rawTherapee/rtdata/languages

This is the directory where all translations should go.

Translations are loaded for a given term at three levels:

  1) default
  2) <Language>
  3) <Language> <Locale/Variant>

Developers who are adding a new feature should add new strings *only* to 
default.  This file should be comprised of basic English text.  It will be used 
in the event that there are no more specific languages specified.  Once you 
have modified default, you should run ./tools/generateTranslationDiffs (Bash 
script) which will re-generate the localizations with commented out additions 
which you have just added.

Translators should in general implement the <Language> file.  This is the 
generic translation for a given language; for instance, 'French', 'German', 
'Norsk', etc.  If a string exists in this file (and the user has specified this 
language), then RawTherapee will override the value in default with the value 
in <Language>.  Please note that the filename for this file must not contain 
any spaces.

In some situations, translations may differ based on region, locale, etc.  A 
good example of this is the difference in spelling between 'color' (American 
English) and 'colour' (British English). In this case, the vast majority of 
strings are identical between English and English (UK); however, to keep the 
proper spelling in Britain, we have a locale file called 'English (UK)' which
contains the differences between the two.  RawTherapee uses locale files when:
  a) The user has selected a language which has a space in the file name
  b) There is another file which is identical to the locale file up until the 
     space (i.e., 'English' to the locale file 'English (UK)').

If a locale file is used, it is applied in the same manner as <Language> is to 
default.  The locale will override any keys present from the ones in the 
language file (and in turn, the default).

After the generateTranslationDiffs has been run, all untranslated terms for 
a given language/locale will exist at the end of the file, prefixed by a ! 
comment marker.  Translators should go through this section of the file and 
translate all terms which they can. After you have translated a line, just 
remove the ! comment marker.  Comments may be included using the #xx comment 
marker, where xx is a numeric prefix used to make sure automated sorting keeps 
comments in the right order, e.g.:
  #00 Comment line 1...
  #01 Line 2...
  #02 3, etc.

Metadata for the language file appear in comments with the following syntax:
  #10 @KEY=VALUE
The @ character must appear immediately after the whitespace following the
numeric prefix. KEY is the metadata name and VALUE is value for that metadata
name. The following key(s) are recognized:
  a) LANGUAGE_DISPLAY_NAME: The language name as shown in preferences.

To create a file with only Latin characters from a non-Latin one, you can use 
sed with the "y" command. For example, to create a latin-only "Polish (Latin 
Characters)" file from the non-latin "Polish" one:
  sed 'y/ĄĆĘŁŃÓŚŹŻąćęłńóśźż/ACELNOSZZacelnoszz/' < Polish > "Polish (Latin Characters)"

You can use this Wikipedia "Character sets" category page to help you find all 
the characters in the language file you want to convert into Latin-only:
  http://en.wikipedia.org/wiki/Category:Character_sets

To convert all line terminators in all language files to CRLF (dos/mac/unix)
you can use vim:
  a) cd rtdata/languages
     vim
  b) In vim, type:
     :set ffs=dos
     :args *
     :argdo w
  c) vim will process all language files. Once done, you can close it:
     :q