[12:16am]<hassox_ is now known as hassox. [01:04am]<hassox_ is now known as hassox. [01:22am]<hassox_ is now known as hassox. [11:07am]<spoob> Damn, switching the engine back to Rhino makes me think of the Java knock-knock joke... "who's there?" "...... . ..... .... ... Java". [11:11am]<tlrobinson1> hehe yeah the jvm startup time is particularly bad [11:37am]<Dantman> That's why we don't use CGI [2:44pm]<Dantman> Yay, http://code.google.com/p/chromium/issues/detail?id=20144#c10 [2:46pm]<Dantman> Can't remember if Rhino still doesn't. But looks like v8 is going to follow the same object property order pattern that the rest of the engines follow. [2:48pm]<ondras> Dantman> actually, your comment in that issue seems irrelevant [2:48pm]<ondras> because noone was questioning Array iteration [2:48pm]<ondras> I believe that they were experiencing non-defined iteration order when some of the (iterated) properties were arrays, nothing more [2:49pm]<ashb> also ES5 mandates an opder now [2:50pm]<ondras> nice [2:50pm]<ondras> so the looping order == definition order, okay? [2:52pm]<ashb> think soi [2:52pm]<ashb> let me check that [2:55pm]<ashb> hmmm where would it bee.. [2:56pm]<ashb> ah perhaps not [2:56pm]<ashb> Object.keys says: [2:56pm]<ashb> If an implementation defines a specific order of enumeration for the for-in statement, that same enumeration order must be used in step 5 of this algorithm. [2:56pm]<ashb> so its not required, but most platforms do [2:58pm]<ondras> hm, so it is basically the same as es3 [2:58pm]<ashb> 'The mechanics and order of enumerating the properties (step 6.a in the first algorithm, step 7.a in the second) is not specified.' [2:58pm]<ashb> thats about for .. in [2:58pm]<ashb> so yeah, not specified [2:58pm]<ashb> probably at v8's urging [2:58pm]<ondras> more anti-v8 activists! [2:58pm]<ondras> [2:59pm]<? ondras wonders if he ever written some code which would fail with non-defined order [2:59pm]<ashb> i know i have [2:59pm]<ondras> baaad ashb! [3:04pm]<Dantman> You get stuff written in the browser that looks strange since it's data stored in an object that ends up differently ordered. [3:07pm]<ondras> one can say the same about a lottery wheel [3:10pm]<spoob> defined order sounds like a desirable condition, but it would be very slow for structures that can be iterated but not in order. [3:10pm]<spoob> for example; iterating through a hash or binary tree [3:11pm]<spoob> (yes, I know how to traverse binary tree in-order. Foot. Blam! [3:58pm]<ryah is now known as ryah_away. [4:01pm]<Dantman> spoob, We're only talking about plain no-frills objects... There shouldn't be anything stopping you from creating specialized optimized Set, Hashtable, etc... structures. [4:01pm]<Dantman> Actually, I'd kind of like those [4:02pm]<Dantman> Sometimes when doing algorithmic stuff I run into a case where I'd like to have a Set, but have to deal with an array since I can't hack it with an object since my data isn't string based. [4:02pm]<Dantman> Or may have values that conflict with js special cased properties [4:03pm]<spoob> Dantman; I mean that having the requirement for guaranteed ordering on iteration could be computationally expensive for some kinds of randomised data structures [4:04pm]<spoob> hence leaving that bit out, resulting in v8 engines being particularly bizarre. [4:05pm]<? Dantman still would like some of the various types of collections [4:21pm]<gmosx_> hello [5:55pm]<Wes-> dantman> you could explore this technique: [5:55pm]<Wes-> const ballSet = { basket> void 0, base> void 0, bowling> void 0 } [6:03pm]<inimino> Wes-> I think that's what "can't hack it with an object since my data isn't string based" was about [6:04pm]<Wes-> frankly that didn't make sense to me [6:04pm]<Wes-> arrays *are* objects [6:05pm]<ashb> i think it was about cant store the object in the keys object keys are only strings [6:06pm]<inimino> right [6:06pm]<Wes-> but then he couldn't store the key in an array either [6:07pm]<ashb> [{sure> "could"}] [6:07pm]<Wes-> right, the key in that case is 0 [6:07pm]<inimino> right, that's the idea [6:08pm]<Wes-> So he wants sets of objects, ah [6:08pm]<Wes-> I was still thinking of pascal sets [6:45pm]<RayMorgan_ is now known as RayMorgan. [6:46pm]<ryah_away is now known as ryah. [10:00pm]<MisterN_ is now known as MisterN.