[CC'd to development list]
marcos boullón magán wrote:
Now I think that the Makefile in v.sample dir (vector
processing sample) is not too useful because it relies in other
Makefiles in the source code tree.
We should really be installing the include/Make/*.make files. Writing
portable Makefiles for add-on modules is likely to be a difficult task
without them (particularly Platform.make and Grass.make, which are
generated by the configure script).
--
Glynn Clements <glynn@gclements.plus.com>
On Fri, 2006-03-24 at 16:02 +0000, Glynn Clements wrote:
We should really be installing the include/Make/*.make files. Writing
portable Makefiles for add-on modules is likely to be a difficult task
without them (particularly Platform.make and Grass.make, which are
generated by the configure script).
IMHO, only files with parameters that cannot go into 'grass.pc' should
be exported.
--
Brad Douglas <rez@touchofmadness.com>
Brad Douglas wrote:
> We should really be installing the include/Make/*.make files. Writing
> portable Makefiles for add-on modules is likely to be a difficult task
> without them (particularly Platform.make and Grass.make, which are
> generated by the configure script).
IMHO, only files with parameters that cannot go into 'grass.pc' should
be exported.
The grass.pc file is of little use; apart from linking in too many
libraries, it won't work with static libraries or if you need
additional -L switches to locate the dependencies.
More generally, the idea of treating GRASS as a monolithic package is
fundamentally flawed. Realistically, you would need a separate .pc
file for each library or group of related libraries.
Also, it should be possible to write Makefiles which work either as
part of GRASS or as standalone modules.
--
Glynn Clements <glynn@gclements.plus.com>