Hello,
I may (still uncertain!) slowly try to get r.le programs to work under 6.X. Here is what I see as problems to be fixed. If anyone knows of other problems, let me know:
1. r.le.setup fails when a sitefile is requested--update needed
2. r.le.setup does not display properly in the display window
3. with gis.m, r.le.setup does not connect to display window
4. r.le.pixel crashes with a memory problem
5. no r.le.dist program
Bill Baker, Univ. of Wyoming
bakerwl@uwyo.edu
On Mon, 2006-09-25 at 14:51 -0600, William L. Baker wrote:
Hello,
I may (still uncertain!) slowly try to get r.le programs to work under
6.X. Here is what I see as problems to be fixed. If anyone knows of
other problems, let me know:1. r.le.setup fails when a sitefile is requested--update needed
2. r.le.setup does not display properly in the display window
3. with gis.m, r.le.setup does not connect to display window
4. r.le.pixel crashes with a memory problem
5. no r.le.dist program
William,
r.le needs major work. It has not been touched in several years and I'm
a bit surprised that it still works at all. I got a start on it last
week by at least resolving compiler errors/warnings for r.le.patch, but
I didn't really attempt to fix anything.
Is there anyone who has the time to deal with r.le in the near term? It
will be a fairly time consuming project if you plan on updating text I/O
(it references stderr, stdout, and files extensively).
What happened to r.le.dist, anyway? Would it be best to remove r.le
from production branches (6.2) until it works reliably again?
--
Brad Douglas <rez touchofmadness com> KB8UYR
Address: 37.493,-121.924 / WGS84 National Map Corps #TNMC-3785
Hello Brad,
Thank you for your thoughts. I wrote the code with assistance, so have some advantages there. But, disadvantage is that I have not followed GRASS updates since 5.2 or so. For example, I did not know there was a problem with using stderr, stdout, or files. Can someone explain what should be done instead?
Yes, it could be removed from the production branches if you like. Until I work on it, there probably won’t be fixes. But, if someone who knows GRASS revisions could help, then maybe this could be done more quickly? But, eventually I would like it to work, and I will slowly get up to speed and fix it even if no one has time (with some occasional help).
One thing that would help me right away would be if someone would quickly scan the source code to see any obvious inconsistencies with 6.xxx.
r.le.dist was never ported to the 5.XXX revision, as I ran out of time to do that. So, it would require a much larger revision.
r.le.patch, r.le.setup, and r.le.pixel I think should not be so hard.
Bill Baker, Univ. of Wyoming
Brad Douglas wrote:
On Mon, 2006-09-25 at 14:51 -0600, William L. Baker wrote:Hello, I may (still uncertain!) slowly try to get r.le programs to work under 6.X. Here is what I see as problems to be fixed. If anyone knows of other problems, let me know: 1. r.le.setup fails when a sitefile is requested--update needed 2. r.le.setup does not display properly in the display window 3. with gis.m, r.le.setup does not connect to display window 4. r.le.pixel crashes with a memory problem 5. no r.le.dist programWilliam, r.le needs major work. It has not been touched in several years and I'm a bit surprised that it still works at all. I got a start on it last week by at least resolving compiler errors/warnings for r.le.patch, but I didn't really attempt to fix anything. Is there anyone who has the time to deal with r.le in the near term? It will be a fairly time consuming project if you plan on updating text I/O (it references stderr, stdout, and files extensively). What happened to r.le.dist, anyway? Would it be best to remove r.le from production branches (6.2) until it works reliably again?
I completely read the original message wrong. I thought you were making
a bug report for some reason. Doh! 
I hadn't realized you were the
author.
I believe the usages of stderr, stdout,etc. are proper, but there's a
push to get the modules to use G_message(), G_warning(),
G_fatal_error(), etc. It'll be a large task that will eventually need
to be done.
If you'd like some help, I can assist. I'll go over the source again
and send you an email off list. If you'll be actively working on it,
there's no need to remove it from CVS.
On Mon, 2006-09-25 at 15:43 -0600, William L. Baker wrote:
Hello Brad,
Thank you for your thoughts. I wrote the code with assistance, so have
some advantages there. But, disadvantage is that I have not followed
GRASS updates since 5.2 or so. For example, I did not know there was a
problem with using stderr, stdout, or files. Can someone explain what
should be done instead?Yes, it could be removed from the production branches if you like.
Until I work on it, there probably won't be fixes. But, if someone who
knows GRASS revisions could help, then maybe this could be done more
quickly? But, eventually I would like it to work, and I will slowly
get up to speed and fix it even if no one has time (with some
occasional help).One thing that would help me right away would be if someone would
quickly scan the source code to see any obvious inconsistencies with
6.xxx.r.le.dist was never ported to the 5.XXX revision, as I ran out of time
to do that. So, it would require a much larger revision.r.le.patch, r.le.setup, and r.le.pixel I think should not be so hard.
Bill Baker, Univ. of Wyoming
Brad Douglas wrote:
> On Mon, 2006-09-25 at 14:51 -0600, William L. Baker wrote:
>
> > Hello,
> >
> > I may (still uncertain!) slowly try to get r.le programs to work under
> > 6.X. Here is what I see as problems to be fixed. If anyone knows of
> > other problems, let me know:
> >
> > 1. r.le.setup fails when a sitefile is requested--update needed
> > 2. r.le.setup does not display properly in the display window
> > 3. with gis.m, r.le.setup does not connect to display window
> > 4. r.le.pixel crashes with a memory problem
> > 5. no r.le.dist program
> >
>
> William,
>
> r.le needs major work. It has not been touched in several years and I'm
> a bit surprised that it still works at all. I got a start on it last
> week by at least resolving compiler errors/warnings for r.le.patch, but
> I didn't really attempt to fix anything.
>
> Is there anyone who has the time to deal with r.le in the near term? It
> will be a fairly time consuming project if you plan on updating text I/O
> (it references stderr, stdout, and files extensively).
>
> What happened to r.le.dist, anyway? Would it be best to remove r.le
> from production branches (6.2) until it works reliably again?
--
Brad Douglas <rez touchofmadness com> KB8UYR
Address: 37.493,-121.924 / WGS84 National Map Corps #TNMC-3785
On Mon, Sep 25, 2006 at 02:20:42PM -0700, Brad Douglas wrote:
On Mon, 2006-09-25 at 14:51 -0600, William L. Baker wrote:
> Hello,
>
> I may (still uncertain!) slowly try to get r.le programs to work under
> 6.X. Here is what I see as problems to be fixed. If anyone knows of
> other problems, let me know:
>
> 1. r.le.setup fails when a sitefile is requested--update needed
> 2. r.le.setup does not display properly in the display window
> 3. with gis.m, r.le.setup does not connect to display window
> 4. r.le.pixel crashes with a memory problem
> 5. no r.le.dist programWilliam,
r.le needs major work. It has not been touched in several years and I'm
a bit surprised that it still works at all. I got a start on it last
week by at least resolving compiler errors/warnings for r.le.patch, but
I didn't really attempt to fix anything.Is there anyone who has the time to deal with r.le in the near term? It
will be a fairly time consuming project if you plan on updating text I/O
(it references stderr, stdout, and files extensively).What happened to r.le.dist, anyway? Would it be best to remove r.le
from production branches (6.2) until it works reliably again?
hi,
please note that there are the r.li efforts:
http://www.faunalia.it/mediawiki/index.php/R_li
Maybe efforts could be joined?
Markus
Yes, as Markus suggests, perhaps before we dig into r.le and start cleaning it up, I’d better look at r.li, which I don’t yet have installed. More once I’ve done that…
Bill Baker, Univ. of Wyoming
Brad Douglas wrote:
I completely read the original message wrong. I thought you were making a bug report for some reason. Doh! :-) I hadn't realized you were the author. I believe the usages of stderr, stdout,etc. are proper, but there's a push to get the modules to use G_message(), G_warning(), G_fatal_error(), etc. It'll be a large task that will eventually need to be done. If you'd like some help, I can assist. I'll go over the source again and send you an email off list. If you'll be actively working on it, there's no need to remove it from CVS. On Mon, 2006-09-25 at 15:43 -0600, William L. Baker wrote:Hello Brad, Thank you for your thoughts. I wrote the code with assistance, so have some advantages there. But, disadvantage is that I have not followed GRASS updates since 5.2 or so. For example, I did not know there was a problem with using stderr, stdout, or files. Can someone explain what should be done instead? Yes, it could be removed from the production branches if you like. Until I work on it, there probably won't be fixes. But, if someone who knows GRASS revisions could help, then maybe this could be done more quickly? But, eventually I would like it to work, and I will slowly get up to speed and fix it even if no one has time (with some occasional help). One thing that would help me right away would be if someone would quickly scan the source code to see any obvious inconsistencies with 6.xxx. r.le.dist was never ported to the 5.XXX revision, as I ran out of time to do that. So, it would require a much larger revision. r.le.patch, r.le.setup, and r.le.pixel I think should not be so hard. Bill Baker, Univ. of Wyoming Brad Douglas wrote:On Mon, 2006-09-25 at 14:51 -0600, William L. Baker wrote:Hello, I may (still uncertain!) slowly try to get r.le programs to work under 6.X. Here is what I see as problems to be fixed. If anyone knows of other problems, let me know: 1. r.le.setup fails when a sitefile is requested--update needed 2. r.le.setup does not display properly in the display window 3. with gis.m, r.le.setup does not connect to display window 4. r.le.pixel crashes with a memory problem 5. no r.le.dist programWilliam, r.le needs major work. It has not been touched in several years and I'm a bit surprised that it still works at all. I got a start on it last week by at least resolving compiler errors/warnings for r.le.patch, but I didn't really attempt to fix anything. Is there anyone who has the time to deal with r.le in the near term? It will be a fairly time consuming project if you plan on updating text I/O (it references stderr, stdout, and files extensively). What happened to r.le.dist, anyway? Would it be best to remove r.le from production branches (6.2) until it works reliably again?
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Yes, I would suggest to look seriously at r.li: it needs development,
but its structure should be more modern and modular, thus easier to
maintain than r.le (in fact, it was after trying to maintain r.le that
we decided to start with r.li). Serena is working on this, developing
new indices. It should be easy from this to add many more indices.
Please those interested note that we also have a mailing list specific
to the project (ready to merge in grass-dev when necessary).
All the best.
pc
William L. Baker ha scritto:
Yes, as Markus suggests, perhaps before we dig into r.le and start cleaning it
up, I'd better look at r.li, which I don't yet have installed. More once I've
done that...Bill Baker, Univ. of Wyoming
Brad Douglas wrote:
I completely read the original message wrong. I thought you were making
a bug report for some reason. Doh!
author.I believe the usages of stderr, stdout,etc. are proper, but there's a
push to get the modules to use G_message(), G_warning(),
G_fatal_error(), etc. It'll be a large task that will eventually need
to be done.If you'd like some help, I can assist. I'll go over the source again
and send you an email off list. If you'll be actively working on it,
there's no need to remove it from CVS.
- --
Paolo Cavallini
email+jabber: cavallini@faunalia.it
www.faunalia.it
Piazza Garibaldi 5 - 56025 Pontedera (PI), Italy Tel: (+39)348-3801953
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFFGLOp/NedwLUzIr4RAn7yAJoDvsIBoS9jGEW/HeEDFF8OnbBb3ACgtNjt
xQflgGg1086vTEWFVKOzrCg=
=Rua1
-----END PGP SIGNATURE-----
William L. Baker wrote:
I may (still uncertain!) slowly try to get r.le programs to work under
6.X.
The raster format & GIS API is unchanged, only vector & sites have
changed. So required fixes for porting a raster module are in fact
small.
Here is what I see as problems to be fixed. If anyone knows of
other problems, let me know:1. r.le.setup fails when a sitefile is requested--update needed
fixed in CVS HEAD + 6.2 branch
- just removed the prompt
In CVS HEAD I've changed the module to use G_parser() instead of custom
prompts.
2. r.le.setup does not display properly in the display window
fixed in CVS HEAD + 6.2 branch
- it was hanging on G_system(" d.frame -e");
changed to use D_setup(1) to clear the screen
3. with gis.m, r.le.setup does not connect to display window
fixed in CVS HEAD + 6.2 branch
- menu.tcl needed "guarantee_xmon;"
4. r.le.pixel crashes with a memory problem
Brad?
William: Can you prepare an example using the spearfish dataset? These
modules are too complex for me to quickly understand.
5. no r.le.dist program
Also, no r.le.para as exists in GRASS 5, nor r.le.null & r.le.rename as
mentioned in the help page. If these should be copied from GRASS 5, let
us know, it's not hard to do. I don't see r.le.dist built in GRASS 5.4
either??
I notice that the GRASS 5 r.le man pages were not ported into GRASS 6
description.html files. That's easy to do and should be done.
Brad:
Would it be best to remove r.le from production branches (6.2) until
it works reliably again?
If r.le is deemed not ready for release, removing it is as simple as
editing raster/Makefile and removing the entries for it in menu.tcl.
Markus:
please note that there are the r.li efforts:
http://www.faunalia.it/mediawiki/index.php/R_li
Maybe efforts could be joined?
Yes please.. but it is simple to copy over missing r.le modules and man
pages, it can be done in a few minutes. So we should do that while we wait
for r.li to be ready.
Hamish
Hamish,
Great. I will get the CVS HEAD version next and have a look. Will try r.le.pixel with sample dataset. I will see if there are other major problems that prevent some use of r.le for now, until r.li is fully developed.
r.le.para, r.le.null, and r.le.rename are not very important programs and for now they can be dropped.
r.le.dist is something I should update to 6.XXX, but that will take quite some time (months).
I will leave it to GRASS developers to decide whether to keep r.le for now or remove it until it is fixed or r.li is ready. If it is removed, I can work on it with some help from you all and then put it back in once it is ready. If it is kept in, maybe Hamish and I and some others can make sure it generally works.
Thank you!
Bill Baker
Hamish wrote:
Hamish wrote:
William L. Baker wrote:
I may (still uncertain!) slowly try to get r.le programs to work under
6.X.The raster format & GIS API is unchanged, only vector & sites have
changed. So required fixes for porting a raster module are in fact
small.Here is what I see as problems to be fixed. If anyone knows of
other problems, let me know:1. r.le.setup fails when a sitefile is requested--update needed
fixed in CVS HEAD + 6.2 branch
- just removed the promptIn CVS HEAD I've changed the module to use G_parser() instead of custom
prompts.2. r.le.setup does not display properly in the display window
fixed in CVS HEAD + 6.2 branch
- it was hanging on G_system(" d.frame -e");
changed to use D_setup(1) to clear the screen
Does it resolve the 2 issues reported in
http://intevation.de/rt/webrt?serial_num=3323?
3. with gis.m, r.le.setup does not connect to display window
fixed in CVS HEAD + 6.2 branch
- menu.tcl needed "guarantee_xmon;"4. r.le.pixel crashes with a memory problem
Brad?
William: Can you prepare an example using the spearfish dataset? These
modules are too complex for me to quickly understand.
For the record - there is bug report for about that:
http://intevation.de/rt/webrt?serial_num=3084
5. no r.le.dist program
Other folks are missing it too:
http://intevation.de/rt/webrt?serial_num=3040
Also, no r.le.para as exists in GRASS 5, nor r.le.null & r.le.rename as
mentioned in the help page. If these should be copied from GRASS 5, let
us know, it's not hard to do. I don't see r.le.dist built in GRASS 5.4
either??
I notice that the GRASS 5 r.le man pages were not ported into GRASS 6
description.html files. That's easy to do and should be done.
I'm sorry but I can't offer any help with documenation. No time.
Please note there is also a wish regarding r.le.pixel:
http://intevation.de/rt/webrt?serial_num=2104
Thanks for trying to revive r.le.* suite. If there are any news in
regard to the 4 BT tickets I mentioned, please put your comments in
them, CCing me tutey at 02.pl, so we can keep the track of it.
Cheers,
Maciek
P.S.
I will be less available these days. Hopefully in 2 weeks I will have
more time.
>> 2. r.le.setup does not display properly in the display window
..
> fixed in CVS HEAD + 6.2 branch
> - it was hanging on G_system(" d.frame -e");
> changed to use D_setup(1) to clear the screenDoes it resolve the 2 issues reported in
http://intevation.de/rt/webrt?serial_num=3323?
yes, I think. I don't know about r.fragstats comments. see report.
>> 4. r.le.pixel crashes with a memory problem
..
For the record - there is bug report for about that:
http://intevation.de/rt/webrt?serial_num=3084
no, that sounds like a slow memory leak not a malloc mistake.
probably the segmentation suggestion is the solution/cause of #3084.
>> 5. no r.le.dist program
Other folks are missing it too:
http://intevation.de/rt/webrt?serial_num=3040
fyi, this wish was spam-bombed 4 days ago.
> I notice that the GRASS 5 r.le man pages were not ported into GRASS
> 6 description.html files. That's easy to do and should be done.I'm sorry but I can't offer any help with documenation. No time.
hopefully someone can, it's just cut & paste stuff.
WRT r.li, I think it'll be great to have that done and agree that all
major development effort probably belongs there. But if we can fix a few
bugs in a few hours to make r.le useful again, I don't think that hurts
any. At least we will have put r.le to rest in a comfortable state.
Hamish