[WG: I18N & L10N] i18n fixes review request

Fri Apr 19 02:27:32 PDT 2013

Hi,

David Haines rencently commented in 
https://jira.sakaiproject.org/browse/SAK-21759:
> The approach used here allocates the collator in the compare method and that leads to unnecessarily high CPU usage when lots of data is involved. The allocation of the collator should be outside of the method. The implementation used in STAT-311 looks appropriate.

Could someone confirm the fix in 
https://jira.sakaiproject.org/browse/STAT-311 is OK?

If it's confirmed it's OK, could someone improve the fix for SAK-21759 
with an implementation similar to STAT-311?

Cheers,

J-F