While it is nice to see everyone all chatty about the mailing list I have
more down to earth questions about a problem in Grass5.0beta on my Mklinux
machine. With Tom's excellent help (and a few key pieces from Malcolm) I
have been able to get all of Grass compiled except for a few minor moudules (I left the list at my office).
Most significant was that I had to go to /etc and do a ln -s /bin/mknod mknod
to get mknod to work correctly. As a result of that everything seems to
be working fine including tcltkgrass but I have run into two problems.
A: (and I really need this one to work!) in s.menu after loading up my
sites and then doing an occurrencre report on them agaist a raster map,
the calculations run but the program halts when it should be giving me
the report and tells me
Can't open report file
/usr/local/grass_data/rjoe/test/rjoe/.tmp/localhost.localdomain/4478.0 no
such file for directory.
each time I run this it gives me a different number after ...localdomain/
I have the feeling this is a permissions problem and so on a whim did a
chmod 777 on localhost.localdomain but that did not fix anything...
comments please?
B: v.digit, which I also really need to work did not compile until after I
fixed the mknod problem. I reran GISGEN starting with v.digit in the nextstep and everything went
fine. I then ran MAKELINKS.head. Now when I enter v.digit it says it can
not find it?
Thanks!
R. Joe Brandon
---------------------------------------------------------------------
Center for Advanced Spatial Technologies (fax) 501 575-5218
Rm. 12 Ozark Hall 501 575-4277
University of Arkansas rjoe@cast.uark.edu
Fayetteville, AR 72701
Partially restored -> http://www.cast.uark.edu/~rjoe
---------------------------------------------------------------------
[on GRASS 5.0]
On Wed, Jun 02, 1999 at 03:18:59PM -0500, R. Joe Brandon wrote:
Most significant was that I had to go to /etc and do a ln -s /bin/mknod mknod
to get mknod to work correctly.
For me and Bill it is a known problem.
src/CMD/generic/GENGIS.sh has the path to mknod hardcoded, which
IMO is a bug.
Unfortunatly the baylor developers didn't pay much attention on
my and Bill's patches.
Can't open report file
/usr/local/grass_data/rjoe/test/rjoe/.tmp/localhost.localdomain/4478.0 no
such file for directory.
each time I run this it gives me a different number after ...localdomain/
I have the feeling this is a permissions problem and so on a whim did a
chmod 777 on localhost.localdomain but that did not fix anything...
no idea on that one.
Are you sure that the exact paths exists?
B: v.digit, which I also really need to work did not compile until after I
fixed the mknod problem. I reran GISGEN starting with v.digit in the nextstep and everything went
fine. I then ran MAKELINKS.head. Now when I enter v.digit it says it can
not find it?
Note that the GRASS5.0b source was subject to unnoticed change.
You should check if you have the last on from the Baylor ftp site.
(Look at the date and the size.)
IIRC v.digit should be superceded by v.digspline, but I did not
seriously try to run one of thoses.
I hope these few clues are useful to you.
Bernhard
On Wed, 2 Jun 1999, Bernhard Reiter wrote:
> Can't open report file
>=20
> /usr/local/grass_data/rjoe/test/rjoe/.tmp/localhost.localdomain/4478.0 no=
such file for directory each time I run this it gives me a different
number after ...localdomain/ I have the feeling this is a permissions
problem and so on a whim did chmod 777 on localhost.localdomain but that
did not fix anything. no idea on that one.
Are you sure that the exact paths exists?
i just checked and the exact path does exist, and I do have read/write
permisions there. It seems to be gumming up a few routines but I am at a
loss for what to do. I have even gone in as myself and manually and created
files there to make sure the permissions were fine, and it was so. Very strange. Anyone have any
more ideas?
As for the v.digit comment, did you mean that there were times when
5.0beta wa sin the directory minus certain modules?
Thanks!
R. Joe
On Wed, Jun 02, 1999 at 11:12:57PM -0500, R. Joe Brandon wrote:
On Wed, 2 Jun 1999, Bernhard Reiter wrote:
i just checked and the exact path does exist, and I do have read/write
permisions there. It seems to be gumming up a few routines but I am at a
loss for what to do. I have even gone in as myself and manually and created
files there to make sure the permissions were fine, and it was so. Very strange. Anyone have any
more ideas?
What I usually do is to look it up in the source code to see what is
happening. But that is no real idea. I think I could do some
configuration regarding GRASS running on different computers.
This might have something to do with it.
As for the v.digit comment, did you mean that there were times when
5.0beta wa sin the directory minus certain modules?
The source was definetly in the Baylor ftp directory in different
versions. At one time IIRC at least one module v.digspline was missing.
Bernhard
We never received any patches from you.
Keep us posted. Cheers!
Bruce
Bernhard Reiter wrote:
[on GRASS 5.0]
On Wed, Jun 02, 1999 at 03:18:59PM -0500, R. Joe Brandon wrote:
> Most significant was that I had to go to /etc and do a ln -s /bin/mknod mknod
> to get mknod to work correctly.
For me and Bill it is a known problem.
src/CMD/generic/GENGIS.sh has the path to mknod hardcoded, which
IMO is a bug.
Unfortunatly the baylor developers didn't pay much attention on
my and Bill's patches.
> Can't open report file
>
> /usr/local/grass_data/rjoe/test/rjoe/.tmp/localhost.localdomain/4478.0 no
> such file for directory.
>
> each time I run this it gives me a different number after ...localdomain/
>
> I have the feeling this is a permissions problem and so on a whim did a
> chmod 777 on localhost.localdomain but that did not fix anything...
no idea on that one.
Are you sure that the exact paths exists?
> B: v.digit, which I also really need to work did not compile until after I
> fixed the mknod problem. I reran GISGEN starting with v.digit in the nextstep and everything went
> fine. I then ran MAKELINKS.head. Now when I enter v.digit it says it can
> not find it?
Note that the GRASS5.0b source was subject to unnoticed change.
You should check if you have the last on from the Baylor ftp site.
(Look at the date and the size.)
IIRC v.digit should be superceded by v.digspline, but I did not
seriously try to run one of thoses.
I hope these few clues are useful to you.
Bernhard
------------------------------------------------------------------------
Part 1.2 Type: application/pgp-signature
--
Bruce Byars
Hydrologist and Senior Research Scientist
Center for Applied Geographic and Spatial Research
Baylor University
On Thu, Jun 03, 1999 at 01:06:09PM -0500, B Byars wrote:
We never received any patches from you.
Keep us posted. Cheers!
My changes are largely incorporated into Bill's, see:
http://grasshopper.grasshoppernet.com/~bhughes/
Mails to the Baylor Address about several issues were never answered.
Bill told me that he also never got an answer from the developers.
The webstructure on Baylor seems to be a bit off, and nobody seems
to care to link Bill's site from there.
The work was mentioned on the user list, too.
Markus Neteler then contacted us and AFAIK mentioned the B. Hughes'
patches to the Baylor developers, which at least partly seem to
incorporate them. But still no comment about that.
Bernhard