Things are not so simple.
This is really a DaveB change so he should take a look. He had made the change in the noquote parse procs for @, but not the "normal" one. In other words he changed things so @x.y.z;noquote@ would be recognized by not @x.y.z@.
He made this change because there was code using vars of that form which had snuck in and weren't being recognized. I added the change to the second parse proc because there was also code that used vars of the @x.y.z@ form as well as @x.y.z;noquote@. Plus, as should be obvious I hope, both forms should except the same syntax for vars.
The one line rollback will ...
1. Break some existing code, though unfortunately I don't remember where (might be bug-tracker, template list builder, or forums since that's where I've been spending time recently).
2. Will cause the syntax for vars accepted by "noquote" to differ from the syntax accepted for un-noquoted vars, which is, of course, absolutely BOGUS.
I suggest Dave remind us as to exactly why he needed to allow @x.y.z@ style vars in the first place and maybe we can figure out a way to fix this that really works in all cases - or we can rewrite the code that caused Dave to make this change in the first place.
