Is anybody summarizing the list of patches and fixes
required to compile 5.0beta? I'm confused
with the many messages on this regard.
Thanks
Agus
Dr. Agustin Lobo
Instituto de Ciencias de la Tierra (CSIC)
Lluis Sole Sabaris s/n
08028 Barcelona SPAIN
tel 34 93409 5410
fax 34 93411 0012
alobo@ija.csic.es
http://pangea.ija.csic.es/alobo
Hello all
On Aug 5, 11:34am, Agustin Lobo wrote:
Is anybody summarizing the list of patches and fixes
required to compile 5.0beta? I'm confused
with the many messages on this regard.
I am compiling the beta2 code on our SGI Origin 2000 running IRIX 6.5 and am
keeping track of the problems and the fixes I have run into. I will post them
if someone wants them when I am finished. However, at the moment, I'm stumped.
I don't know much about how configure works and I have the following problem.
In the module grass5.0beta2/src/general/g.help I have error messages for the
file menu.c as follows:
line 742: error(1565): struct "_win_st" has no field "NONE"
Maxlines = ((short)Numlines <= Windoname->CURSES_MAXY) ?
I have tracked the problem down to an entry in
grass5.0beta2/src/include/config.h which is as follows:
/* define curses.h WINDOW structure component */
#define CURSES_MAXY NONE
This line was generated from the following line in
grass5.0beta2/src/include/config.h.in
/* define curses.h WINDOW structure component */
#define CURSES_MAXY @CURSES_MAXY@
What is happening is that CURSES_MAXY is supposed to reference a member of the
_win_st struct defined in /usr/include/curses.h. However, the configure script
came up with "NONE" on my machine. So, does anyone know what should be the
proper value, and how I can get configure to find it?
Any help or suggestions will be greatly appreciated.
--
Sincerely,
People who think they know everything are very irritating to those
of us who do. ---Anonymous
Hello Jazzman,
I am interested in the your compilation experience for SGI Origin 2000.
Are you compiling with switches that make use of
multiple processors?. We have one with 10 processors, and
it will usefull to take advantege of the multi-thread processing,
specially for the heavy stuff like r.watershed, r. water.fea , etc.
Regards
Alejandro Hinojosa
At 06:02 PM 8/5/99 +0000, you wrote:
Hello all
On Aug 5, 11:34am, Agustin Lobo wrote:
Is anybody summarizing the list of patches and fixes
required to compile 5.0beta? I'm confused
with the many messages on this regard.I am compiling the beta2 code on our SGI Origin 2000 running IRIX 6.5 and am
keeping track of the problems and the fixes I have run into. I will post them
if someone wants them when I am finished. However, at the moment, I'm stumped.
I don't know much about how configure works and I have the following problem.In the module grass5.0beta2/src/general/g.help I have error messages for the
file menu.c as follows:
line 742: error(1565): struct "_win_st" has no field "NONE"
Maxlines = ((short)Numlines <= Windoname->CURSES_MAXY) ?I have tracked the problem down to an entry in
grass5.0beta2/src/include/config.h which is as follows:
/* define curses.h WINDOW structure component */
#define CURSES_MAXY NONEThis line was generated from the following line in
grass5.0beta2/src/include/config.h.in
/* define curses.h WINDOW structure component */
#define CURSES_MAXY @CURSES_MAXY@What is happening is that CURSES_MAXY is supposed to reference a member of the
_win_st struct defined in /usr/include/curses.h. However, the configure script
came up with "NONE" on my machine. So, does anyone know what should be the
proper value, and how I can get configure to find it?Any help or suggestions will be greatly appreciated.
--
Sincerely,Jazzman (a.k.a. Justin Hickey) e-mail: jhickey@hpcc.nectec.or.th
High Performance Computing Center
National Electronics and Computer Technology Center (NECTEC)
Bangkok, ThailandPeople who think they know everything are very irritating to those
of us who do. ---AnonymousJazz and Trek Rule!!!
Hello all,
We've been onitoring the users group here and have setup
email for you to send your patch info to. Send to:
patches@bluefin.baylor.edu
Hopefully this can serve to keep us all on the same page
and allow us to provide better support to the community.
Also, if you have patcheas you'd like to share, upload
them to keck.baylor.edu and put them in the
/pub/grass/incoming folder. Be sure to include a readme
file. Thanks.
Bruce
--
Bruce Byars
Hydrologist and Senior Research Scientist
Center for Applied Geographic and Spatial Research
Baylor University
I am trying to install grass5.0beta2 on my MkLinux system which did have
grass5.0beta running on it succsesfully.
I placed grass5.0beta2.tar.gz in /usr/local and did a gzip -d. I then did
tar -xvf on the resulting grass5.0beta.tar. After untarring I cd'd into
the new directory grass5.0beta and ran ./configure both as a user, and then
as root (after I had the problem below) everything seemed to go well with
the last few lines of the data parsed to the screen from configure
reading...
<SNIP>
checking for xdr_int in -lsun... (cached) no
checking for xdr_int in -lnsl... (cached) yes
creating ./config.status
creating Makefile
creating src/CMD/head/head
creating src/scripts/shells/create_fifos.sh
creating src/include/config.h
src/include/config.h is unchanged
I then ran "make install" and received this error?...
install -d -m 755 /usr/local/grass-5.0b /home/grass/.bashrc
UNIX_BIN=/usr/local/bin SRC=`pwd` /bin/sh src/CMD/generic/CREATE_GMAKE.sh
/bin/sh: /home/grass/.bashrc: Permission denied
make: *** [pre-install] Error 126
Any ideas?
R. Joe Brandon
mailto:rjoe_brandon@iname.com
http://www.cast.uark.edu/~rjoe
I had problems similar to below on an intel/Linux machine. The solution
was to compile and install the latest version of ncurses (ncurses-4.2)
from the gnu site. I don't know about SGI...
On Thu, 5 Aug 1999, Justin Hickey wrote:
Hello all
On Aug 5, 11:34am, Agustin Lobo wrote:
> Is anybody summarizing the list of patches and fixes
> required to compile 5.0beta? I'm confused
> with the many messages on this regard.I am compiling the beta2 code on our SGI Origin 2000 running IRIX 6.5 and am
keeping track of the problems and the fixes I have run into. I will post them
if someone wants them when I am finished. However, at the moment, I'm stumped.
I don't know much about how configure works and I have the following problem.In the module grass5.0beta2/src/general/g.help I have error messages for the
file menu.c as follows:
line 742: error(1565): struct "_win_st" has no field "NONE"
Maxlines = ((short)Numlines <= Windoname->CURSES_MAXY) ?I have tracked the problem down to an entry in
grass5.0beta2/src/include/config.h which is as follows:
/* define curses.h WINDOW structure component */
#define CURSES_MAXY NONEThis line was generated from the following line in
grass5.0beta2/src/include/config.h.in
/* define curses.h WINDOW structure component */
#define CURSES_MAXY @CURSES_MAXY@What is happening is that CURSES_MAXY is supposed to reference a member of the
_win_st struct defined in /usr/include/curses.h. However, the configure script
came up with "NONE" on my machine. So, does anyone know what should be the
proper value, and how I can get configure to find it?Any help or suggestions will be greatly appreciated.
--
Sincerely,Jazzman (a.k.a. Justin Hickey) e-mail: jhickey@hpcc.nectec.or.th
High Performance Computing Center
National Electronics and Computer Technology Center (NECTEC)
Bangkok, ThailandPeople who think they know everything are very irritating to those
of us who do. ---AnonymousJazz and Trek Rule!!!
Hello Alejandro
On Aug 5, 9:41am, Alejandro Hinojosa wrote:
Subject: Re: patches to 5.0beta + I'm stumped
Hello Jazzman,
I am interested in the your compilation experience for SGI Origin 2000.
Are you compiling with switches that make use of
multiple processors?. We have one with 10 processors, and
it will usefull to take advantege of the multi-thread processing,
specially for the heavy stuff like r.watershed, r. water.fea , etc.
We are not using any switches for multiple processing. The reason why is that
in order to use multiple processors efficiently, the code must be written with
parallel programming techniques. Simply telling the compiler that you want a
program (originaly written as a sequential program) to use multiple processors
does not usually produce good parallel code. In fact I have seen instances
where the resulting parallel code was slower than the sequential version. If
any of the grass modules are to be parallelized (which would be great for some
of them), then the code for those modules needs to be re-written.
Unfortunately, we don't have the time to do that. So for now, we continue to
use sequential code.
Sorry I couldn't help you on this, but I will post the problems and the fixes I
had compiling grass when I finish (if I ever finish :).
--
Sincerely,
People who think they know everything are very irritating to those
of us who do. ---Anonymous
Justin Hickey wrote:
We are not using any switches for multiple processing. The reason why is that
in order to use multiple processors efficiently, the code must be written with
parallel programming techniques. Simply telling the compiler that you want a
program (originaly written as a sequential program) to use multiple processors
My experience with the SGI Power C and Power F77 compilers is that you will get
very near the speed up expected for the number of processors, simply by using the
parallel code switch. This is particularly the case when dealing with arrays, and
hence most of the raster processing tools would show a noticable improvement.
I suggest all that needs to be done (if one has the Power C compiler) is to modify
the definition of the CC command in the basic makefile (one would, of course, need
to suss out what that was, and where to modify the basic makefile). It would
certainly be worth a try, and if it wasn't successful, simply go back to original
CC definition and rebuild.
Cheers,
--
Terry Duell, Senior Mobility Engineer
Army Engineering Agency
Maribyrnong, Victoria, Australia
ph:61-3-93195837 fax:61-3-93195830
Hello Bruce
On Aug 5, 2:15pm, B Byars wrote:
Subject: GRASS 5.0 beta patches
We've been onitoring the users group here and have setup
email for you to send your patch info to. Send to:patches@bluefin.baylor.edu
Is there any special format for sending this information (eg, diff output or
something like that)? Or do we just send a message with the changes made to the
code?
Thanks for all your help and effort.
--
Sincerely,
People who think they know everything are very irritating to those
of us who do. ---Anonymous
Hello Terry
On Aug 6, 12:19pm, Terry Duell wrote:
My experience with the SGI Power C and Power F77 compilers is that you will
get very near the speed up expected for the number of processors, simply by
using the parallel code switch.
OK, I will give it a try. My experience was years ago, so they must have
improved their compilers significantly.
Thanks for sharing your experience.
--
Sincerely,
People who think they know everything are very irritating to those
of us who do. ---Anonymous
Done my compile of grass5.0beta3 on the alpha here.
Found the usuall bunch of little obvious bugs, which needed fixing.
GRASS Programmers: Please do not use C++ style like
comments, they are not ANSI-C and a resonable C compiler
will reject them.
On Thu, Aug 05, 1999 at 02:15:10PM -0500, B Byars wrote:
We've been onitoring the users group here and have setup
email for you to send your patch info to. Send to:patches@bluefin.baylor.edu
Okay.
Also, if you have patcheas you'd like to share, upload
them to keck.baylor.edu and put them in the
/pub/grass/incoming folder.
Doesn't work tonight, it gives me a "no space left on device" message.
I attach my patch notes and the patch.
Have fun,
Bernhard
--
Research Assistant, Geog Dept UM-Milwaukee, USA. (www.uwm.edu/~bernhard)
Association for a Free Informational Infrastructure (ffii.org)
Intevation GmbH
grass5.0beta3.bernhard.patch1.txt (1.84 KB)
grass5.0beta3.bernhard.patch1.diff (12.6 KB)