[Gdal-dev] Failure in opening a pci disk file

[Frank Warmerdam]

N. Farah wrote:
> Ok Frank...An instance of the raster dataset is created and returned 
> even if the dataset is a vector one.
> Is this the desired behaviour ?
> Once you have a instance of the raster dataset returned one may assume 
> that this was a valid raster dataset. Posting an error or returning a 
> null pointer might be an alternative for non raster data.

Noureddine,

It is arguable whether it makes sense to successfully return a dataset
if there are no bands of raster data.  I don't mind what the PCIDSK
driver does now, but would agree that it would be helpful to the user if
it at least returned a warning if there were no bands.

Andrey, could you add that?

> Returning an instance of a non raster dataset would make probably more 
> sense (to me) if the data model of vector dataset was the same as the 
> raster one. OGR is using a bit different model to handle data (data 
> source and data layers instead of dataset and bands).

If we had support for PCIDSK vectors it would belong in OGR.  But it is
out of scope for the client who funded the PCIDSK development.

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent