WinGRASS not deleting temp ppm files from map display

[Isaac Ullah]

Before I submitted this as abug, I wanted to run it past the list first. I've been doing  lot of zooming, panning, etc. with WinGRASS map display lately, and I just noticed that my remaining 5gigs of free disk space had been eaten up in the last few days. I haven't downloaded anything, and have mainly just been woring in grass. I checked the size of the mapsets I've been working in and I discovered that the main one is 4.5 gig! It only has a few maps in it, so I checked the size of cell, and it is small. I check all the other folders and they were all small except for ".tmp" which was more than 4Gigs! When I looked inside, I noticed that ALL the ppm files from the map display from the last three days were still in there! I checked other mapsets that I have used in that time frame, and all of them have ppm files in them. mapsets I have not used have empty .tmp folders. Three days ago is when I installed RC4 binaries. Is this an issue with RC4 for anyone else? Is it a Windows thing? I'm using XP, on a laptop. I HAVE gone into hibernation with GRASS still up. Does dong so cause GRASS to "forget" the current PID, and therefore not be able to find the old ppm files to flush? Or is this just a bug in RC4?

Here are contents of the .tmp directory by ls command from MYSYS shell:

iullah@BENGAL1 ~/My Documents/GIS DataBase/Palestine_Grid_1923/PERMANENT/.tmp
$ ls
1240.0.pgm 2796.3.pgm 3672.1.pgm 3696.18.pgm 3696.8.pgm 4328.0
1240.0.ppm 2796.3.ppm 3672.1.ppm 3696.18.ppm 3696.8.ppm 4328.1
1376.0.pgm 2796.4.pgm 3696.0.pgm 3696.19.pgm 3696.9.pgm 4372.0.pgm
1376.0.ppm 2796.4.ppm 3696.0.ppm 3696.19.ppm 3696.9.ppm 4372.0.ppm
1376.1.pgm 3208.0.pgm 3696.1.pgm 3696.2.pgm 3756.0 460.0.pgm
1376.1.ppm 3208.0.ppm 3696.1.ppm 3696.2.ppm 3836.0.pgm 460.0.ppm
1396.0 3284.dbf 3696.10.pgm 3696.20.pgm 3836.0.ppm 484.0.pgm
1644.0.pgm 3288.0 3696.10.ppm 3696.20.ppm 3836.1.pgm 484.0.ppm
1644.0.ppm 3316.0 3696.11.pgm 3696.21.pgm 3836.1.ppm 5108.0
2128.0.pgm 3380.0 3696.11.ppm 3696.21.ppm 3880.0.pgm 5452.0
2128.0.ppm 3380.1 3696.12.pgm 3696.22.pgm 3880.0.ppm 5452.1
2244.0.pgm 3380.2 3696.12.ppm 3696.22.ppm 3900.0.pgm 5452.2
2244.0.ppm 3464.0.pgm 3696.13.pgm 3696.3.pgm 3900.0.ppm 5452.3
2480.0 3464.0.ppm 3696.13.ppm 3696.3.ppm 3932.0.pgm 5452.4
2548.0.pgm 3480.0.pgm 3696.14.pgm 3696.4.pgm 3932.0.ppm 6104.0
2548.0.ppm 3480.0.ppm 3696.14.ppm 3696.4.ppm 4012.0.pgm 672.0.pgm
2796.0.pgm 3596.0.pgm 3696.15.pgm 3696.5.pgm 4012.0.ppm 672.0.ppm
2796.0.ppm 3596.0.ppm 3696.15.ppm 3696.5.ppm 4052.0 672.1.pgm
2796.1.pgm 3656.0 3696.16.pgm 3696.6.pgm 4104.0.pgm 672.1.ppm
2796.1.ppm 3656.1 3696.16.ppm 3696.6.ppm 4104.0.ppm
2796.2.pgm 3672.0.pgm 3696.17.pgm 3696.7.pgm 4236.0.pgm
2796.2.ppm 3672.0.ppm 3696.17.ppm 3696.7.ppm 4236.0.ppm

 

--

Isaac I Ullah, M.A.

Archaeology PhD Student,
ASU School of Evolution and Social Change

Research Assistant,
Mediterranean Landscape Dynamics Project
***************************************************
[hidden email]
[hidden email]

http://www.public.asu.edu/~iullah
***************************************************

_______________________________________________
grass-windows mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/grass-windows

[Isaac Ullah]

I just looked through ALL my mapsets, and I have some other mapsets that also have many ppm files in them (several hunderd Megabites worth). These are mapsets that I heve not used since switching to RC4 (I was using RC3), so it no longer seems to me to be a new poblem with RC4. I DO remember that I let my laptop hibernate while keeping GRASS open while using these other mapsets as well. At the moment, that seems to me to be the most likely culprit. I wonder if this only happen on Windows?

Cheers,

Isaac

On Thu, Apr 16, 2009 at 6:02 PM, Isaac Ullah <[hidden email]> wrote:

Before I submitted this as abug, I wanted to run it past the list first. I've been doing  lot of zooming, panning, etc. with WinGRASS map display lately, and I just noticed that my remaining 5gigs of free disk space had been eaten up in the last few days. I haven't downloaded anything, and have mainly just been woring in grass. I checked the size of the mapsets I've been working in and I discovered that the main one is 4.5 gig! It only has a few maps in it, so I checked the size of cell, and it is small. I check all the other folders and they were all small except for ".tmp" which was more than 4Gigs! When I looked inside, I noticed that ALL the ppm files from the map display from the last three days were still in there! I checked other mapsets that I have used in that time frame, and all of them have ppm files in them. mapsets I have not used have empty .tmp folders. Three days ago is when I installed RC4 binaries. Is this an issue with RC4 for anyone else? Is it a Windows thing? I'm using XP, on a laptop. I HAVE gone into hibernation with GRASS still up. Does dong so cause GRASS to "forget" the current PID, and therefore not be able to find the old ppm files to flush? Or is this just a bug in RC4?

Here are contents of the .tmp directory by ls command from MYSYS shell:

iullah@BENGAL1 ~/My Documents/GIS DataBase/Palestine_Grid_1923/PERMANENT/.tmp
$ ls
1240.0.pgm 2796.3.pgm 3672.1.pgm 3696.18.pgm 3696.8.pgm 4328.0
1240.0.ppm 2796.3.ppm 3672.1.ppm 3696.18.ppm 3696.8.ppm 4328.1
1376.0.pgm 2796.4.pgm 3696.0.pgm 3696.19.pgm 3696.9.pgm 4372.0.pgm
1376.0.ppm 2796.4.ppm 3696.0.ppm 3696.19.ppm 3696.9.ppm 4372.0.ppm
1376.1.pgm 3208.0.pgm 3696.1.pgm 3696.2.pgm 3756.0 460.0.pgm
1376.1.ppm 3208.0.ppm 3696.1.ppm 3696.2.ppm 3836.0.pgm 460.0.ppm
1396.0 3284.dbf 3696.10.pgm 3696.20.pgm 3836.0.ppm 484.0.pgm
1644.0.pgm 3288.0 3696.10.ppm 3696.20.ppm 3836.1.pgm 484.0.ppm
1644.0.ppm 3316.0 3696.11.pgm 3696.21.pgm 3836.1.ppm 5108.0
2128.0.pgm 3380.0 3696.11.ppm 3696.21.ppm 3880.0.pgm 5452.0
2128.0.ppm 3380.1 3696.12.pgm 3696.22.pgm 3880.0.ppm 5452.1
2244.0.pgm 3380.2 3696.12.ppm 3696.22.ppm 3900.0.pgm 5452.2
2244.0.ppm 3464.0.pgm 3696.13.pgm 3696.3.pgm 3900.0.ppm 5452.3
2480.0 3464.0.ppm 3696.13.ppm 3696.3.ppm 3932.0.pgm 5452.4
2548.0.pgm 3480.0.pgm 3696.14.pgm 3696.4.pgm 3932.0.ppm 6104.0
2548.0.ppm 3480.0.ppm 3696.14.ppm 3696.4.ppm 4012.0.pgm 672.0.pgm
2796.0.pgm 3596.0.pgm 3696.15.pgm 3696.5.pgm 4012.0.ppm 672.0.ppm
2796.0.ppm 3596.0.ppm 3696.15.ppm 3696.5.ppm 4052.0 672.1.pgm
2796.1.pgm 3656.0 3696.16.pgm 3696.6.pgm 4104.0.pgm 672.1.ppm
2796.1.ppm 3656.1 3696.16.ppm 3696.6.ppm 4104.0.ppm
2796.2.pgm 3672.0.pgm 3696.17.pgm 3696.7.pgm 4236.0.pgm
2796.2.ppm 3672.0.ppm 3696.17.ppm 3696.7.ppm 4236.0.ppm

 

--

Isaac I Ullah, M.A.

Archaeology PhD Student,
ASU School of Evolution and Social Change

Research Assistant,
Mediterranean Landscape Dynamics Project
***************************************************
[hidden email]
[hidden email]

http://www.public.asu.edu/~iullah
***************************************************

--

Isaac I Ullah, M.A.

Archaeology PhD Student,
ASU School of Evolution and Social Change

Research Assistant,
Mediterranean Landscape Dynamics Project
***************************************************
[hidden email]
[hidden email]

http://www.public.asu.edu/~iullah
***************************************************

_______________________________________________
grass-windows mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/grass-windows

[Glynn Clements]

Isaac Ullah wrote:

> Before I submitted this as abug, I wanted to run it past the list first.
> I've been doing  lot of zooming, panning, etc. with WinGRASS map display
> lately, and I just noticed that my remaining 5gigs of free disk space had
> been eaten up in the last few days. I haven't downloaded anything, and have
> mainly just been woring in grass. I checked the size of the mapsets I've
> been working in and I discovered that the main one is 4.5 gig! It only has a
> few maps in it, so I checked the size of cell, and it is small. I check all
> the other folders and they were all small except for ".tmp" which was more
> than 4Gigs! When I looked inside, I noticed that ALL the ppm files from the
> map display from the last three days were still in there! I checked other
> mapsets that I have used in that time frame, and all of them have ppm files
> in them. mapsets I have not used have empty .tmp folders. Three days ago is
> when I installed RC4 binaries. Is this an issue with RC4 for anyone else? Is
> it a Windows thing? I'm using XP, on a laptop. I HAVE gone into hibernation
> with GRASS still up. Does dong so cause GRASS to "forget" the current PID,
> and therefore not be able to find the old ppm files to flush? Or is this
> just a bug in RC4?

AFAICT, the GUI removes the PPM/PGM files for a layer when you remove
the layer from the display, or when you close the map window.

One possibility is that the close event doesn't get handled on Windows
for whatever reason.

Can you conduct some tests:

1. Start the GUI from the command prompt, display a map, then quit.
Report any error message and whether any PPM/PGM files are left
behind.

2. Start the GUI, add some maps to the list, display them, remove some
maps from the list, check whether the PPM/PGM files are deleted as the
maps are removed.

--
Glynn Clements <[hidden email]>
_______________________________________________
grass-windows mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/grass-windows

[Isaac Ullah]

Glynn, I tried your tests with no replication of the error. Then I had an epiphany. I realized that all the mapsets that had the extra ppm files are mapsets where I was digitizing vectors from features in raster images using the tcltk v.digit module in RC3. I did a test using spearfish, and I was able to replicate the problem. When v.digit loads, it creates it's own ppm and ppg files for the background of the v.digit display (seperate from the ppm and ppg files created for the regular tcltk map display). When one exits from v.digit, those ppm and ppg stay. When one exits from grass, the ppm and ppg files for the map display are erased, but NOT those created by the v.digit display. I don't have access to my linux box at the moment, but this could potentially be a cross-platform issue with the tcltk v.digit. I assume the issue will be fixed in the forthcoming wxpython v.digit as AFAIK the wx gui does not write ppm files for it's display... Shall I submit a bug ticket on this?

Cheers 

Isaac

On Thu, Apr 16, 2009 at 8:00 PM, Glynn Clements <[hidden email]> wrote:

Isaac Ullah wrote:

> Before I submitted this as abug, I wanted to run it past the list first.
> I've been doing  lot of zooming, panning, etc. with WinGRASS map display
> lately, and I just noticed that my remaining 5gigs of free disk space had
> been eaten up in the last few days. I haven't downloaded anything, and have
> mainly just been woring in grass. I checked the size of the mapsets I've
> been working in and I discovered that the main one is 4.5 gig! It only has a
> few maps in it, so I checked the size of cell, and it is small. I check all
> the other folders and they were all small except for ".tmp" which was more
> than 4Gigs! When I looked inside, I noticed that ALL the ppm files from the
> map display from the last three days were still in there! I checked other
> mapsets that I have used in that time frame, and all of them have ppm files
> in them. mapsets I have not used have empty .tmp folders. Three days ago is
> when I installed RC4 binaries. Is this an issue with RC4 for anyone else? Is
> it a Windows thing? I'm using XP, on a laptop. I HAVE gone into hibernation
> with GRASS still up. Does dong so cause GRASS to "forget" the current PID,
> and therefore not be able to find the old ppm files to flush? Or is this
> just a bug in RC4?

AFAICT, the GUI removes the PPM/PGM files for a layer when you remove
the layer from the display, or when you close the map window.

One possibility is that the close event doesn't get handled on Windows
for whatever reason.

Can you conduct some tests:

1. Start the GUI from the command prompt, display a map, then quit.
Report any error message and whether any PPM/PGM files are left
behind.

2. Start the GUI, add some maps to the list, display them, remove some
maps from the list, check whether the PPM/PGM files are deleted as the
maps are removed.

--
Glynn Clements <[hidden email]>

--

Isaac I Ullah, M.A.

Archaeology PhD Student,
ASU School of Evolution and Social Change

Research Assistant,
Mediterranean Landscape Dynamics Project
***************************************************
[hidden email]
[hidden email]

http://www.public.asu.edu/~iullah
***************************************************

_______________________________________________
grass-windows mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/grass-windows

[Glynn Clements]

Isaac Ullah wrote:

> Glynn, I tried your tests with no replication of the error. Then I had an
> epiphany. I realized that all the mapsets that had the extra ppm files are
> mapsets where I was digitizing vectors from features in raster images using
> the tcltk v.digit module in RC3. I did a test using spearfish, and I was
> able to replicate the problem. When v.digit loads, it creates it's own ppm
> and ppg files for the background of the v.digit display (seperate from the
> ppm and ppg files created for the regular tcltk map display). When one exits
> from v.digit, those ppm and ppg stay.

That would explain it.

> When one exits from grass, the ppm and
> ppg files for the map display are erased, but NOT those created by the
> v.digit display. I don't have access to my linux box at the moment, but this
> could potentially be a cross-platform issue with the tcltk v.digit.

It's likely to be less of an issue on Unix, as the temp directory gets
cleaned when GRASS exits. Some of the cleaning code doesn't work on
Windows (it can't determine whether a file is owned by the current
user, or whether the process which created it is still running), so
the cleaning is less aggressive, only removing files which are more
than 4 days old.

> I assume
> the issue will be fixed in the forthcoming wxpython v.digit as AFAIK the wx
> gui does not write ppm files for it's display...

The wx GUI generates PPM/PGM files, but removes them when they're no
longer required.

> Shall I submit a bug ticket on this?

Please do.

--
Glynn Clements <[hidden email]>
_______________________________________________
grass-windows mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/grass-windows