VM source bug/missing call?

Dan Ingalls DanI at wdi.disney.com
Sun Feb 8 20:40:09 UTC 1998


>Hi Dan,
>Thanks for the explanation.  That's why I asked and didn't assume; I'm 
>fairly new to Squeak/Smalltalk and thought I should find out instead of 
>just dropping the call in ;-)  thanks.  Does this mean once Jitter is 2x 
>that you're going to rewrite the VM to get rid of the "warts"? ;-)

Well, it will be a judgement call.  If Jitter requires a lot more space, then the interpreter will continue to be important as a practical vehicle, regardless of how fast Jitter becomes.  My hope is that if we get Jitter up to 2x-3x, then it might run at 1x without a small cache, which would mean that the current interpreter would cease to be so important.  At that point it would be nice to rewrite it cleanly, for educational purposes if nothing else, and as a part of Ian's new hierarchy, thus reducing the current duplication of interpreter methods.

However, it must be said that my personal motivation for that project will probably not be as high as for other more forward-looking projects.  My hope is that someone else, possibly a professor with students interested in interpreter design, might do this and do a nice job documenting it as well.

	- Dan





More information about the Squeak-dev mailing list