[geos-devel] geos-config

Ferdinando Villa ferdinando.villa at uvm.edu
Fri Jun 25 10:48:57 EDT 2004


Ugh... ok that's not too bad, but would it be terribly messy to put all
headers under a geos/ directory and include all of them with the same
notation everywhere? The build process for postgis etc wouldn't change
as long as geos-config is used... the BUILDING_GEOS conditional sounds a
bit perverse to me... and I think we really need to prefix header files
to avoid ambiguities. If folks think we should do it this way, I can
volunteer to code that in if it's too much work for one person.

On Fri, 2004-06-25 at 10:47, strk wrote:
> On Fri, Jun 25, 2004 at 10:31:18AM -0400, Ferdinando Villa wrote:
> > That's debatable, but my vote would go to cflags pointing to /includes,
> > and prefixing all includes as geos/. I don't see
> > a problem in having geo{s|m}.h containing lines like
> > 
> > #ifndef __geos_platform_h
> > #include <geos/platform.h>
> > #endif
> >  
> > using the bracket notation should keep us safe from ambiguity and it
> > would work as long as the include path is given as /includes. Looks
> > elegant to me, too, but that's me...
> 
> Note that geom.h is included by GEOS modules themself, whose headers
> do not currently reside under a geos/ directory, so the switch should
> be something like
> 
> #ifdef __BUILDING_GEOS
> #include "platform.h"
> #else
> #include <geos/platform.h>
> #endif
> 
> The switch would not be needed for geos.h since it is a front-end only
> header file....
> 
> --strk;
> 
> 
> > 
> > cheers
> > f
> > 
> > On Fri, 2004-06-25 at 10:29, strk wrote:
> > > On Fri, Jun 25, 2004 at 09:16:08AM -0400, Ferdinando Villa wrote:
> > > > Yes - --cflags should produce the -I line, --includes should just list
> > > > the path. The current script is buggy - strk can you fix it? 
> > > 
> > > The fix is in current CVS.
> > > Ferdinando, do you think includes should be
> > > .../includes or .../includes/geos ?
> > > I'd like application to use #include "geos/geom.h" or
> > > similar (which would work with current geos-config output)
> > > but then geom.h includes platform.h (and not geos/platform.h)
> > > and it should not work then....
> > > What do you think should be done about this ?
> > > 
> > > --strk;
> > > 
> > > > 
> > > > The procedure to use the geos.m4 macro is as follows:
> > > > 
> > > > put GEOS_INIT(1.0.0) in configure.int
> > > > 
> > > > run aclocal -I macros instead of aclocal 
> > > > the rest is the same
> > > > 
> > > > your CXXFLAGS will include the geos path automatically and your LIBS
> > > > will link geos in.
> > > > 
> > > > ferdinando
> > > > 
> > > > On Fri, 2004-06-25 at 04:40, strk wrote:
> > > > > I don't like this also.
> > > > > I think it should produce just the path without the -I switch.
> > > > > Anyway, you should find a geos.m4 file in the GEOS distribution
> > > > > handy for application's configure scripts.
> > > > > 
> > > > > --strk;
> > > > > 
> > > > > On Thu, Jun 24, 2004 at 11:38:23PM -0500, Stephen Lime wrote:
> > > > > > Hi folks: Trying to integrate geos into a configure script based on the 
> > > > > > output from geos-config. I'm curious if the --includes and --cflags 
> > > > > > options should be producing the same output (they do). I expected one 
> > > > > > of them to produce "-I/usr/local/include/geos". I can work around this 
> > > > > > using the --prefix but was wondering if the behavior is as intended.
> > > > > > 
> > > > > > Steve
> > > > > > 
> > > > > > _______________________________________________
> > > > > > geos-devel mailing list
> > > > > > geos-devel at geos.refractions.net
> > > > > > http://geos.refractions.net/mailman/listinfo/geos-devel
> > > > > _______________________________________________
> > > > > geos-devel mailing list
> > > > > geos-devel at geos.refractions.net
> > > > > http://geos.refractions.net/mailman/listinfo/geos-devel
> > > > -- 
> > > > 
> > > > _______________________________________________
> > > > geos-devel mailing list
> > > > geos-devel at geos.refractions.net
> > > > http://geos.refractions.net/mailman/listinfo/geos-devel
> > > _______________________________________________
> > > geos-devel mailing list
> > > geos-devel at geos.refractions.net
> > > http://geos.refractions.net/mailman/listinfo/geos-devel
> > -- 
> > 
> > _______________________________________________
> > geos-devel mailing list
> > geos-devel at geos.refractions.net
> > http://geos.refractions.net/mailman/listinfo/geos-devel
-- 




More information about the geos-devel mailing list