[UR] Support platform dependant typedefs #705
Conversation
veselypeta
force-pushed
the
petr/704/platform-dependant-types
branch
from
a01d624
to
ab75ef2
Compare
| @@ -490,6 +495,14 @@ typedef struct ur_rect_region_t { | |||
|
|
|||
| } ur_rect_region_t; | |||
|
|
|||
| /////////////////////////////////////////////////////////////////////////////// | |||
| @@ -32,6 +32,11 @@ from templates import helper as th | |||
| // standard headers | |||
| #include <stdint.h> | |||
| #include <stddef.h> | |||
| #if defined(_WIN32) | |||
| #define NOMINMAX | |||
| #include "windows.h" | |||
There was a problem hiding this comment.
windows.h is a bit notorious for being a big nasty mess, as noted by the need for NOMINMAX. I have a feeling HANDLE is actually defined in a different header, I'm gonna poke around on a Windows machine to see if we can avoid including windows.h.
There was a problem hiding this comment.
At the very least we should #define WIN32_LEAN_AND_MEAN before including windows.h.
veselypeta
force-pushed
the
petr/704/platform-dependant-types
branch
from
6e9c995
to
7b47297
Compare
| - an ISO-C standard identifier | ||
| - a list of objects | ||
| - `if` must be by one of {`Linux`, `Windows`, `Darwin`} | ||
| - `else` must be None |
There was a problem hiding this comment.
maybe empry instead of None?
There was a problem hiding this comment.
also this doesn't say if and else should be mutually exclusive, it doesn't make sense to have when both at the same time.
veselypeta
force-pushed
the
petr/704/platform-dependant-types
branch
from
7b47297
to
24ce983
Compare
veselypeta
force-pushed
the
petr/704/platform-dependant-types
branch
from
24ce983
to
9d03fe4
Compare
|
Bindless images will implement this with |
This PR will enable platform dependent types to be defined as part of the UR specification. The supported platform types are: "Windows", "Linux" & "Darwin". This is useful to model platform specific types in UR such as a file descriptor.
TODO
typedefvalue in YaML.mdwindows.hhas to be included since we refer toHANDLECloses #704