Load does not correctly recognize some raw files on MERLIN and Linux machines

[Alex Buts]

if one tries to load some merlin file (e.g.) MER14336 using generic Load algorithm, the file is processed by incorrect loader (LoadPDFgetNFile)

The problem is observed on RHEL6 and Ubuntu and with some raw files (not all of them)

LoadRaw works correctly but generic load dialogue does not looks right.

[Martyn Gigg]

So a quick look at the code suggests to me that the problem is with LoadPDFgetNFile's quickFileCheck & fileCheck methods. It opens the file and starts trying to read it line-by-line looking for a line starting with #L". On linux there is no distinction between text & binary data from the point of view of fstream so I think it just gets unlucky when some files have a sequence that looks like "#L" in the file.

I would think the fix would be a proper check of whether the file is ascii or not. See LoadAscii::isAscii method.

[Nick Draper]

Moved to r2.6 at the end of r2.5

[Martyn Gigg]

My changes to the way the loader selection works have fixed this so it's no longer valid.

[Martyn Gigg]

This has been fixed as part of #7263 and should be tested after that has been verified and merge to master.

[Samuel Jackson]

This needs to wait until ticket #7263 has finished testing.

[Gesner Passos]

The problem has 'disapeared' ;)

[Stuart Campbell]

This ticket has been transferred to github issue ​7442