[Gdal-dev] IdrisiDataset::SetGeoTransform

[Mateusz Loskot]

Ari Jolma wrote:
> Mateusz Loskot wrote:
>>
>> Even if the prototype of IdrisiDataset::SetGeoTransform() does not
>> explicitly inform about constness of input parameter,
>> I think it's safe to assume the following:
>>
>> IdrisiDataset::SetGeoTransform( double const* padfGeoTransform );
>>   
> 
> I'm not being very clear I believe. In my eyes the fact that
> SetGeoTransform does not change the adfGeoTransform property (note *adf*
> not *padf*) of the object looks like a bug. It only changes some
> NameValues (?). When you ask the transformation parameters from the
> object using GetGeoTransform, it does not query these NameValues but the
> adfGeoTransform, which has not been changed. adfGeoTransform is
> initialized if a new dataset is created and it is filled (from the
> NameValues) in Open method, but not if SetGeoTransform is called.

I think Frank has explained it in previous post (bug).
Also, I have to admit I wasn't sure if it is a bug or not, but sounds
suspicious.

>>> Then, I see the buildbot is running a csharp-test. Maybe it would be
>>> a good idea to run the Perl-bindings test? Regards,
>>
>> Ari, please send me instructions how to configure/run Perl bindings test
>> and I will add it to the buildbot, no problem.
> 
> in swig/perl and in Linux (can be done in Windows but it is really not
> so nice since it is difficult to delete the temporary files the test
> creates and it is much slower also):
> 
> make generate
> make build
> make test
> 
> The last step returns a report saying that 3 out of 1301 tests failed
> (the Idrisi set/getGeoTransform tests) and all ogr tests were ok. The
> report is standard Perl format.

Ari, thank you I will try to add it today.

Cheers
-- 
Mateusz Loskot
http://mateusz.loskot.net