarth
|
3d740f13e4
* ensure that the data of the temp node is freed correctly in case of an error (the error might be non-fatal in case of a generic passing an untyped type to Write/Read)
|
7 years ago |
svenbarth
|
781303c0d8
* ensure that the file parameter for the Read/Write family of intrinsics is (and can be) freed in tinlinenode.handle_read_write()
|
7 years ago |
florian
|
9af9658fa7
* first parameter of SetLength must be valid, as it is read
|
7 years ago |
svenbarth
|
9b45f58c0b
+ add flags that allow checking for overloads with isbinaryoverloaded() and isunaryoverloaded() without modifying the passed in node or even checking for normally non-overloadable operators
|
7 years ago |
svenbarth
|
0e5a598b6a
* don't use a vecnode to convert a dynamic array or an array constructor to a pointer as that will lead to problems with range checking if the array should be empty; fixes run of tarray12 with -Cr
|
7 years ago |
svenbarth
|
916ff0b92c
+ extend Concat() with support for dynamic arrays
|
7 years ago |
florian
|
194c0c7073
* compilation fixed
|
7 years ago |
florian
|
4752b0ef96
+ tinlinenode calls cpu specific routines for unknown inline numbers
|
7 years ago |
svenbarth
|
f6a867ef04
* move handling of Concat to tinlinenode so that it can be easily extended for dynamic arrays
|
7 years ago |
svenbarth
|
b3ee9339b8
* also report the dynamic array "overloads" for the Insert() intrinsics
|
7 years ago |
svenbarth
|
c5b33f51f9
* fix for Mantis #32412: correctly handle an incorrect parameter count for Delete() and Insert() intrinsics
|
7 years ago |
svenbarth
|
0b02dab684
+ new Delphi-compatible intrinsic GetTypeKind() which returns the TTypeKind of a type as a constant value (and thus can be optimized away in If- and Case-statements)
|
8 years ago |
svenbarth
|
29d9beb40c
* initialize datatemp
|
8 years ago |