Thoughts on a mapserver-config script...

[Umberto Nicoletti]

On 6/7/07, Tamas Szekeres <szekerest at gmail.com> wrote:
> Umberto,
>
> My primary intention with that ticket was to remove the explicit
> duplication of the various libraries in the csharp Makefile.in like:
>
> PROJ_INC= @PROJ_INC@
> PROJ_LIBS= @PROJ_LIBS@
> PROJ=     @PROJ_ENABLED@
>
> that should not be needed. The actual solution requires to make some
> additional efforts to keep the csharp and the core makefile in sync.

That is also true: it is always very expensive and troublesome to sync
the java and csharp makefiles with the core.
We could look at how other OS projects are doing this.

Umberto

> I think we might have a different selection of the defines to compile
> the core mapserver related stuff than to pass the variables to the
> SWIG command line.
>
> Best regards,
>
> Tamas
>
>
>
> 2007/6/7, Umberto Nicoletti <umberto.nicoletti at gmail.com>:
> > On 6/6/07, Tamas Szekeres <szekerest at gmail.com> wrote:
> > > Umberto,
> > >
> > > Isn't that Daniel's comments at
> > > http://trac.osgeo.org/mapserver/ticket/1884 may be relevant to this
> > > demand?
> >
> > Yes they are.
> > My point is that some laguages like perl, python and ruby do not have
> > makefiles and at the moment they resort to reading the mapscriptvars
> > file, while java and csharp do have a Makefile and if mapserver-config
> > could have an option to just output the same defines that go into
> > mapscriptvars we could think of a way of unifying the whole process.
> >
> > The mapserver-config could even do something as simple as:
> >
> > head -2 < mapscriptvars | tail -1
> >
> > to get the needed swig defines. Of course if mapserver-config is to be
> > installed somewhere like /usr/local it must know where to find the
> > mapscriptvars file.
> > The alternative is to add the substitution of ALL_ENABLED like
> > suggested in #1884.
> >
> > This also raises the question if mapscriptvars is still needed and if
> > mapserver-config should replace it.
> >
> > Umberto
> >
> > >
> > > Tamas
> > >
> > >
> > > 2007/6/6, Umberto Nicoletti <umberto.nicoletti at gmail.com>:
> > > > It would also be very cool if the script could spit out only the
> > > > -DXXXX params for usage with swig.
> > > >
> > > > Umberto
> > > >
> > > > On 6/6/07, Frank Warmerdam <warmerdam at pobox.com> wrote:
> > > > > Steve Lime wrote:
> > > > > > I checked on how GDAL did this and copied that process over to mapserver in
> > > > > > the main branch. Doing a 'make' will create a mapserver-config script now
> > > > > > although it likely needs tweaking...
> > > > >
> > > > > Steve,
> > > > >
> > > > > This looks promising on a preliminary review.
> > > > >
> > > > > I notice that "mapserver-config --libs" does not include -lmap.  Shouldn't it?
> > > > > I also think the value of this is somewhat limited yet since we don't have
> > > > > a concept of installing mapserver yet.  So we can't actually point things
> > > > > at the installed location of the mapserver library (or include files).
> > > > >
> > > > > Also, I think a config switch that reports the full version string like
> > > > > the following would be nice.  Though I can't see exactly how this would
> > > > > be accomplished without some substantial changes since this is normally
> > > > > constructed based on macros in maperror.c.
> > > > >
> > > > >   OUTPUT=GIF OUTPUT=PNG OUTPUT=JPEG OUTPUT=WBMP OUTPUT=SVG SUPPORTS=PROJ
> > > > > SUPPORTS=FREETYPE SUPPORTS=WMS_SERVER SUPPORTS=WFS_SERVER SUPPORTS=WCS_SERVER
> > > > > INPUT=EPPL7 INPUT=OGR INPUT=GDAL INPUT=SHAPEFILE DEBUG=MSDEBUG
> > > > >
> > > > > Best regards,
> > > > > --
> > > > > ---------------------------------------+--------------------------------------
> > > > > I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
> > > > > light and sound - activate the windows | http://pobox.com/~warmerdam
> > > > > and watch the world go round - Rush    | President OSGeo, http://osgeo.org
> > > > >
> > > >
> > >
> >
>