Taxonomy module: Fix form maxlengths

-1. We keep going back and forth on this. We've got one group of people (webchick is part of 'em, I think) that keeps creating patches to standardize all the form inputs to the same number (64 or similar) and then another group of people who goes around and standardizes them back to database column length. That neverending cycle just isn't working anymore, and more substantial research needs to be done on the right, and final, approach.

Shouldn't the length of database fields and input text fields now be the same even for multibyte languages after we switched to uft-8 for the database? What is mysql's notion of a character wrt uft-8?

Listen closely to what killes asked: it appears, and this is certainly news to me, that part of the reason for the disparity was that databases stored multibyte languages as MORE than one character - so a five letter username in a five length field would work, but not if one of those letters were multibyte (six characters in a five length field). Anyways, the 64 *is* attempting to be consistency - that was the whole goal of that previous patch (whom I've yet to find) - to standardize all the maxlength's of all FAPI textfields to one length, being 64.

Personally, I do agree that maxlength's should be set to the length of the database column, but I've just seem these patches go in, get undone, and then get re-patched again, so many times that I'd rather have some demonstrably "look, this is why we do, and don't you dare change it". And for that, I need a bit more "concrete" proof than just "standardization". If someone could investigate killes' comment, we'd be able to say "In the past, the disparity was due to multibyte languages. Nowadays, however, we match column length to maxlength in every circumstance." Until that happens, we're just going in circles.

I'll be refraining from comment on your other patch - while I agree it's needed, I don't agree it's needed RIGHT NOW. I'd much rather have proper research done and have *all textfields and all column lengths* reconsidered in one central issue/patch. Until then, we're just gonna have all these minor patches, like yours, getting in, and then suddenly they'll be patches like "Username doesn't respect 64 maxlength standardization". The same ol', same ol', all over again. This is the circle.

Fixed in CVS/HEAD.