[Date Index][Thread Index]
[Date Prev][Date Next][Thread Prev][Thread Next]
Re: c) or d)
- From: Stefan `Sec` Zehl <nospam@thanx>
- Date: Sat, 27 May 2000 22:58:37 +0200
On Sat, May 27, 2000 at 07:47:16PM +0200, Denis Barbier wrote:
[...]
> > > > > 6) <BLA <BLB>></BLB>></BLA> or <BLA <BLB>></BLB></BLA>
[...]
> > I don't think so. It would be violating POLA. A tag should close with
> > the first 'available' >.
> What is POLA?
"Principle of least astonishment"
> > <BLA <BLB>> Is clearly BLA with attribute <BLB> no matter what follows.
>
> I strongly disagree. In my mind innermost tags must be closed first.
> So here is another question, which one is preferred?
> c) first open, first close
> d) last open, first close
Oh, so now I understand your way of parsing that line.
If anything, this makes a strong point of disallowing 'stray' >
altogether.
> My preferred one is (d) (and to be honest i never thought of (a) before)
> because it prevents nesting problems, e.g.
> <define-container bla>%body</define-container>
> <bla><blb></bla></blb>
> is valid with (c).
Why should it be more or less valid in case c and/or d? If both are
containers, this line is simply incorrect. Either you start matching the
<bla> with the </bla> and notice that <blb> is not closed. Or you Start
with <bla>, recurse into <blb>...</blb> and notice that <bla> is never
closed.
> And it allows
> <define-tag foo>
> Before definition: <bar>
> <define-tag bar>
> baz
> </define-tag>
> After definition: <bar>
> </define-tag>
Hm. Thinking my way the first </define-tag> would end the outmost
<define-tag>. Yes. I can see your point. I agree that d) ist most
probably better for this :)
CU,
Sec
--
UNIX for the masses: <!--#exec cmd="/vmunix.cgi" --> -- Seen in the sdm.
______________________________________________________________________
Website META Language (WML) www.engelschall.com/sw/wml/
Official Support Mailing List sw-wml@engelschall.com
Automated List Manager majordomo@engelschall.com