Squeak Socket Primitives

Craig Latta Craig.Latta at NetJam.ORG
Thu Nov 11 22:20:21 UTC 1999


	Damn; I'm violating my one-message-per day restriction. I should wake up even later. :)

> FWIW, all the documents about Berkely sockets on Win32 concur
> that one should use non-blocking sockets.

	References, please? I suspect people have said this due to an apprehension to using threads, either because of their previoius performance on win32, or because it used to be tricky.

> An aside about polling/blocking/non-blocking et al. To what degree
> can these issues be filtered up to the Squeak level? (Acutally, to what 
> degree *are* they filtered up :))

	Nail on head. They have indeed been filtered up, to the detriment of the system in my opinion. I feel a little weird saying this, since I'm usually the first to say "Do everything in Smalltalk!", but I think "use the right tool for the job" is more compelling here. In the back of my mind the whole time I'm thinking "Throw away the host sockets library and rewrite the IP stack in Slang!". This is followed by fits of laughter (I'm far too busy to actually do this, although it's entirely feasible).

> I.e., do we basically have the Berkely Socket api directly callable
> from Squeak? Then one should be able to decide whether to use 
> blocking or non-blocking sockets from Squeak, yes?

	It is directly callable. Unfortunately, the question of where to wait for events is inextricably related to performance and the simplicity of implementation and debugging.


-C


--
Craig Latta
composer and computer scientist
craig.latta at netjam.org
www.netjam.org
latta at interval.com
Smalltalkers do: [:it | All with: Class, (And love: it)]





More information about the Squeak-dev mailing list