[gdal-dev] GeoPackage fails after touching it with Spatialite-gui

[Even Rouault]

Jukka,

I highly suspect that the ALTER TABLE must imply a rewriting of the file by
spatialite/spatialite-gui, and when doing so, it doesn't preserve the
"application id" (4 bytes in the header of sqlite file), that geopackage
specification mentions to be set to a particular value. Consequently the
geopackage driver later fails when checking the signature.

A few inline comments below

> Hi,
>
> There are no GUI-tools for managing GeoPackage databases at the moment and
> because after all they are only SQLite db files just like Spatialite files I
> thought it would be a good idea to do some maintenance work for GPKG
> database with Spatialite-gui.
>
> I could do a few things with Spatialite-gui without troubles: create a view,
> create an index, add column. But then I renamed one table and after that I
> cannot use the DB file with GPKG driver any more. SQLite driver does open it
> but db is unusable because the geometry BLOBs are different. I have received
> two king of error messages from ogrinfo and also a sudden program crash. One
> error message is:
>
> ogrinfo test.gpkg
> ERROR 1: bad application_id on 'test.gpkg'
> ERROR 1: bad application_id on 'test.gpkg'
> FAILURE:
> Unable to open datasource `test.gpkg' with the following drivers.

Yes, that's expected if the application id has been changed. That one doesn't
cause a program crash, right ?

>
> Another message is:
> ogrinfo test.gpkg
> ERROR 1: (null)
> INFO: Open of `test.gpkg'
>       using driver `GPKG' successful.

--> Hum, "ERROR 1: (null)" means a NULL pointer used as a string somewhere and a
potential source for crashing.

>
> To reproduce the error create a GeoPackage with GDAL, open it with
> Spatialite-gui and rename one table. I know it is perhaps not a good idea to
> rename a GeoPackage table with "ALTER TABLE ... RENAME TO ..." because it
> breaks the integrity, but on the other hand, because it seems to be possible
> to do so without any triggers and constraints preventing the operation and
> the result is an unusable database it could be worth having a look on what
> happens.
>
> -Jukka Rahkonen-
>
>
> _______________________________________________
> gdal-dev mailing list
> gdal-dev at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/gdal-dev
>