Open ocornut opened 2 years ago
Errors preventing cimgui_internal.h to be parsed (this is before attempting to compile cimgui_internal.cpp)
[X] cimgui_internal.h(313,138): warning C4229: anachronism used: modifiers on data are ignored
#ifndef ImQsort
CIMGUI_API void cImQsort(void* base, size_t count, size_t size_of_element, int (* IMGUI_CDECL compare_func)(void const*, void const*));
#endif // #ifndef ImQsort
For now I've removed this function (the "right" fix would be to remove both IMGUI_CDECL
and const
) but it is unlikely someone would need this wrapper from C.
[ ] cimgui_internal.h(522,9): error C2061: syntax error: identifier 'ImVector'
// Helper: ImGuiTextIndex<>
// Maintain a line index for a text buffer. This is a strong candidate to be moved into the public API.
typedef struct ImGuiTextIndex_t
{
ImVector<int> LineOffsets;
int
EndOffset
} ImGuiTextIndex;
Removed the type, however the "right" fix here would be to emit a ImVector_int
instance in the structure?
[ ] error C2061: syntax error: identifier 'ImVec2ih_ImVec2ih'
Constructor not removed due to using explicit
?
CIMGUI_API explicit ImVec2ih_ImVec2ih(ImVec2ih* self, ImVec2 rhs);`
I hackily removed this however it seems to be parsing error chocking on explicit
.
Was this abandoned? If not, is there any way to test it out? Doesn't seem like headers for imgui_internal.h are generated as of now
The idea is that you’d need to run dear bindings on this specific header rather than imgui.h.
It doesn’t work yet. IMHO this is the most important feature ahead but this is being developed in spare time. Feel free to help fixing things if you can.
Extra thought: due to the nature of internals i guess the generator could be tweaked to not emit the non-Ex/Ex functions.
Extra thought: due to the nature of internals i guess the generator could be tweaked to not emit the non-Ex/Ex functions.
Had the same thought. The non-Ex/Ex split makes it right more annoying, since C# supports default arguments just fine, and the variant with implied default values won't probably even be referenced by the wrapper.
Did you meant to close this? I guess it's part of #47 ? When #47 makes progress it may make sense to enable it on CI ?
Oops, no, I didn't notice it's linked to #46 as well so when I merged that it got closed. I'll reopen until we get #47 all sorted out!
Untested but Ben mentions this is still WIP