Problem opening PNG templates - new problem in v9.5

[MartinHore]

Steps to reproduce

1. I have LiDAR derived and other templates in PNG format together with associated PGW files with geo-referencing data which is set to align with the UK OSGB grid. I geo-reference my maps with CRS set to "Local" for preference. In v9.4 and earlier these templates reliably opened correctly positioned. This is what I want.

Actual behaviour

I've just upgraded to v9.5. These templates are now opening centered on the map origin and enlarged by a factor which depends on the map scale: 10 times for 1:10000, 4 times for 1:4000 etc. This is not what I want. It looks like OOM v9.5 is ignoring the PGW file for some reason. SHP Shapefiles still open correctly, and there is not a problem if I geo-reference the map to EPSG 27700. But geo-referencing to "Local" keeps my templates and map upright which I prefer. It looks like one of the updates in v9.5 has had an unintended consequence.

Expected behaviour

Tell us what should happen instead

Configuration

Mapper Version: 9.5 Operating System: Windows 10

[dg0yt]

Thank you very much for reporting this issue.

I geo-reference my maps with CRS set to "Local" for preference. [...] templates are now opening centered on the map origin and enlarged by a factor which depends on the map scale [...] It looks like OOM v9.5 is ignoring the PGW file for some reason.

IIUC there is a problem with existing maps (georeferencing "local") and their PNG templates (with EPSG 27700 world file)? I normally don't test this configuration where geospatial data including GPX tracks is ubiquituous. But it doesn't mean it should break.

SHP Shapefiles still open correctly.

In any case, Shapefiles are vector, not raster, thus using a very different code path. And they do include the CRS. PGW files do not.

there is not a problem if I geo-reference the map to EPSG 27700.

This is supported and tested, even with EPSG 27700.

But geo-referencing to "Local" keeps my templates and map upright which I prefer.

I don't think this is related to choosing between EPSG 27700 and "local". It is a matter of setting a (fake) declination in order to force grivation to zero.