## LATEX-L@LISTSERV.UNI-HEIDELBERG.DE

 Options: Use Forum View Use Monospaced Font Show Text Part by Default Show All Mail Headers Message: [<< First] [< Prev] [Next >] [Last >>] Topic: [<< First] [< Prev] [Next >] [Last >>] Author: [<< First] [< Prev] [Next >] [Last >>]

Lars =?iso-8859-1?Q?Hellstr=81=F6m?= writes:

> OK, you cannot in general assume that the container for one key value is
> assigned before the container for another (this wasn't what I thought but
> it is probably logical if you think about what \DeclareInstance does). This
> fact is probably important enough to be stressed in the documentation (kind
> of "Note to people who write their own templates: You cannot make any
> assuptions about the order of assignments carried out by
> \DoParameterAssignments.").

oh yes (if it is indeed that way), but it would be worth checking if there is
a clear relation of a describable nature and if so that could be described
alternatively.

problem really is like this

k1 =n \key,
l1 =l \key2

then you can't do

k1 =12pt
l1 = \key + 10pt

in an instance declaration since the \key would not be defined until the
instance is used why l1 is evaluted when the instance is created

at least i think this is how it behaves currently, maybe that's bad and one
should \def or \edef all keys at declaration time.

(ignore for the moment that one shouldn't use \key in the instance since this
is internal to the template body so to speak; one could imagine something
like \keyval{k1} instead if one would go for this kind of feature --- one
problem with \keyval or the whole thing in generls is for which types should
it be allowed and with what semantics?  )

frank