[GRASS-dev] make in g.extension (G6 vs G7)

[Martin Landa]

Hi all,

comparing g.extension's compiling approach in G6 (bash) and G7
(python). In G6 is set external GISBASE which requires to define bunch
of ARCH_ variables.

make MODULE_TOPDIR="$GISBASE" ARCH_INC="-I$GISBASE/include
-I${MYINST_DIR}/include -I$TMPDIR/dist/include"
ARCH_LIBPATH="-L$GISBASE/lib -L${MYINST_DIR}/lib -L$TMPDIR/dist/lib"
ARCH_LIBDIR="$GISBASE/lib" ARCH_DISTDIR="$TMPDIR/dist"

Notes: there are also some unsolved problems

if [ "/home/landa/grassdata/nc_spm_08/sqlite/.tmp/geo102/9408.0/dist/bin/r.area"
!= "" ] ; then GISRC=/home/landa/grassdata/nc_spm_08/sqlite/.tmp/geo102/9408.0/dist/demolocation/.grassrc64
GISBASE=~/smetiste/grass6/grass-6.4.2svn
PATH="/home/landa/grassdata/nc_spm_08/sqlite/.tmp/geo102/9408.0/dist/bin:$PATH"
LD_LIBRARY_PATH="/home/landa/grassdata/nc_spm_08/sqlite/.tmp/geo102/9408.0/dist/bin:/home/landa/smetiste/grass6/grass-6.4.2svn/lib:/home/landa/smetiste/grass6/grass-6.4.2svn/lib"
LC_ALL=C /home/landa/grassdata/nc_spm_08/sqlite/.tmp/geo102/9408.0/dist/bin/r.area
--html-description < /dev/null | grep -v '</body>\|</html>' >
r.area.tmp.html ; true ; fi
ERROR: G_getenv(): Variable LOCATION_NAME not set

and

GRASS_PERL=/usr/bin/perl VERSION_NUMBER=6.4.2svn sh
/home/landa/smetiste/grass6/grass-6.4.2svn/tools/g.html2man
/home/landa/grassdata/nc_spm_08/sqlite/.tmp/geo102/9408.0/dist/docs/html/r.area.html
/home/landa/grassdata/nc_spm_08/sqlite/.tmp/geo102/9408.0/dist/man/man1/r.area.1
1
make[2]: *** [/home/landa/grassdata/nc_spm_08/sqlite/.tmp/geo102/9408.0/dist/man/man1/r.area.1]
Error 127

In G7 no extra GISBASE is set, in the result

        ret = grass.call(['make',
                          'MODULE_TOPDIR=%s' % gisbase.replace(' ', '\ '),
                          'BIN=%s' % bin_dir,
                          'HTMLDIR=%s' % html_dir,
                          'MANDIR=%s' % man1_dir,
                          'SCRIPTDIR=%s' % script_dir],


g.extension in G7 should work without problems compared to g.extension in G6.

It will be good at least to sync these approaches.

Martin

-- 
Martin Landa <landa.martin gmail.com> * http://geo.fsv.cvut.cz/~landa