I know that the Windows Vista grass users community is not so wide, anyway I would like to know if someone is thinking about the problem I posted sometime ago and if someone else can reproduce it.
Whenever I try to create a new Location (in my case from epsg code) g.proj.exe crashes and returns the following message:
“child killed: segmentation violation”
and Grass returns, obviously, to it’s startup situation, waiting for the location/mapset selection to open…
I don’t know Windows system so well to debug it by myself. It would enough, for now, to know if someone else have faced it.
Giovanni
PS: I attach the picture with the Windows crash message
(attachments)
On Tue, 17 Mar 2009, G. Allegri wrote:
I know that the Windows Vista grass users community is not so wide, anyway I
would like to know if someone is thinking about the problem I posted
sometime ago and if someone else can reproduce it.
Whenever I try to create a new Location (in my case from epsg code)
g.proj.exe crashes and returns the following message:
"child killed: segmentation violation"
and Grass returns, obviously, to it's startup situation, waiting for the
location/mapset selection to open...
I don't know Windows system so well to debug it by myself. It would enough,
for now, to know if someone else have faced it.
OK I've eventually got OSGeo4W installed on a Vista machine (no idea if it's Vista related though) and tried it and I can reproduce the problem. It's not required to attempt to create a new location, simply
g.proj -p epsg=4326
is enough to reproduce the crash.
g.proj -w
prints the projection information in WKT format and *then* crashes. I'm guessing the problem is somewhere in the linkage with GDAL/OGR. Perhaps even related to the issue of freeing pointers with the correct function that was used to create them, that was reported a couple of months ago. But I thought we fixed that.
The next step is to compile GRASS from scratch on Vista to see if the problem is in GRASS or related to the OSGeo4W packaging. That could take a while but it needs done so hopefully it will get done.
Interesting for now would be if someone could test the same commands in an OSGeo4W install on a Windows-XP machine.
Paul
The next step is to compile GRASS from scratch on Vista to see if the
problem is in GRASS or related to the OSGeo4W packaging. That could take a
while but it needs done so hopefully it will get done.
I hope someone can do it (Marco Pasetti?), because I have little
experience on Windows (I'm mostly a Linux user).
Interesting for now would be if someone could test the same commands in an
OSGeo4W install on a Windows-XP machine.
AFAIK it works fine on Windows XP as stated by Michael Barton [1]
[1] http://n2.nabble.com/Re%3A-osgeo4w-on-Windows-Vista%3A-g.proj.exe-crashes-when setting-the-initial-location-td2458405.html#a2458405
I’ve finally compiled a basic grass-6.4.0RC3 on Windows Vista from scratch (also libraries), without wxpython and geos (I have to solve build errors). Now g.proj.exe doesn’t crash, so it seems a problem related to the actual OSGeo4W stack…
IMPORTANT: I forgot to say that I've built gdal with MinGW, but I
suppose that in OSGeo4W it's built with MSVC, right?
A curiosity: has grass been built against the MSVC gdal?
2009/3/19 G. Allegri <giohappy@gmail.com>:
I've finally compiled a basic grass-6.4.0RC3 on Windows Vista from scratch
(also libraries), without wxpython and geos (I have to solve build errors).
Now g.proj.exe doesn't crash, so it seems a problem related to the actual
OSGeo4W stack...