We are using Globalize for some time now, but only for messages without plurals, so we only needed supplemental/likelySubtags in our application.
Now that we want to use all modules, we also need to add the CLDR data.
And we discovered that loading all required cldr datasets ( supplemental/likelySubtags, supplemental/numberingSystems, supplemental/plurals, supplemental/ordinals, supplemental/currencyData, supplemental/timeData, supplemental/weekData, main/{locale}/numbers, main/{locale}/currencies, main/{locale}/ca-gregorian, main/{locale}/timeZoneNames, main/{locale}/dateFields, main/{locale}/units )
We end up with 150K for the main data and 150K per locale. This is quite big and we feel it's not optimal.
Is there a way to reduce this ? strip some parts of these files that are never used by globalize anyway ?
We tried Globalize compiler, it doesn't seem to help for that problem as we need to provide it our CLDR data, and it seems it's not able to strip out what isn't actually needed.