Defines in swig interface files

[Umberto Nicoletti]

I think we should get rid of those defines and move them into the
mapserver files, rather than add compilation switches for all
languages. Most languages require swig to be launched manually and
don't have a Makefile.

I found a reference by Sean to this design choice here:

http://mapserver.gis.umn.edu/bugs/show_bug.cgi?id=895#c2
(end of first para)

but there should be a more explicit statement somewhere else.
Compilation switches could maybe be useful for the compilation of
libmapscript.so/dll?

Umberto

On 8/28/06, Tamas Szekeres <szekerest at gmail.com> wrote:
> In my previous post i have noted that suppressing the compilation
> swiches may result in unexpected interface creation.
> For example missing -DUSE_POINT_Z_M will prevent from generating z and m members
> for pointObj etc.
>
> I've added a bug for this issue:
> http://mapserver.gis.umn.edu/bugs/show_bug.cgi?id=1881
>
> Tamas
>
>
> 2006/8/28, Umberto Nicoletti <umberto.nicoletti at gmail.com>:
> > I have noticed that some swig interface files have #ifdef in them
> > (list follows).
> > Unfortunately they are completely useless as swig cannot pick them up
> > because the needed defines are not passed to swig in the Makefile.
> > The only exception to this rule is c# mapscript. Java mapscript used
> > to pass the defines to swig, but this was removed some time ago.
> >
> > swiginc/image.i:#if GD2_VERS > 1   Not sure about this one, maybe it
> > is right.....
> > swiginc/image.i:#ifdef USE_GD_GIF
> > swiginc/image.i:#ifdef USE_GD_PNG
> > swiginc/image.i:#ifdef USE_GD_JPEG
> > swiginc/image.i:#ifdef USE_GD_WBMP
> > swiginc/point.i:#ifdef USE_POINT_Z_M
> > swiginc/point.i:#ifdef USE_POINT_Z_M
> > swiginc/point.i:#ifdef USE_POINT_Z_M
> > swiginc/point.i:#ifdef USE_POINT_Z_M
> > swiginc/point.i:#ifdef USE_POINT_Z_M
> > swiginc/point.i:#ifdef USE_POINT_Z_M
> > swiginc/shape.i:#ifdef USE_GEOS
> > swiginc/shape.i:#ifdef USE_GEOS
> >
> > The convention which was never written and if it was I cannot find
> > where is that defines must only go into the .c and .h files to keep
> > the swig interface as simple as possible.
> > Sean Gillies can probably confirm this.
> >
> > At this point it is too late to fix this issue for 4.10. I think in
> > the meanwhile we'll should decide if we are to stick with this
> > convention and after 4.10 is out move the defines (if needed).
> >
> > I can open a bug to track this issue and maybe we should write an rfc
> > with coding conventions for mapscript developers?
> >
> > Umberto
> >
>