[gdal-dev] gdalsrsinfo "ERROR - failed to load SRS definition"

[Even Rouault]

Piero,

> Dear list,
> (PREFACE: I was looking for a *gdal-users* list but I could see it does not
> exists; I hope I am in the right place)

You're in the right place. There's only one mailing list for GDAL users,
developers, followers, fans, etc... ;-)

> Any hint about the source of error?
> Is there some restriction on the CRS definition in the XML form or am I
> missing some parameter or is this a bug in the utility?

Nothing wrong on your side or in gdalsrsinfo, but looking at the code for
import/export support of the XML representation of SRS in the library, I can see
that it only supports the Transverse Mercator projection in import, and
Transverse Mercator and LCC 1SP in export. Here you have LAEA, hence the error.

There's also another error, due to the +a and +b parameter being equal, this
leads to an inverse flattening equal to 0. And in the code that deals with
import, there's a test that forbids 0 as inverse flattening. I believe this test
could be safely removed, as 0 inverse flattening is usually used for the sphere.
Or perhaps the cleaner way would be to export the semiMinorAxis instead of the
inverse flattening in that case.

Well, the conclusion is that XML support for SRS has currently a few
restrictions. So, unless you have strong reasons to use the XML representation
of SRS, I'd suggest using WKT instead.

Best regards,

Even