yury
|
9a01bcf3dc
* Changed Integer to Longint in definitions of RTTI structures. It fixes typinfo for 16-bit CPUs.
|
5 years ago |
svenbarth
|
a00be912aa
* fix regression introduced with r42240: packed sets for the Integer based variants of SetToString/StringToSet need to be shifted on Big Endian systems
|
5 years ago |
svenbarth
|
7e1607e095
* correct alignment for the property table
|
6 years ago |
svenbarth
|
2dc771b7ba
* TClassData also needs alignment dummies
|
6 years ago |
svenbarth
|
2002650a23
* small corrections (grammar, casing, typo) for the comment for TCallConv
|
6 years ago |
svenbarth
|
707e05d131
* check for CPUM68k instead of M68K; the latter should only be defined when building the compiler, so I wonder how that ever worked...
|
6 years ago |
florian
|
1a67ffdd22
* fix alignment of ParamList
|
6 years ago |
florian
|
23585ded15
* fix access to ParamFlags of the tkMethod branch of type data for CPUs requiring proper alignment
|
6 years ago |
florian
|
73acb4d1c1
* fixes TInterfaceRawData to match TTypeData
|
6 years ago |
nickysn
|
a3a39e1640
* i8086 compilation fix after changeset 42390
|
6 years ago |
florian
|
217ae6e4bb
* more fixes to rtti after attribute branch merging
|
6 years ago |
svenbarth
|
71fa4d1fe2
* the RTTI of an attribute not only consists of the creation function, but also of the attribute's type, a pointer to the constructor and a data blob containing the constant parameters
|
6 years ago |
svenbarth
|
8cd1f1973b
- remove unneeded local variable
|
6 years ago |
svenbarth
|
affa9f9d9c
* to be on the safe side: don't include AttributeTable fields when cycling
|
6 years ago |
svenbarth
|
3ad24c9db8
* as attributes can be part of any type they are best suited in a common part of TTypeData
|
6 years ago |
svenbarth
|
de15025a47
* streamline retrieval of attributes (both in TypInfo and RTTI units)
|
6 years ago |
svenbarth
|
e060a43d70
* rename TAttributeData to the more appropriate TAttributeTable
|
6 years ago |
svenbarth
|
c17dafc2f9
* fix TClassData by including the AttributeTable at the right location
|
6 years ago |
svenbarth
|
e296b26e9e
- remove unit info and related code again as that will be handled similar, but differently with dynamic packages
|
6 years ago |
svenbarth
|
ee940c8270
* always write a class' reference to the attribute table, even if the unit does not have attributes
|
6 years ago |
svenbarth
|
788d976f3a
* use a Word instead of a Byte for the attribute index (after all the compiler writes a Word to the count field)
|
6 years ago |
svenbarth
|
5153a288dc
* it's more correct for GetPropAttribute to return a TCustomAttribute instead of a TObject
|
6 years ago |
svenbarth
|
d137e06ade
* instead of embedding the attributes inside the class' and property's type data only store a reference to a table of attributes
|
6 years ago |
svenbarth
|
84848df18d
Reintegration of Custom Attributes branch by Joost van der Sluis with patches reworked by Svetozar Belic [patch 3/3]
|
6 years ago |
svenbarth
|
b2932393df
Reintegration of Custom Attributes branch by Joost van der Sluis with patches reworked by Svetozar Belic [patch 1/3]
|
6 years ago |
svenbarth
|
cc5549f3e7
* change Value argument of SetToString and result of of StringToSet from Integer to LongInt so that small sets are handled on 8/16-bit platforms as they are on 32- and 64-bit
|
6 years ago |
svenbarth
|
15d0deff81
* fix SetToString/StringToSet for 8/16-bit platforms
|
6 years ago |
svenbarth
|
89e454aca8
* rework/extend SetToString/StringToSet so that sets with a size > 4 can be converted as well (this is Delphi compatible)
|
6 years ago |
svenbarth
|
a076677336
* instead of a flags field generate a table with all Initialize operators of all fields (nested or not, parent class or not) for class types; objects and records don't use this (yet)
|
7 years ago |
svenbarth
|
2e1327b6b2
* use pooled constant references for the parameter locations as more often than not they're equal between multiple methdos (e.g. the Self argument is always the same)
|
7 years ago |