[gdal-dev] Possible bug in OGRSpatialReference:: importFromWkt()

[Frank Warmerdam]

On Thu, Jan 12, 2012 at 6:30 PM, Martin Chapman <chapmanm at pixia.com> wrote:
> Frank,
>
>
>
> I think there is a bug in OGRSpatialReference::importFromWkt().  I did the
> following:
>
>
>
> char* wkt = NULL;
>
> sourceSRS.exportToWkt(&wkt);
>
> targetSRS.importFromWkt(&wkt);
>
> CPLFree(wkt);
>
>
>
> It blows up on the CPLFree().  Why you might ask.  Because it looks like the
> importFromWkt() increments the pointer as it’s parsing it so by the time it
> returns the pointer you passed in no longer points to the beginning of the
> array of char* pointers. Instead, it points to the end of the memory block.
> Shouldn’t you be assigning a temp pointer to the start of the array before
> you start parsing it?  If you made it a const char** parameter this wouldn’t
> happen.


Martin,

The behavior is intentional though quite possibly ill advised.

The original intent was that the modified pointer after the import
would indicate what had not been consumed by the import
parser.  But in practice this has just make for a messy to use
API and generally speaking callers have to setup their own
temporary pointer for the import.

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 Software Developer