Hi,
adding VASKLIB to IMAGERYLIB breaks compilation of some modules
(r.in.gdal) on windows or --without-curses. VASKLIB is not used by in modules
which use IMAGERYLIB, IMO VASKLIB should only be used in modules Makefiles
which realy use it.
http://freegis.org/cgi-bin/viewcvs.cgi/grass6/include/Make/Grass.make.in.diff?r1=1.60&r2=1.61
Radim
On Fri, 2006-10-06 at 14:25 +0200, Radim Blazek wrote:
Hi,
adding VASKLIB to IMAGERYLIB breaks compilation of some modules
(r.in.gdal) on windows or --without-curses. VASKLIB is not used by in modules
which use IMAGERYLIB, IMO VASKLIB should only be used in modules Makefiles
which realy use it.
http://freegis.org/cgi-bin/viewcvs.cgi/grass6/include/Make/Grass.make.in.diff?r1=1.60&r2=1.61
We should be moving away from VASKLIB, too... Unfortunately, some parts
of IMAGERYLIB still require VASKLIB functions, IIRC.
--
Brad Douglas <rez touchofmadness com> KB8UYR
Address: 37.493,-121.924 / WGS84 National Map Corps #TNMC-3785
On 10/6/06, Brad Douglas <rez@touchofmadness.com> wrote:
On Fri, 2006-10-06 at 14:25 +0200, Radim Blazek wrote:
> Hi,
> adding VASKLIB to IMAGERYLIB breaks compilation of some modules
> (r.in.gdal) on windows or --without-curses. VASKLIB is not used by in modules
> which use IMAGERYLIB, IMO VASKLIB should only be used in modules Makefiles
> which realy use it.
> http://freegis.org/cgi-bin/viewcvs.cgi/grass6/include/Make/Grass.make.in.diff?r1=1.60&r2=1.61
We should be moving away from VASKLIB, too... Unfortunately, some parts
of IMAGERYLIB still require VASKLIB functions, IIRC.
The functions which require VASKLIB are compiled only if HAVE_CURSES_H.
Radim
--
Brad Douglas <rez touchofmadness com> KB8UYR
Address: 37.493,-121.924 / WGS84 National Map Corps #TNMC-3785
Radim Blazek wrote:
adding VASKLIB to IMAGERYLIB breaks compilation of some modules
(r.in.gdal) on windows or --without-curses. VASKLIB is not used by in modules
which use IMAGERYLIB, IMO VASKLIB should only be used in modules Makefiles
which realy use it.
That approach fails if IMAGERYLIB is a shared library.
The most robust solution would be to move the parts of IMAGERYLIB
which depend upon VASKLIB into a separate library.
--
Glynn Clements <glynn@gclements.plus.com>