Unlike with some other approaches, with “built-in” Phoenix localisation mechanisms when a translation is missing the msgid
is used. Seems nice, but is this good or bad?
- For example – do you supply all the “translated” strings also for default locale? Seems like it doesn’t make much sense because every
msgid
andmsgstr
is in fact the same string ANDmsgid
will be substituted anyway if one doesn’t supplymsgstr
, but maybe there are some gotchas? Any non-negligible performance penalties? - Finding missing translations might be more tricky as it is easier to omit this or that and still no exception is risen… how do you ensure that all the relevant strings are covered/translated?