Commit Graph

167 Commits

Author SHA1 Message Date
Jeremy Ashkenas
ae4f6309e8 Self-compiler: can now compile functions. 2010-02-09 21:31:07 -05:00
Jeremy Ashkenas
001c915c21 Self-compiler: object literals. 2010-02-09 20:53:25 -05:00
Jeremy Ashkenas
91a7102f11 Self-compiler: array slice literals. 2010-02-09 19:30:28 -05:00
Jeremy Ashkenas
c6a6788694 removing the constructor safety check -- it wasn't safe enough (Issue 36) 2010-02-09 07:59:48 -05:00
Jeremy Ashkenas
0b5b6113ee CoffeeScript-in-CoffeeScript just had it's first self-compiled snippet. 2010-02-08 22:22:59 -05:00
Jeremy Ashkenas
d5c98165ea removed an old now-impossible branch of execution from ReturnNode 2010-02-08 21:16:14 -05:00
Jeremy Ashkenas
b8d22bc572 finished up the CoffeeScript version of the Scope object 2010-02-08 21:10:48 -05:00
Jeremy Ashkenas
69808ba523 trying out a new organizational scheme, with a quickie 'inherits' helper function 2010-02-08 20:20:11 -05:00
Jeremy Ashkenas
aabfba9599 nodes.coffee is continuing to roll along -- maybe a tenth implemented 2010-02-08 19:49:39 -05:00
Jeremy Ashkenas
decaea0f5f adding 'by' to array comprehensions 2010-02-06 10:12:57 -05:00
Jeremy Ashkenas
b795ae7fe1 removing arguments as a keyword -- we can detect its use at code-generation time. 2010-02-05 22:01:11 -05:00
Jeremy Ashkenas
dc7d0f1568 fixing assigning to @properties within an expression 2010-02-03 18:16:31 -05:00
Jeremy Ashkenas
cb7a1033fa adding @property for this.property 2010-02-02 20:36:46 -05:00
Jeremy Ashkenas
c5c841f2fc fixing bug with mixed dot and soak accessors 2010-02-02 10:39:44 -05:00
Jeremy Ashkenas
af53a04932 added test for lexical scope sharing through generated closure wrappers, something uncommonly used, but that was a regression 2010-01-24 13:39:27 -05:00
Jeremy Ashkenas
817e8deb27 adding soaked method calls, with caching 2010-01-24 12:52:15 -05:00
Jeremy Ashkenas
d728c3d669 added existence chains with '?.' -- soaks up attempts to access undefined properties, returning 'undefined' 2010-01-23 23:30:55 -05:00
Jeremy Ashkenas
9160500e84 removing 'this' rewriting in favor of correctly calling generated closures 2010-01-23 21:11:27 -05:00
Jeremy Ashkenas
c3ce2ea9b1 added automatic safety closure wrapper for functions declared within for loops. 2010-01-23 17:53:07 -05:00
Jeremy Ashkenas
2d90a751f7 edits for clarity 2010-01-17 14:55:06 -05:00
Jeremy Ashkenas
8647b54a61 rename compile_double_reference to compile_reference 2010-01-17 14:26:00 -05:00
Jeremy Ashkenas
8e1f3c0eca generating multiple calls to the same function should use compile_double_reference to ensure a single evaluation of the call itself. 2010-01-17 14:23:41 -05:00
Jeremy Ashkenas
c4d0903e6a fixing assignment-in-condition 2010-01-17 10:40:59 -05:00
Jeremy Ashkenas
e72ef1a61a reverting change 2010-01-17 10:28:04 -05:00
Jeremy Ashkenas
d7d9cb8d28 only let returns stop an expression from being closure-ified -- breaks and continues may be valid 2010-01-17 10:21:24 -05:00
Jeremy Ashkenas
f6c8e81ea6 the existential operator can now be used infix as well 2010-01-16 23:03:54 -05:00
Jeremy Ashkenas
52539ae7d2 abbreviating the existential operator 2010-01-16 22:26:34 -05:00
Jeremy Ashkenas
95b362499f added the conditional existence operator 2010-01-16 22:17:55 -05:00
Jeremy Ashkenas
0bc4da2b51 ensure that functions are only called once, when chaining comparators 2010-01-16 22:04:08 -05:00
Jeremy Ashkenas
9679fc0b52 removing redundant unary check 2010-01-16 16:49:03 -05:00
Jeremy Ashkenas
9cb0564972 added Python's chainable comparisons, like: 10 > 5 > 1 2010-01-16 16:37:49 -05:00
Jeremy Ashkenas
c6c0c7d059 simplification of function and prototype naming -- last_assign, immediate_assign, and proto_assign are gone, in favor of 'name' and 'proto' properties on CodeNodes 2010-01-16 15:44:07 -05:00
Jeremy Ashkenas
62e946b8ce purely empty functions at the top level should be wrapped in parens, so as not to cause a JS syntax error 2010-01-16 15:02:04 -05:00
Jeremy Ashkenas
9eff443032 arguments no longer is just a find-and-replace -- it'll fix the arguments variable at the top of scope if you use it in a function body 2010-01-16 12:52:26 -05:00
Jeremy Ashkenas
8957feedb4 expression closure wrappers are now safer -- they won't be generated if there's a statement_only inside 2010-01-16 12:10:31 -05:00
Jeremy Ashkenas
1cd7fa8ebe added children macro to Node, using it so that all nodes now have a 'children' method -- used for safe references to 'this' within closure wrappers 2010-01-16 11:24:10 -05:00
Jeremy Ashkenas
701cdb4c13 never try to push a statement_only 2010-01-15 19:47:16 -05:00
Jeremy Ashkenas
001cc29deb slightly shorter generated code for ==> 2010-01-14 08:55:09 -05:00
Jeremy Ashkenas
e77e520607 CoffeeScript 0.2.5 is on the books 2010-01-13 23:24:45 -05:00
Jeremy Ashkenas
ed8a54995d with splats allowed in destructuring assignment 2010-01-13 22:25:58 -05:00
Jeremy Ashkenas
2d206e7b60 pulling out pushes into a pushnode 2010-01-13 21:33:46 -05:00
Jeremy Ashkenas
bb9fdd3015 while loops can now be used as expressions -- they return an array containing the computed result of each iteration. 2010-01-13 21:27:22 -05:00
Jeremy Ashkenas
1e7d638435 adding bound functions, with test 2010-01-13 20:59:57 -05:00
Jeremy Ashkenas
0ceca0778c adding when clauses with multiple values 2010-01-13 19:56:35 -05:00
Jeremy Ashkenas
abd9ab5c71 unified ParamSplatNode and ArgSplatNode into SplatNode 2010-01-12 23:49:47 -05:00
Jeremy Ashkenas
ea349a1a59 more safety type-checks in nodes.rb 2010-01-12 23:26:35 -05:00
Jeremy Ashkenas
8dfbd1a2a8 using Object.prototype.hasOwnProperty.call instead of obj.hasOwnProperty, with an alias, for Rhino and java objects 2010-01-12 17:35:37 -05:00
Jeremy Ashkenas
186797a745 got compile_pattern_match about as small as its going to get 2010-01-11 22:16:23 -05:00
Jeremy Ashkenas
d54fa2f2a1 a whole chunk of compile_pattern_match was redundant, axed it -- along with the array/vs/object split 2010-01-11 22:12:18 -05:00
Jeremy Ashkenas
6c980d8adc first draft of destructuring assignment -- working out the wrinkles -- not sure if we want to do the full spec 2010-01-11 21:44:47 -05:00