Character set usage in structural elements

[Kevin Hawkins]

Currently our spec is a little vague about the character set we can
use in structural element values - the values in the headers and the body
section names for example.

We have a restriction on the character set for the name which reads
as follows..

"This keyword uses alpha-numeric printable characters a..z (a through
z)
0-9 (zero through nine), _ (underscore), - (minus) and, additionally,
embedded _ (space)."

Was it implicit that his should apply to the structural elements too
?

Eg the source address ??

I ask because it would appear the following is possible

Source=ukusa.awkward.[;&,£.usage

I would like to restrict the character set as above in these
elements - does anyone have any thoughts - or did they already believe it
was ? This would apply to everything within the header and the body section
names. We would still permit the '*' and '>' wildcard in target
addresses.

This has come up as we translate addresses to routine names in VB
(auto match of a source address to a subroutine) as the extra characters
are not allowed. A '/' was not handled correctly as it appeared in an
existing xAPp source address. It was auto generated as an instance value
from /EGNM

One issue I see this might effect is the indexing of body names
using [1] [2] or (1) (2) etc - maybe the brackets should be allowed and
could be transposed into an allowable character elsewhere as special cases
-
as the whitespace character might have to be. Note that I think they are
not
currently permitted in the names either light(1)=OFF is not legal :-(

Kevin