SGML: Treatment of RS/RE

Subject: Re: Treatment of RS/RE

Date: 1 Oct 1996 12:50:09 GMT

From: davep@ACM.ORG

Newsgroup: comp.text.sgml


----------------------------------------------------------------------

In article <32503723.4B07@mit.edu>, Andrew Greene <amgreene@mit.edu> writes:
>I was under the (mis?)impression that REs were significant
>if a decision had to be made about whether the parser had
>encountered #PCDATA or an element.

>Yet from the exception in Charles's message, "element-start  nondata* 
>RE", I'd expect the RE to be ignored.

Unfortunately, it has to be recognized as data at one point in the
process, and later that _data_ character is ignored.  The first
happens before the check against the content model; the second
after.

I expect that the revision of 8879 now (but I hope not forever) in the
works will use a different parsing model that will not require such
an ignored RE to be counted against a PCDATA token in the content
model.  That'll help

Dave Peterson
SGMLWorks!

davep@acm.org