Fix .raw files being detected as text or other formats

[masterAllen]

Reason

Opener.getFileType() identified files by magic numbers and an ASCII heuristic before looking at the .raw extension:

// ImspectorPro MSR
if (name.endsWith(".msr")||name.endsWith(".MSR"))
	return UNKNOWN; // Open with Bio-formats plugin

// Combined TIFF and DICOM created by GE Senographe scanners
if (buf[128]==68 && buf[129]==73 && buf[130]==67 && buf[131]==77
&& ((b0==73 && b1==73)||(b0==77 && b1==77)))
	return TIFF_AND_DICOM;

// ...

// RAW 
if (name.toLowerCase(Locale.US).endsWith(".raw") && !Prefs.skipRawDialog)
	return RAW;

Raw dumps have no header. The first bytes are just pixel data, so they can accidentally look like another format. For example, in low-light photography the pixel values are often small enough to fall in the printable ASCII range, so the original code could easily treat a .raw file as text.

Fix

In ij/io/Opener.java, classify .raw first (case-insensitive), before magic-number and ASCII checks.