Re: Proposal: remove “.” from namefollower

On Thursday 17 March 2022 12:36:00 (+01:00), Tom Hillman wrote:


What about rules at the end of a file?
The S is a separator, so there would be no requirement of space after the last rule.


Steven




Tom

_________________
Tomos Hillman
eXpertML Ltd
+44 7793 242058
On 17 Mar 2022, 11:33 +0000, Steven Pemberton <steven.pemberton@cwi.nl>, wrote:

Despite our recent resolution to accept this proposal, I will mention in passing that if we changed the top-level rule of ixml from
ixml: s, rule+s, s.
-s: (whitespace; comment)*.
to
ixml: s, rule+S, s.
-S: (whitespace; comment)+.
-s: (whitespace; comment)*.

(that is demanding at least one whitespace or comment after a rule), then we could retain "." in names.

This would invalidate no existing ixml grammar that I know of (except the ones designed to test the absence of spaces after rules).

Steven
On Monday 28 February 2022 14:23:21 (+01:00), Norm Tovey-Walsh wrote:


Hello,

On balance, I think we have to fix the ambiguity problem and the
simplest, most direct solution is to remove full stop from the
namefollower production.

Yes, “.” is a valid XML name character, but we already accept that our
names and XML names are not exactly the same. Restricting the space is
if anything less problematic than the fact that we allow characters that
can’t appear in XML names. But I think that’s a consequence of Unicode
character classes, so I can live with that for the simplicity of the
expression in namefollower.

I was concerned that we were picking up “.” from one of those classes,
but that doesn’t appear to be the case.

Be seeing you,
norm

--
Norm Tovey-Walsh
Saxonica