[nilason]

Fix uncaught exception of PointView::getFieldAs.

Second part of Coverity Scan CID 1685510

(#7883 only fixed one part of the issue).

[echoix]

Pure code move, AI review only finds pre-existing behaviour that I don't judge it is worth it

Details

Reviewed PR 7887 (r.in.pdal: fix uncaught exception), a single-file, pure code-motion change in raster/r.in.pdal/info.cpp: the min/max scan over the reprojected point view is moved inside the existing try block so that pdal::PointView::getFieldAs exceptions are caught and turned into G_fatal_error instead of escaping as an uncaught exception.

The move itself is correct: the loop only reads reprojected values, no state is needed after the try, G_fatal_error is noreturn so partially-updated outputs are never observed, and no other control flow or caller (get_extent at line 94) is affected. The only cost is that the catch message ("Reprojection of the data extent failed") now also covers field-read failures, which is acceptable.

Findings (1):

• raster/r.in.pdal/info.cpp:187 — If the reprojected view is empty (PDAL's reprojection filter drops points it cannot transform rather than throwing), the loop never runs and the function silently returns the unchanged source-CRS extent as though it were reprojected; a post-execute() size check would surface this. Pre-existing behavior, but it now sits on the moved lines.