[Webkit-unassigned] [Bug 34019] Custom-written JavaScript parser

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=34019





--- Comment #23 from Zoltan Herczeg <zherczeg at webkit.org>  2010-01-27 06:27:34 PST ---
Thank you for the feedback!

> (it is ~ 1.6 time better)

great! I hope that was in release mode :)

>   "new f(x++)"

Thanks. I also found some bugs myself, and the fast/js layout tests also shown
me some others, like: (1++).x; a(x++).b; for (new a in c) { }, and "a + b = c"
should fail on the syntax checker as well. All of them are fixed in the newest
patch.

The constructor of the parser extended to:
Grammar(JSGlobalData* globalData, int maxFunctionDepth = 1, bool strict =
false)

I started to work on maxFunctionDepth first, and it is supported now. Tested
both on small examples and all JSCore tests are passed with maxFunctionDepth =
3 . The 'strict' is just added for future use. I really want to implement the
strict mode checks in a different patch :)

There are only two fast/js tests which fail:
fast/js/reparsing-semicolon-insertion.html - the comment says:
// According to the ECMA spec, these should all be syntax errors. However, the
// pre-existing behaviour of JavaScriptCore has always been to accept them.

fast/js/function-declaration-statement.html - crashes. No idea what is
happening here (yet), since the JS code snippets are parsed correctly by the
command line tool.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.