[tei-council] rule contexts for attr constraints

[Syd Bauman]

> Should be */@attname, I assume. I am surprised its not, I thought I
> had fixed that.

Uhh... I'd say elementname/@attname, no? Don't want a constraint
written for div/@type firing on name/@type.

I plan to tuck this fix into the other work on extract-isosch I'm
doing.

> > If a <constraintSpec> is a child of <attDef> and its child
> > <constraint> does *not* have a <sch:rule>, i.e., the path is
> > elementSpec/attList/attDef/constraintSpec/constraint/sch:assert,
> > then the context= that extract-isosch.xsl generates is the parent
> > element. This seems plain wrong.