View previous topic :: View next topic |
Author |
Message |
JarrettBillingsley
Joined: 20 Jun 2006 Posts: 457 Location: Pennsylvania!
|
Posted: Sun Jun 22, 2008 11:20 am Post subject: New MiniD 2 implementation |
|
|
I'm working on it! I know it hasn't seemed like I've been doing anything but I've been doing a LOT over the past 2 weeks, none of which I've committed yet. I've reimplemented most of the MiniD2 interpreter, API, and other implementation details. It runs code! Still to go are the stdlibs and compiler. Ugh, the compiler.
Probably the biggest changes are the new Lua-like API, which is because of the other big change: it now has its own GC and does its own memory allocation. Also, preliminary tests have shown the new interpreter to be faster than the old one in virtually every way, though I'll have to run some more thorough benchmarks for statistical significance |
|
Back to top |
|
|
Kosmonaut
Joined: 22 Jun 2008 Posts: 10
|
Posted: Sun Jun 22, 2008 6:16 pm Post subject: |
|
|
Nice!
I have been a lurker in this forum. But nonetheless I have followed the MiniD development with interest. Lots of thing to learn from the source code.
Do you have an idea when we could see this new implementation committed in the source code repository? There's a timeline? or something like that?
Thanks for the hard word, and congratulations for this great project.
Kosmonaut |
|
Back to top |
|
|
JarrettBillingsley
Joined: 20 Jun 2006 Posts: 457 Location: Pennsylvania!
|
Posted: Mon Jun 23, 2008 7:34 am Post subject: |
|
|
Thanks!
Things are progressing a bit faster than I anticipated. If all goes well, I hope to have the standard library converted over to the new API (which might imply some new API constructs being created) within the next week or two, at which point I'll upload it. Or I might just upload it sooner, since I'm starting to get antsy. But right now, it's by no means in any shape for serious use. I have it compiling stuff by piping the output of the old minidc into my test program X)
Eee hee hee, you'll like the new options for coroutine support. |
|
Back to top |
|
|
Ligustah
Joined: 21 Oct 2007 Posts: 45 Location: Berlin, Germany
|
Posted: Mon Jun 23, 2008 12:54 pm Post subject: |
|
|
Oh Great!
I can't wait to get my hands on that new implementation.
It's pretty nice that you managed to get the interpreter even faster!
Mfg Ligustah |
|
Back to top |
|
|
Kosmonaut
Joined: 22 Jun 2008 Posts: 10
|
Posted: Mon Jun 23, 2008 3:19 pm Post subject: |
|
|
JarrettBillingsley wrote: | Thanks!
Things are progressing a bit faster than I anticipated. If all goes well, I hope to have the standard library converted over to the new API (which might imply some new API constructs being created) within the next week or two, at which point I'll upload it. Or I might just upload it sooner, since I'm starting to get antsy. But right now, it's by no means in any shape for serious use. I have it compiling stuff by piping the output of the old minidc into my test program X)
Eee hee hee, you'll like the new options for coroutine support. |
Well, that's great! You just take your time, it was just curiousity what drove me to ask the timeline, but its nice to know that its near completion. I'll check this forum periodically. |
|
Back to top |
|
|
JarrettBillingsley
Joined: 20 Jun 2006 Posts: 457 Location: Pennsylvania!
|
Posted: Fri Jun 27, 2008 9:22 pm Post subject: |
|
|
Yeah, it's up now, in /branches/v2new. No, it's not complete. No, it probably doesn't work out of the box right now. I have it on life support, hooked up to hard-coded executables and temporary loading mechanisms on my end. |
|
Back to top |
|
|
Kosmonaut
Joined: 22 Jun 2008 Posts: 10
|
Posted: Sun Jun 29, 2008 8:40 am Post subject: |
|
|
Wow! Completely unespected. Well, at least we get a glimpse of what will become the new MiniD 2. Thanks! |
|
Back to top |
|
|
JarrettBillingsley
Joined: 20 Jun 2006 Posts: 457 Location: Pennsylvania!
|
Posted: Thu Jul 17, 2008 9:43 am Post subject: |
|
|
Yeah so, I've been working (job). Things just keep getting pushed back. Redesigning the compiler is a lot less fun than redesigning the interpreter X| |
|
Back to top |
|
|
Kosmonaut
Joined: 22 Jun 2008 Posts: 10
|
Posted: Sun Jul 20, 2008 3:35 pm Post subject: |
|
|
JarrettBillingsley wrote: | Yeah so, I've been working (job). Things just keep getting pushed back. Redesigning the compiler is a lot less fun than redesigning the interpreter X| |
Well, it seems we will have to wait more
Don't worry, take your time, real life sometimes interfers with everything else. |
|
Back to top |
|
|
h3r3tic
Joined: 30 Mar 2004 Posts: 261 Location: Torun, Poland
|
Posted: Wed Jul 23, 2008 5:40 am Post subject: |
|
|
*force persuasion* You want to drop your job and return to MiniD hacking |
|
Back to top |
|
|
JarrettBillingsley
Joined: 20 Jun 2006 Posts: 457 Location: Pennsylvania!
|
Posted: Wed Jul 23, 2008 6:44 am Post subject: |
|
|
I'm not doing nothing, you know. I'm up to the codegen phase of the compiler, which is certainly the trickiest part in terms of memory management. Once that's done, things should move much more smoothly. |
|
Back to top |
|
|
JarrettBillingsley
Joined: 20 Jun 2006 Posts: 457 Location: Pennsylvania!
|
Posted: Thu Jul 24, 2008 2:53 pm Post subject: |
|
|
Codegen is done. By that I mean "I've run it through some tests and it seems to be behaving itself." There are undoubtedly bugs in it somewhere. |
|
Back to top |
|
|
|