The NeoMutt program is written in English, but it has been through a process of Internationalization and Localization.
Internationalization (abbreviated to i18n):
Localization (abbreviated to l10n):
The translation team is currently managed by Rich Russon (flatcap).
It would be a great help to the project (and Rich) if someone volunteered to help with this.
NeoMutt builds upon Mutt, so it has inherited its translations. Mutt hasn’t changed much (in years), so most of them are still correct.
Here are some recent statistics for the Swedish translation:
|1321||Total number of strings to be translated|
|219||Fuzzy – need checking by a human|
|104||To do – need translating|
The “fuzzy” translations appear when the
msgfmt tool finds a translation that
it thinks is correct. These translations are often wrong.
No, it’s quite simple – you just need to be multi-lingual.
The translations are stored in “Portable Object” text files (.po)
Here’s a sample from the Japanese translation (ja.po)
#: main.c:905 msgid "Mailbox is empty." msgstr "メールボックスが空。"
There are plenty of tools to help you with the translation process. One of them is “poedit” – it’s quite simple to use. This screenshot shows some translated, some fuzzy and some untranslated strings.
The translation mechanism uses GNU’s “gettext”. Here are some of the relevant parts of their docs:
If you’re technical you could look at the NeoMutt source code:
git clone -b translate https://github.com/neomutt/neomutt
NeoMutt has new releases about once a month. Before each release, the
translators will be emailed and asked for their latest
.po file. This will
be merged into NeoMutt and an updated copy will be mailed back.
There’s a pretty good editor for translation files: ‘poedit’, but some of us are only really happy at the command line.
Here are some hints to improve your translate-fu!
This tool will extract a subset of translations, e.g. fuzzy or untranslated:
msgattrib --no-wrap --only-fuzzy de.po > de-fuzzy.po msgattrib --no-wrap --untranslated de.po > de-untrans.po
Now you can focus on just bits that need some work.
This is a powerful search tool. You can find regexes in the msgid, msgstr or comments. Recently, I wanted to examine the overlap between the IDN messages, so I used:
msggrep --no-wrap --msgid -e IDN fr.po > fr-idn.po
Now that we’ve finished working, we want to re-combine the files. We can use
msgcat to merge them:
msgcat --no-wrap --use-first de-fuzzy.po de.po > de-new.po
Note: Some reordering might occur, so
--sort-by-file might help.
To check the validity of your translation files, and to get some stats, we can use
msgfmt. We don’t the compiled message file though.
msgfmt --statistics -c -o /dev/null fr.po 1087 translated messages, 185 fuzzy translations, 74 untranslated messages.
Ideally, the L10n comments should be enough to tell the translator everything they need to know. Sometimes, though, looking at the source can make things much clearer much more quickly.
Each of the tranlations is preceded by a location line:
#: compose.c:118 compose.c:970 send.c:292
Vim users can use the
gF (goto file) command which will open the file under the cursor and goto the exact source line you want.
Here are some tips from @mgedmin.
Find the next untranslated message with
\n\n avoids false matches from multiline translations, the
/e position the cursor between the double quotes.
This can be wrapped up in a custom command: Find untranslated:
command! FindUntranslated normal /msgstr ""\ze\n\n/e<cr>
Find untranslated or fuzzy:
command! FindUntranslatedOrFuzzy normal /msgstr ""\ze\n\n\|^#.*, \zsfuzzy\>/e<cr>
command! -bar PoStats echo system("msgfmt --statistics -c -o /dev/null -", bufnr("%"))
Finally a function to remove the
fuzzy flag: ftplugin/po.vim