This patch is to add support for the proposed addition of tags to the OCS standard version 1.7, as seen in T6133
It is supposed to:
| Automatic diff as part of commit; lint not applicable. |
| Automatic diff as part of commit; unit tests not applicable. |
+1 looks sensible to me.
What's the OCS state in this regard? When will 1.7 be a thing?
Sweet :)
What's the OCS state in this regard? When will 1.7 be a thing?
OCS 1.7 will be a thing hopefully in the not too distant future... Incidentally, i need to regain write access to the fdo wiki, where said standard is hosted, as i lost that when they changed all the access methods in... some long time ago.
Ping, if we could get this merged soon, it would be terribly handy :) (it's the basis for moderating the ghns content, and needed for the knewstuff patch in D6513)
T6133 suggests that tags are formatted as "group##key=value" or something similar. Wouldn't it make sense to handle parsing that format here as well? Or are tags more intended as exact matches?
The group is more a convenience concept than anything else - i think it'd be nice to have something to do something with that, but not in this patch (the "something" needs a more proper definition, and i'd rather not have this functionality, which we need for the curating work that's being undertaken on the store, wait any longer than we have to...).
To answer the specific question, during the matching phase the group##key is supposed to be considered canonical, and the groupings are more for pairing of similar content when putting things together. As a random sort of idea, one might imagine having something like a validator which matches the concept "must have something defined in group x", but that's why i've done the validators the way i have, there's nothing stopping us from putting that in when we have a purpose for the functionality :)
Right, so leave it to application users how to deal with that, maybe later on add some convenience functions. Sounds sensible. :)