Re: Re: Jing and libxml2 is right, Trang is wrong

>
> It's done on purpose because
>
> <quote src="
> http://lists.oasis-open.org/archives/relax-ng/200301/msg00006.html">
>
> The reason why the latter is forbidden is that there's 
> a tendency to use it assuming it means
>
> <start>
>   <choice>
>     <ref name="foo"/>
>     <ref name="bar"/>
>   </choice>
> </start>
>
> </quote>
>
> Eric

Eric,

1) at the URL below, a patch for trang-20030619 that generates correct
patterns for start with implicit grouping.

http://ftp.davidashen.net/misc/trang-20030619-sig.patch

2) I tend to disagree with James. Either implicit grouping is not
allowed anywhere, or it is consistently allowed; 

start = foo,bar

is perfectly correct and normally used in nested grammars. Nested
grammars are very useful in conjunction with parentRef elements.

Trying to fix a hypothetic bug James introduced a real one.

3) In cases where 

start = foo, bar 

is a bug,

<start>
  <ref name="foo"/>
  <ref name="bar"/>
</start>

does not have a chance to be misused as it were a choice, because
this pattern is forbidden in top-level grammars.

Cases where

start = foo, bar

is correct is used in nested grammars; and in most cases nested
grammars are used by people who know what they are doing.

I think this artificial limitation should be lifted.

David Tolpin
http://davidashen.net/