Re: stackless question
"Mark Hahn" <[email protected]>
| Newsgroups | gmane.comp.lang.prothon.user |
|---|---|
| Message-ID | <[email protected]> |
Christian Tismer wrote: > I will have a look into your source, to see how stackless > it really is. You finally realize that, when you try to > switch and find out that something is still hindering you > > The question is: Where do you need to jump at what time? > What is the set of your possible jump targets? > These must be stackless, that means, whenever a jump > to a certain object is allowed, it must be able to do that > in the current context. The C stack must either be cleared, > or in a state that makes this equivalent to being cleared. > > Well, I'll have a closer look and tell you. I haven't implemented any of the Thread or ThreadMgr (channel) stackless stuff so there is only normal function calling right now. There won't be much to see right now. I'm still waiting for everyone to comment on the proposal. I'm feeling pretty good about the proposal right now though since you approved my other email. >> P.S. I've been thinking of looking at using Parrot. Do you know >> anything good or bad about Parrot? > > Yes. I had a look, but it was about two years ago. Later, I took > my bad impression from hearsay. I was looking for a good VM > for my own Python, looked into Java VMs for a long time, and > also into Parrot. It turns out to be a quite bulky thing, > with lots of special structures just to make Perl fast. > The time I looked, there was just one developer. This might have > changed. > But I don't support people who say "we support any scripting > language, provided it is Perl". (You know the Model T gag? > "You can have it in any color, provided it is black".) They have four banks of registers and only one is objects. Their bytecodes work directly with non-object primitives. I think three-fourths of Parrot would be wasted on Python or Prothon. Also they don't seem to be making much progress on it. > Writing your own VM is much easier than to use something that > carries a lot of baggage with it. I did a few tiny machines > for my own testing, until PyPy started, and Armin convinced me > that a VM is a nice thing to have, but useless if you have > a real compiler. So that's probably my way, although I always > like to have my own toy interpreter I think a JIT is best of all worlds. begin 666 smile.gif M1TE&.#EA#P`/`)$!`````+^_O___`````"'Y! $```$`+ `````/``\```(N MC V9QY$"X6(@6GGJO0!)+3RA$XDA:&Y6JGXMIX$K%G,8^2EE]G:4?&ID%+Y# #`0`[ ` end