Multilanguage sites

[OndrejH]

I see there was a discussion on handling multilanguage sites.

I had minor trouble what I consider (yet not sure) to be caused by Joomla core:
When using override language ini file created by fabrik, the file got deleted by Joomla after I touched the Language Manager and tried to add new label record.
I omitted this behaviour after (1) I had let joomla create override ini file by itself (i.e. got into conflict of joomla possessed owner privileges and web ownership privileges), and (2) changed the locking of file from LOCK_EX.
Can you maybe comment on locking trouble possible cause? Now everything work great and smooth.
I am really satifisfied with the tutorial and this solution.

I am pretty sure this problem with joomla overwriting language file did not exist when I tried to translate this way with earlier version of J few months ago.

 

[OndrejH]

Just to stay consistent, the tutorial mentioned above is this one (most of you may know it already):
http://fabrikar.com/forums/index.php?wiki/make-fabrik-multilingual-with-fabrik-itself/

 

[OndrejH]

One more little hint for non English users. Override language file must follow certain convention.
Such as (https://docs.joomla.org/Specification_of_language_files ):

The characters {}|&~![()^" cannot be used in keys. There are also reserved words for keys: null, yes, no, true, false, on, off, none. Note: the COM_EXAMPLE_ prefix is not mandatory as of Joomla! 1.6/1.7, but it is advisable to follow this convention to avoid naming collisions with other extensions. ...

If you start building your labels around character sets which would be used as language constant and converted from UTF-8 to ISO-8859-1(or UNICODE?) encoding and would include any of those prohibited characters, your override ini file would not get processed. Therefore it might be a good idea to start with labels in English and translate those ones. Though, this re-encoding problem does not necessarily apply to every (your requested) language and character set encoding.

e.g.:
Translate label "Greece" to Czech, i.e. en-GB to cs-CZ
GREECE="?ecko" - this is processed without error

If someone decided to start the opposite way, for example cs-CZ to en-GB (as in my 1st attempt), we get following error:
?ECKO="Greece" ... this becomes unrecognized as encoded as
??X
ECKO="Greece", this stops reading language file and you eventually end up with spending time on searching on the cause of problem.

I hope I saved some time to anyone who might be going through the same learning curve.