[Squeakfoundation]re: Network stuff

[Craig Latta]

> I'm in favour of anything that cleans up the ugliness around
> filenames, sockets, streams etc.

	As I recall, Michael's changes are at a higher level, mostly for
"session protocol" support (given streams already, how to speak HTTP,
etc.). They do continue to use the current low-level external-resource
support, though. Flow does change that lower-level support, so there
would need to be some further modification of Michael's work in a system
in which, e.g., Flow is the only low-level external-resource support
available.

> I also believe (maybe incorrectly) that Flow requires some thread
> capability in order to function propelry.

	That's correct currently. One could modify it so that weren't the case,
but I have no interest in that myself.

> Since I can't support threads on RISC OS...

	What's the ETA for the RISC-OS POSIX threads project?

> ...and I don't think one can guarantee thread support on embedded
> OSs...

	Some do provide that support. I don't think it's appropriate to lump
them all together as "providing no guarantee". Can you cite some
specific examples of embedded operating systems that don't provide
thread support?

	I think Michael's changes make sense now, regardless of the Flow
schedule.

	I also think all the packages we've mentioned so far should appear
optional to the user. If it seems like something will "rot" because it
is merely available from SqueakMap and not in a release snapshot, then
there's something wrong with the SqueakMap system (and we should fix
it). None of this stuff should be in a release snapshot.


	thanks,

-C

--
Craig Latta
http://netjam.org/resume
craig at netjam.org
Smalltalkers do: [:it | All with: Class, (And love: it)]