[On GRASSLIST] H. Bowman wrote:
Grass5.0.0pre4 is out apparently.
Kudos to all. Time to beat on the new shape file importers.. 
The configure summary is a nice touch.
Configured it (with everything but NLS support), compiled it. No errors.
Upon 'make install', it hums though until it gets to this:
test -d /usr/src/grass5source/grass5.0.0pre4/dist.i686-pc-linux-gnu/locale
&&
/bin/sh -c "cd
/usr/src/grass5source/grass5.0.0pre4/dist.i686-pc-linux-gnu ;
tar cBf - locale | (cd /usr/local/grass5 ; tar xBf - ) 2>/dev/null ;
true"
make[1]: *** [real-install] Error 1
make[1]: Leaving directory `/usr/src/grass5source/grass5.0.0pre4'
make: *** [install] Error 2
Trouble is, there is no ./dist.X/locale directory.
We need to release an update. This bug is probably going to bite
*most* users, as NLS is disabled by default. The net result is that
the following section of the "install" target won't get executed:
${SHELL} -c "cd ${GISBASE} ; tar cBf - man | (cd ${INST_DIR} ; tar xBf - ) 2>/dev/null ; true"
${SHELL} -c "cd ${GISBASE} ; tar cBf - include | (cd ${INST_DIR} ; tar xBf - ) 2>/dev/null ; true"
${SHELL} -c "cd ${GISBASE} ; tar cBf - lib | (cd ${INST_DIR} ; tar xBf - ) 2>/dev/null ; true"
${SHELL} -c "$(INSTALL) ${GISBASE}/dev/create_fifos.sh ${INST_DIR}/dev ; true"
if [ ${XDRIVER} = fifo ] ; then \
${SHELL} -c "${SHELL} ${SRCDIR}/src/scripts/shells/create_fifos.sh ${INST_DIR} 2>/dev/null ; true"; \
fi
${SHELL} -c "sed 's#'${GISBASE}'#'${INST_DIR}'#g' ${GISBASE}/etc/monitorcap > ${INST_DIR}/etc/monitorcap ; true"
${SHELL} -c "chmod -R 1777 ${INST_DIR}/locks 2>/dev/null ; true"
${SHELL} -c "chmod -R a+rX ${INST_DIR} 2>/dev/null ; true"
The bit regarding monitorcap and the final chmod could be critical.
--
Glynn Clements <glynn.clements@virgin.net>
On Tue, May 21, 2002 at 05:33:27PM +0100, Glynn Clements wrote:
[On GRASSLIST] H. Bowman wrote:
> Grass5.0.0pre4 is out apparently.
>
> Kudos to all. Time to beat on the new shape file importers.. 
> The configure summary is a nice touch.
>
>
> Configured it (with everything but NLS support), compiled it. No errors.
> Upon 'make install', it hums though until it gets to this:
>
> test -d /usr/src/grass5source/grass5.0.0pre4/dist.i686-pc-linux-gnu/locale
> &&
> /bin/sh -c "cd
> /usr/src/grass5source/grass5.0.0pre4/dist.i686-pc-linux-gnu ;
> tar cBf - locale | (cd /usr/local/grass5 ; tar xBf - ) 2>/dev/null ;
> true"
> make[1]: *** [real-install] Error 1
> make[1]: Leaving directory `/usr/src/grass5source/grass5.0.0pre4'
> make: *** [install] Error 2
>
>
> Trouble is, there is no ./dist.X/locale directory.
We need to release an update. This bug is probably going to bite
*most* users, as NLS is disabled by default. The net result is that
the following section of the "install" target won't get executed:
${SHELL} -c "cd ${GISBASE} ; tar cBf - man | (cd ${INST_DIR} ; tar xBf - ) 2>/dev/null ; true"
${SHELL} -c "cd ${GISBASE} ; tar cBf - include | (cd ${INST_DIR} ; tar xBf - ) 2>/dev/null ; true"
${SHELL} -c "cd ${GISBASE} ; tar cBf - lib | (cd ${INST_DIR} ; tar xBf - ) 2>/dev/null ; true"
${SHELL} -c "$(INSTALL) ${GISBASE}/dev/create_fifos.sh ${INST_DIR}/dev ; true"
if [ ${XDRIVER} = fifo ] ; then \
${SHELL} -c "${SHELL} ${SRCDIR}/src/scripts/shells/create_fifos.sh ${INST_DIR} 2>/dev/null ; true"; \
fi
${SHELL} -c "sed 's#'${GISBASE}'#'${INST_DIR}'#g' ${GISBASE}/etc/monitorcap > ${INST_DIR}/etc/monitorcap ; true"
${SHELL} -c "chmod -R 1777 ${INST_DIR}/locks 2>/dev/null ; true"
${SHELL} -c "chmod -R a+rX ${INST_DIR} 2>/dev/null ; true"
The bit regarding monitorcap and the final chmod could be critical.
Let's go for the update. I can do that (welcome pre5).
Glynn, could you check if we can use the pre4 branch? It should
be still in a good condition.
Luckily there are no binaries posted yet (hi Andreas, any
news on your bin-build script?).
Markus
On Tue, May 21, 2002 at 06:50:46PM +0200, Markus Neteler wrote:
On Tue, May 21, 2002 at 05:33:27PM +0100, Glynn Clements wrote:
> > Grass5.0.0pre4 is out apparently.
> > Trouble is, there is no ./dist.X/locale directory.
> We need to release an update. This bug is probably going to bite
> *most* users, as NLS is disabled by default.
Let's go for the update. I can do that (welcome pre5).
Glynn, could you check if we can use the pre4 branch? It should
be still in a good condition.
We must to use the release branch, that's what it is for.
Do you really think we should release pre5 that soon?
We probably find even more release critical bugs...
Bernhard Reiter wrote:
> > > Grass5.0.0pre4 is out apparently.
> > > Trouble is, there is no ./dist.X/locale directory.
> > We need to release an update. This bug is probably going to bite
> > *most* users, as NLS is disabled by default.
> Let's go for the update. I can do that (welcome pre5).
> Glynn, could you check if we can use the pre4 branch? It should
> be still in a good condition.
We must to use the release branch, that's what it is for.
Do you really think we should release pre5 that soon?
We probably find even more release critical bugs...
I don't care whether we call it pre5 or pre4a or whatever. But we
*need* to supersede the current "pre4"; "make install" fails, which is
genuinely *critical*, IMHO.
--
Glynn Clements <glynn.clements@virgin.net>
Markus Neteler wrote:
Let's go for the update. I can do that (welcome pre5).
Glynn, could you check if we can use the pre4 branch? It should
be still in a good condition.
The only changes between pre4 (release_13_may_2002_grass5_0_0_pre4)
and the current state of the release branch (releasebranch_26_april_2002_5_0_0)
are:
documents/cygwin_grass50bininstall.html
Additional information regarding the "sunrpc" package
html/index.html
html/searchhtml.html
Change "search" form URL from uni-hannover.de to itc.it.
src/raster/r.sun/main.c
Remove debugging fprintf()
src/raster/r.tiff/r.out.tiff.c
Fix buffer allocation bug
Note: the actual issue (command failing if "locale" doesn't exist)
still needs to be fixed. I suggest simply omitting the "test", i.e.
${SHELL} -c "cd ${GISBASE} ; tar cBf - locale | (cd ${INST_DIR} ; tar xBf - ) 2>/dev/null ; true"
This will print an error message, but will succeed, and so won't
terminate the installation prematurely.
To eliminate the error message, simply have GISGEN.sh create the
locale directory.
--
Glynn Clements <glynn.clements@virgin.net>
On Tue, May 21, 2002 at 06:23:15PM +0100, Glynn Clements wrote:
Bernhard Reiter wrote:
I don't care whether we call it pre5 or pre4a or whatever. But we
*need* to supersede the current "pre4"; "make install" fails, which is
genuinely *critical*, IMHO.
Okay, then we have to call it "pre5".
Bernhard
I was trying to merge two vector files after performing the v.rmedge
function. The v.merge function never got throught and I decided to take a
look at it and found that the mapset say 'root' is being added to the end
of the input and hence the file that was supposed to be opened was say
like /usr/local/grass5/data/root/dig_cats/fields@root. I took care of this
using a strtok and took off the rest of the string starting from the @
symbol. It again crashed and I found that the sscanf was behaving
erroneously..I had to relplace this...was there anyone else who had
experienced the same or is it my configuration???
Thanks
Venkat