Good point, I'll add a warning for that.

On Tue, Sep 14, 2010 at 10:17 AM, P T Withington <[email protected]>wrote:

> In fact, it surprised me that the compiler would not complain if it saw
> both a text attribute _and_ non-LZX content in a tag that supports text.
>
> On 2010-09-14, at 10:07, Henry Minsky wrote:
>
> > That sounds like a good heuristic
> >
> > On Mon, Sep 13, 2010 at 11:18 PM, P T Withington <[email protected]
> >wrote:
> >
> >> At least in this particular case, it's interesting that the programmer
> >> explicitly said `text=""` in the open tag.  I think we could say that if
> you
> >> explicitly assign the text attribute, whether in the open tag as is done
> >> here, or as an explicit attribute tag, then the compiler should _not_
> try to
> >> interpret the body of the tag as the text value.  That would seem to
> give an
> >> easy out for the case where you want to make a one-off text instance
> with
> >> methods -- you just specify the content as the initial value of the text
> >> attribute instead.
> >>
> >> How does that sound?
>
>


-- 
Henry Minsky
Software Architect
[email protected]

Reply via email to