#18 Compile warning in 2.0.0.44

v2.0
closed-invalid
build (28)
5
2001-10-03
2001-09-08
Richard Jones
No

My glext.h has

extern void APIENTRY glMultiModeDrawArraysIBM
(GLenum, const GLint *, const GLsizei *, GLsizei,
GLint);

whereas GL/IBM/multimode_draw_arrays.i has

%name(glMultiModeDrawArraysIBM) void
_glMultiModeDrawArraysIBM(const GLenum *mode, const
GLint *first, const GLsizei *count, GLsizei
primcount);

i.e. the first arg is by-value in the glext header,
but by-ref in the swig interface file.

My glext.h has "GL_GLEXT_VERSION 7".

Discussion

  • Richard Jones
    Richard Jones
    2001-09-08

    • labels: --> build
    • milestone: --> v2.0
     
  • Richard Jones
    Richard Jones
    2001-09-08

    Logged In: YES
    user_id=6405

    I also had a lot of API mismatches for the GLE.i interface
    file when compared to src/gle/src/gle.h.

     
  • Logged In: YES
    user_id=21784

    The GLE warning are harmless warnings about double[] !=
    *double

    The multimode_draw_arrays problem has been noted before.
    The current proto matches spec. Whether the spec has a
    typo or the typo is in the header is unresolved.

     
    • assigned_to: nobody --> twburton
     
  • Richard Jones
    Richard Jones
    2001-09-08

    Logged In: YES
    user_id=6405

    My C is a bit rusty, bit isn't double* different to
    double[][3]? Normally compilers don't complain if you
    mix-n-match doule* and double[]...

     
  • Logged In: YES
    user_id=21784

    There might be some subtlety that I don't know of, but I
    just can't see what the difference could possible be.
    Anyways using double* is nesc with the current typemap
    system. Using double[][3] makes things really complex and
    since all the protos work now (try the GLE demos) I don't
    see the point in changing the current system.

     
    • status: open --> closed-invalid
     
  • Logged In: YES
    user_id=21784

    I'm closing this since after looking at the docs it appears
    that the bug is in glext.h