[Debian GNUstep maintainers] Failed to load strings file - bad character encoding

[Yavor Doganov]

Paul Chany wrote:
> Say I save this file with C-x RET f utf6be-with-signatures and get
> this file:

Hmm, which GNUstep documentation/guide/article advised you to do this?
It's definitely wrong.

> - <NSException: 0x975da50> NAME:NSGenericException REASON:Parse failed

No wonder; this junk is really hard to parse ;-)

> How can one get the proper encoded strings file?

GNUstep expects the .strings file to be in US-ASCII only, or any
non-ASCII characters to be escaped, or in UTF-8 with a BOM marker.
Very annoying limitation; I think it is because of OpenStep
compatibility.

So, just edit and save the file as UTF-8 and insert a BOM at the
beginning.  With Emacs 23:

M-<
C-x 8 RET #xfeff RET
C-x C-s

If you don't put a marker, there is a runtime warning when the
.strings file is loaded, but TTBOMK it is correctly parsed and used on
GNU systems at least (assuming you have the corresponding locale
available, of course).

> Which editor and how can one use for this?

See my Organization: header, brother :-)