On 2007-05-02, at 11:23 EDT, Jim Grandy wrote:
On May 2, 2007, at 3:33 AM, P T Withington wrote:
On 2007-05-01, at 23:26 EDT, Jim Grandy wrote:
Comment: Should the synopsis box for LFC classes give their LZX
tag?
They do, in fact I look for the tagname class property -- e.g.
http://www.openlaszlo.org/lps4/docs/reference/LzNode.html. Did
you see a case where that wasn't so?
I was fooled by the fact that there is an `event` tag but it is
not implemented by `LzEvent`...
That's a gray area in the language, I think, or at least a gray
area in my understanding. Do you see the <event> tag as a
declarative statement which has the side-effect of introducing an
LzEvent object at runtime (which is how I've been viewing it), or
do you see <event> as a direct alias for LzEvent in the same way
that <view> is a direct alias for LzView?
Right. Only subclasses of LzNode have a direct mapping from tag to
class. `event`, like `attribute` and `method` are tags but are not
nodes.
On the one hand, <event> is handled by the tag compiler and used to
generate code that may or may not look simply like an allocation of
an LzEvent -- and it's important to maintain that flexibility, I
think. On the other hand, I presume there is an 'lz.event'
registered in the system as an alias for LzEvent.
Er, actually not. Since again, that is driven by node-ness (as a
kludge until we have a real namespace).
