terms-of-use stored language specific

[Cantor, Scott]

> Is this intended behaviour? This applies to terms-of-use consent
> only, attribute consent is not affected. 

Attribute data isn't i18n-aware, we never implemented that (it's not standardized for one thing, though there are fairly evident ways one might do it).

I suspect what's missing is a third message property for each TOU that would override and control what got stored.

So instead of basing it on the message content (so that it invalidates if the message changes), it could be based on a separate "fixed" property value that would match across all languages, and could also be controlled to dictate whether a change to the message "matters" or not.

Seems like a useful RFE.

-- Scott