Jeff McKenna suggested (privately) that i try MS4W. He didn't really say why, 
and i suspected it wouldn't work because both OSGeo4W and MS4W seem to use the 
same version of the FileGDB driver (at least, both FileGDBAPI.dll files are 
identical).

However, MS4W is still using GDAL 1.9.1 (2012/05/15) while OSGeo4W is now using 
GDAL 1.9.2 (2012-10-08), and, running the same ogrinfo and ogr2ogr tests (via 
OSGeo4W) on GDB files, there are no more crash problems!

Thus, there seems to be a problem with GDAL in the change from 1.9.1 
(2012/05/15) to 1.9.2 (2012-10-08), when using FileGDB.

I'm very new here -- should i be reporting this elsewhere?

Martin Feuchtwanger  [email protected]  604-254-0361
On 06/11/2012 4:25 PM, Martin Feuchtwanger wrote:

Using GDAL-OGR 1.9.2 (2012-10-08) with the FileGDB add-on (not sure of the version, but i got it at the same time i got GDAL), there's a random crash bug upon using ogrinfo or ogr2ogr. I've tried the following on several different GDB files and there's almost always a Windows error message:
ogrinfo.exe has stopped working
    A problem caused the program to stop working correctly....
but it's not always at the same time during the processing.

    ogrinfo --debug on -al <gdb_file>
    OGR: OGROpen(<gdb_file>/02DAE9F8) succeeded as FileGDB.
    OGR: GetLayerCount() = <N>

It usually appears here, and no layers are listed, but sometimes they do get listed

    1: <layer1>     (Multi Line String)
    2: <layer1>     (None)
    :
    N: <layerN>     (Multi Polygon)

before the crash. Sometimes the crash occurs part-way through.
The same crash happens when using ogr2ogr to write a GDB file to Postgres:

    ogr2ogr -overwrite --debug on -f "PostgreSQL" PG:<DB_connection>
    <gdb_file>

    ogr2ogr.xe has stopped working
    A problem caused the program to stop working correctly....

I've seen references to using a patch to make 4GB available (instead of 2GB) for \Program Files (x86)\ programs http://ntcore.com/4gb_patch.php but the problem persists even for small GDB files.

I'm guessing it /is /a memory issue, though, because of it's varying behavior on my machine and, presumably, has not occurred on the programmer's machine.

BTW, the problem seems quite /independent /of the "Invalid function arguments." error resulting from the GDB having a non-standard CRS -- it happens with GDBs having a standard CRS too.

Is it possible the FileGDB driver is broken?
--
Martin [email protected]   604-254-0361  Vancouver, BC


_______________________________________________
Osgeo4w-trac mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/osgeo4w-trac

_______________________________________________
gdal-dev mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/gdal-dev

Reply via email to