You can clone with
HTTPS or Subversion.
Simplify attribute handling.
Trade flexibility for speed... for now.
Fix line numbering using Parslet::Slice.
Fix "error" rule handling -- semicolon insertion works!
Default case doesn't have a left.
More rounding out of the parser; runs eg/richards.js again.
I think this returns the parser to its previous working state... though
much slower than the old one, even on parslet's "tuning" branch.
Fix some places the parser would choose the wrong option.
Also, "//" comments go to the end of the line, not the file.
Check inside blocks for variable declarations. Oops!
Loops and switch/case: eg/cond.js works again.
obj.js working again
Better reserved-word handling.
The previous approach wouldn't have treated 'instanceof' as reserved,
because it starts with 'in', another reserved word.
Fix dependency to point to Parslet instead of RKelly.
Initial work to use new Parslet parser.
Lots of stuff unimplemented that was previously working... but
eg/math.js and eg/nest.js run.
New parser, using Parslet.
Not translating to an AST yet.
Start converting to a Citrus parser.
Rudimentary RegExp support.
Hack towards something that can run the SpiderMonkey tests.
Push get/set/etc conditions into Node classes.
Even though they aren't ours, so it's not particularly neat... it's
better than big case/when blocks.
Rearrange line numbering.
CaseClauseNode goes away, because I'm being evil instead.
Implement switch, continue, break, ?:.
A few random kernel additions.
Don't rely on #call; use #js_call.
Comparison operators treat null/true/false as 0/1/0.
Omit js_key calls for constant fixnums.
Oops... && and || aren't quite *that* safe.
Rearrange runner to almost support running multiple files.
And add a few more globals.
Implement proper get+set handling.
Improve argument handling, avoid hashes, look at call_custom.