[GRASS-dev] grass6.2?

[Glynn Clements]

Michael Barton wrote:

> >> I noted the difference because i.rectify reads from the current location and
> >> writes to the location specified by i.target.
> > 
> > Ah. That is something which its replacement shouldn't imitate.

I had forgotten that i.rectify itself doesn't need replacing, only
i.points.

> > Reading from a different location or mapset is reasonable enough, but
> > nothing should write outside of the current mapset. The case of
> > r.in.gdal creating a new location is probably OK, but writing to an
> > existing mapset (other than the current one) isn't.
> > 
> > BTW, does i.rectify bother to lock the destination mapset?
> 
> Is there a reason to change i.rectify at some point so that it also will
> read from a different directory and write to the current one? This is how
> the georectifier works, but I have to switch back and forth between
> locations/mapsets to make i.rectify work.

We probably need to keep the historical behaviour, but it should
provide the option to rectify /into/ the current mapset.

Without having looked at the code too deeply, I would have thought
that it should be possible to add location= and mapset= options to
i.rectify (analogous to [rv].proj) and to use the current mapset as
the target if the location= option is given.

Presumably, many of the same issues apply to i.ortho.photo, although
that really needs to be completely re-written anyhow.

-- 
Glynn Clements <glynn at gclements.plus.com>