Re: Polymorphic luaL_checkudata?

[Peter Odding <xolox@...>]

Mark Edgar wrote:
> > You don't think it's bad style to add (get|set)_timeout to the file 
> > userdata
> > methods? Not trying to criticize, just interested in your opinion.
>
> Not as long as your new methods fail properly when called with
> non-pipes.
Fair enough :)

> If you want to integrate pipes so that they can be used
> wherever a file is expected (and this seems a good thing) then you
> must use the same metatable.
Must? You /are/ talking about code I'm writing, right? As of yesterday[1] my 
tofile() helper[2] checks for either of the file/pipe metatables. Does this not 
give the same functionality? If so, I have overlooked something...

> This is basically the same thing I've done with pipes in the "ex" API
> too, by the way.  It might be nice if there was also an API which did
> the same thing for network connections too.
Not sure I can make APR's sockets as easy as files and pipes but I'm planning to 
at least give it a try. I think it would really fit Lua to make APR's 
file/pipe/socket i/o interface obey the same interface *subset* (using separate 
but `compatible' metatables allows additional methods so I don't have to dumb 
down the interface just to make it compatible).

> Note that I know nothing at all about APR and its abstractions.  It
> might not make sense to attempt to unify the APR API and the Lua API.
I can always try! We'll see how it works out.

 - Peter Odding

[1] http://lua-users.org/lists/lua-l/2007-04/msg00324.html
[2] the name is really a misnomer now, obviously