Wrong layer CRS recognition ( the 'Hartebeesthoek94' issue)

[borysiasty]

Unfortunately r15361 have introduced one huge problem: all layers with correct WGS-84 prj definition are recognized as the 'EPSG:4148 - Hartebeesthoek94' crs. It can be really confusing and annoying for users and I believe we can't release 1.7 with it. Especially if it will be the last release of the 1.x series.

What about trying to find the exact definition and then fallback to the less restrictive comparision?

[cgsbob]

When I first encountered this problem, I was told that I can use

OVERRIDE_PROJ_DATUM_WITH_TOWGS84=NO 

to bypass this problem. Take a look at  http://trac.osgeo.org/gdal/ticket/3450

[borysiasty]

I only mean that all the files existing around, that was recognised as well known 4326 should be still recognised as it. With all due respect for EPSG:4148, it's unknown for most users.

[jef]

please retry with r15502.

[borysiasty]

Seems to be fixed for OGR layers. Rasters are still affected.

[jef]

Replying to borysiasty:

Seems to be fixed for OGR layers. Rasters are still affected.

please try with r15507 - unfortunately I don't have a good set of rasters to test with - but it correctly picks EPSG:32633 for tests/landsat.tif...

[borysiasty]

Seems it's fixed now. Thanks!

[borysiasty]

oooops :( Still the same with this file:  http://trac.osgeo.org/qgis/attachment/ticket/3508/l2.zip

[jef]

one more take in r15527.