Permalink
Commits on Jul 24, 2010
  1. Fixes for the newer parrot. Doesn't correctly parse code due to nqp-r…

    …x incompatibilities - but it compiles!!
    Jarrod committed Jul 24, 2010
Commits on Dec 5, 2009
Commits on Dec 2, 2009
  1. add a Plumage description

    fperrad committed Dec 2, 2009
Commits on Jul 9, 2009
Commits on Mar 8, 2009
  1. Updated the build process (fperrad++). Also added pred/succ builtins …

    …for better compatibility with 'joy' code.
    Jarrod committed Mar 8, 2009
Commits on Feb 19, 2009
Commits on Feb 7, 2009
  1. Added lor and land. The 'l' stands for both 'list' and 'lazy'. They t…

    …ake list arguments and run them, instead of pulling two value types off the stack. The big advantage is that these are short circuiting expressions, so using these should save a few cpu cycles.
    Jarrod committed Feb 7, 2009
Commits on Feb 5, 2009
  1. Added range and genrange functions, kinda handy for things. Also upda…

    …ted fun.ops, parrot renamed one of its C functions.
    Jarrod committed Feb 5, 2009
Commits on Feb 3, 2009
  1. A couple of quick fixups.

    Jarrod committed Feb 3, 2009
  2. Added a couple of examples, changed the todo.. That's about it for no…

    …w, I think fun is pretty much finished and a new implementation is needed...
    
    There are a lot of problems with this language design, mostly in terms of speed. The language needs some smart optimizations that I don't have the ability to write before it can even be considered a toy language.
    Things like Memoization are crucial to this language. If a 'while' statement has to run a block of code from a prior continuation every loop, it is just a huge waste of CPU cycles, enough to be a show-stopper. Memoization would fix this easily but it's difficult to implement; Functions may be pure or impure, I would need to use static analysis to group atoms into functions, and it's difficult to see how many arguments a function even takes since arguments aren't defined..
    I think that it's time to work on other things.
    Jarrod committed Feb 3, 2009
Commits on Jan 30, 2009
  1. Modified the todo. Already having second thoughts about it. This is a…

    …n annoying design decision.
    Jarrod committed Jan 30, 2009
  2. Modded the recusrion depth limit for fun. Pity that parrot segfaults …

    …for some reason when you delve too deep anyway. agh.
    Jarrod committed Jan 30, 2009
  3. Looks like I got it right the first try. I'm amazed. Removed the debu…

    …g prints from binrec.
    Jarrod committed Jan 30, 2009
  4. Working on making binrec nonrecursive.

    Jarrod committed Jan 30, 2009
Commits on Jan 28, 2009
  1. Fixed a strange bug in delayedsub, was probably a memory leak or some…

    …thing. Also found condlinrec was trying to get the stack from the global namespace. Fixed.
    Jarrod committed Jan 28, 2009
  2. Merge branch 'newdispatch'

    Conflicts:
    
    	fun.pir
    Jarrod committed Jan 28, 2009
  3. Moved a few private vars to the correct namespace.. Time to merge and…

    … push! I guess that means this is the 1.0 release.
    Jarrod committed Jan 28, 2009
  4. Well it looks like its all done in one giant commit. Still need to ch…

    …eck all the functions using private variables are now using the namespace, but so far so good..
    Jarrod committed Jan 28, 2009
Commits on Jan 27, 2009
  1. Ok I guess thats all the dispatcher needs to be. Now to tinker with a…

    …ctions.pm
    Jarrod committed Jan 27, 2009
  2. Took out a couple of lines I added that were preventing a clean build…

    …. master will stay here until the namespace refactor and new user function stuff is done.
    Jarrod committed Jan 27, 2009
Commits on Jan 24, 2009
  1. Added getundeferror/setundeferror. They don't work yet, I need to rew…

    …rite how user functions are called. Workin' on it.
    Jarrod committed Jan 24, 2009
  2. Added construct.

    Jarrod committed Jan 24, 2009
  3. Fixed a bug in primrec.

    Jarrod committed Jan 24, 2009
Commits on Jan 23, 2009
  1. Changed stack/unstack a little. Stack once again only returns the cur…

    …rent continuations' stack. I have no idea why I thought the other way was a better way to do it. I also added 'infra'. Might change the way it works, regarding its use of EOSMarker.
    Jarrod committed Jan 23, 2009
Commits on Jan 22, 2009
Commits on Jan 21, 2009
  1. Renamed (and implemented) autoput/setautoput. Use setdot/getdot. The …

    …idea is that the dot represents the function that you pass it.
    Jarrod committed Jan 21, 2009
  2. Added a hook for the dot so that it can do things other than 'put' wh…

    …en execution has finished. Had to hack in a workaround using a new PMC since the dot is meant to be triggered once execution has finished.. and adding a hook means more execution.
    Jarrod committed Jan 21, 2009