[QGIS-Developer] QGIS 3.8 Transformations dialog

[Richard Duivenvoorde]

On 27/06/2019 10.04, Nyall Dawson wrote:
>> Is it the old one? Or a default one?
>> Because reading the texts there, it appears that the last item is to be
>> used?
> Well - (don't read this as snark, I'm just trying to accurately convey
> the situation): if you want to make the correct choice, fire up proj
> 6.1's projinfo tool and see what it suggest for transforming between
> the two systems. On QGIS < 3.8 or proj <  6.1, this dialog is using a
> custom database of "transformations", which was ad hoc, unreliable,
> and basically untrustworthy.
> 
> The good news is that the solution is already done -- you just need a
> QGIS 3.8 build based on Proj 6.1. And then you can trust QGIS'
> transformation capabilities as being the best there are available
> anywhere (thanks entirely to the proj team).

Ok, by accident I have a QGIS with proj6 here :-)

That one shows 'just' 2 options (added to the qgis.nl post)
IF the green line is just the first one, but we are not sure if it is
the right one... should we make it green anyway?

Other remark (again, see qgis.nl post): in the new transformation
dialog, I do not see the information that I saw in the other (proj5)
dialog, information like:

EPSG Transformation Code: 1672
Source CRS: Amersfoort
Destination CRS: WGS 84
Remarks: Parameter values from Amersfoort to ETRS89(1) (code 1751)
assuming that ETRS89 is equivalent to WGS84 within the accuracy of the
transformation.
Replaces Amersfoort to WGS84 (1) (code 1112).
Replaced by Amersfoort to WGS84 (3) (code 15934).

Is that info not available (any more?), or did we conclude that it was
not valuable? Because I think it gives some context about the validity
of the given transformations? I'd be ok to still show this (if available)?

Will have to compile proj6 again to make projinfo work to check (proj6
was patched to work with QGIS).

Plz do not see this as critiques: I understand that actually QGIS and
proj are getting better and better, and we just are getting more
precise! Thanks all for this work!

Regards,

Richard Duivenvoorde