[CORE][CMAKE] Fix Linux compilation of WWLib #698
Conversation
feliwir
added
Platform
feliwir
force-pushed
the
linux-wwlib
branch
3 times, most recently
from
ee80f88 to
0b7ef34
Compare
| unsigned buffer_base=reinterpret_cast<unsigned>(m_Buffer-sizeof (StringClass::_HEADER)); | ||
| unsigned temp_base=reinterpret_cast<unsigned>(m_TempStrings+MAX_TEMP_BYTES*MAX_TEMP_STRING); | ||
| unsigned long buffer_base=reinterpret_cast<unsigned long>(m_Buffer-sizeof (StringClass::_HEADER)); | ||
| unsigned long temp_base=reinterpret_cast<unsigned long>(m_TempStrings+MAX_TEMP_BYTES*MAX_TEMP_STRING); |
There was a problem hiding this comment.
Is this meant to take the size of pointer? long is 4 bytes on Windows x64. See:
https://en.cppreference.com/w/cpp/language/types
https://en.cppreference.com/w/cpp/language/types#Data_models
There was a problem hiding this comment.
I know, we have no integer type for pointer size yet. Will add one
There was a problem hiding this comment.
All custom integer types are defined in BaseTypeCore.h, however wwstring.cppand all of WWLib does not make use of BaseType.h
What shall we do here?
There was a problem hiding this comment.
Perhaps replicate the c++11 stdint types in a "Dependencies/Utility/stdint_adapter.h"
Then include that adapter in BaseTypeCore.h and always.h
There was a problem hiding this comment.
I think i can do that in a separate PR? I didn't break anything here and this works fine for Linux. It didn't work for windows 64-bit before
| @@ -38,11 +38,25 @@ | |||
| #ifndef _SYSTIMER_H | |||
|
|
|||
| #include "always.h" | |||
| #ifdef _WIN32 | |||
| #else | ||
| #define DEFINE_AUTO_POOL(T,BLOCKSIZE) \ | ||
| template<>\ | ||
| ObjectPoolClass<T,BLOCKSIZE> AutoPoolClass<T,BLOCKSIZE>::Allocator = {} |
There was a problem hiding this comment.
Is this separation really necessary? Cannot just add template<> for all compilers?
There was a problem hiding this comment.
No template<> doesn't work when not using = {} as the initializer here. VC6 doesn't support that
There was a problem hiding this comment.
What is the compile error that this causes? Something about dependent name? It is odd that this requires a initializer because ObjectPoolClass does have a default constructor.
There was a problem hiding this comment.
If i remember it was: missing implementation .....
Co-authored-by: Patrick Zacharias <[email protected]>
Partially implements #522
This PR does the following things to fix Linux compilation:
_UNIXmacro for non windows platformsosdep.h. We replaced this with out compat library