Points of focus
Dan Sugalski
It's time to be looking towards a 0.1.1 release. There's been some
overhaul of the internals and fleshing out of some features, so I
think we're well-warranted to be thinking about another point
release. What I'd like to do this time is:
*) Get continuations all nailed down. There seems to be some
lingering problems in the system I'd like identified with tests and
fixed
*) Get lexical pad operations spec'd out and possibly working
*) Fix hash.c. (Though it may not be broken. Signs are good, though)
*) MMD vtable ops in bytecode
*) ICU building, if Jeff gets things integrated in time
Standard doc updates and bug fixes are in order too, I think.
--
Dan
--------------------------------------"it's like this"-------------------
Dan Sugalski even samurai
d...@sidhe.org have teddy bears and even
teddy bears get drunk
Leopold Toetsch
> Or something equally manager-speaky.
Ok all ...
> *) Fix hash.c. (Though it may not be broken. Signs are good, though)
which indications do you have that something is broken here?
leo
Tim Bunce
> Or something equally manager-speaky.
>
> It's time to be looking towards a 0.1.1 release. There's been some
> overhaul of the internals and fleshing out of some features, so I
> think we're well-warranted to be thinking about another point
> release. What I'd like to do this time is:
>
> *) Get continuations all nailed down. There seems to be some
> lingering problems in the system I'd like identified with tests and
> fixed
> *) Get lexical pad operations spec'd out and possibly working
> *) Fix hash.c. (Though it may not be broken. Signs are good, though)
> *) MMD vtable ops in bytecode
> *) ICU building, if Jeff gets things integrated in time
>
> Standard doc updates and bug fixes are in order too, I think.
Is IMCC method call syntax spec'd, implemented, and reasonably stable?
Tim.
Dan Sugalski
I'm not sure, but folks have been reporting odd errors when working
with lexical pads, ordered hashes, and hashes in general. It's
possible that this is all independent, or indicative of a problem
elsewhere, but they all share the hash.c code, so it seems a good
place to look.
Might be a red herring, of course, but we certainly would be
well-served by giving that code a good going over, as even small
speedups would give us a good boost...
Jens Rieks
On Wednesday 31 March 2004 23:27, Tim Bunce wrote:
> Is IMCC method call syntax spec'd, implemented, and reasonably stable?
I think yes. But the method (and sub) declaration stuff needs a bit work.
For example, you can not declare a "new" or "end" method at the moment.
> Tim.
jens
Dan Sugalski
Yes, yes, and yes. In that order. :)
Documented, on the other hand... that needs some work. But what's in
is simple and stable.
Jens Rieks
On Wednesday 31 March 2004 20:42, Dan Sugalski wrote:
> *) Get continuations all nailed down. There seems to be some
> lingering problems in the system I'd like identified with tests and
> fixed
> *) Get lexical pad operations spec'd out and possibly working
> *) Fix hash.c. (Though it may not be broken. Signs are good, though)
> *) MMD vtable ops in bytecode
> *) ICU building, if Jeff gets things integrated in time
Exceptions will be fixed later?
jens
Dan Sugalski
Yes. They still need some thought and work, and there's a limit to
what I want to deal with in one go. After 0.1.1 beat me up about it
and we'll get them fully specified and implemented, then we can
release 0.2.0 :)
Matt Fowles
There was some discussion a while ago about having a whole class of
array pmc, some for each type as well as some that auto expand and
others that don't. I am about to go on a trip and will thus have time
on the plane to implement such things; however, I do not recall any
official decision about what each should be called and which should do what.
If someone would provide me with this basic information, I will go about
implementing these things. Having looked at doing this before, I am not
entirely sure which vtable functions should be defined in these classes,
or if they should have an inheritance hierarchy of any type. But given
a little guidance I can put in all the grunt work here.
Matt
Leopold Toetsch
> *) Fix hash.c. (Though it may not be broken. Signs are good, though)
Adam Thomason (zhanks) sent me a precise description of one hash bug.
It occured during freezing big hashes--fixed.
leo