Yeah could be the beers, when I said clone… I meant clone. Thanks for the interest though. cwright replied to me privately at one point so the conversation may have got confusing. I got into the muddle because i was trying to avoid duplicating the whole structure 'by hand' and then I tried just editing the data I I wanted to. I assume editing the Input Structure receving the structure from the first JS patch (and thus avoiding a duplication loop) was a no_no, but I'l try it just to see.
Best Alastair 2010/4/12 Boštjan Čadež <[email protected]> > for some reason (prolly the couple of beers i had) i think you got some > terminology mixed up. what do you mean when you say clone? for me a clone > would mean a copy, which in this case would be a good thing. you can edit a > copy. the problem you encountered happens with references. to be sure of > what im trying to say ill go into to much detail. > > lets say you have a structure like (in this case an object): > > poo={}; > poo.color="brown"; > poo.weight=2.3; > poo.shape=["sphere","sphere","cylinder"]; > > lets say you have this in one JS patch and output it to another one via > noodle. in the one that receives you have: > > pooCopy=poo; > pooCopy.weight=2.1; FAIL > > this doesnt work because "pooCopy=poo;" just copies the memory address (a > number) for the object (aka reference) not the whole object (many numbers). > so "pooCopy.weight=2.1" still tries to change the object you cant change, > because inputs are read only. > > second try: > > pooCopy={}; > pooCopy.color=poo.color; > pooCopy.weight=poo.weight; > pooCopy.shape=poo.shape; > > pooCopy.color="grey"; > pooCopy.weight=2.1; > pooCopy.shape=["sphere","sphere","sphere"]; this works > > but if you tried this instead: > > pooCopy.shape[2]="sphere"; FAIL > > this is because poo.shape is an object also. and "pooCopy.shape=poo.shape;" > puts a reference for poo.shape at pooCopy.shape which you cant change > because its read only. > > > pooCopy={}; > pooCopy.color=poo.color; > pooCopy.weight=poo.weight; > pooCopy.weight=[]; > for(var i=0;i<poo.shape.length;i++) > { > pooCopy[i]=poo[i]; > } > > this would do the job. basically, you have to copy each property of an > object of an object of an object / element of an array of an array of an > array "by hand". > > i once found a recursive function that would add the copy ability to the > object prototype, but that does ugly stuff to the output ports. IMHO its > better to make your own copy function for your custom structure, or in the > case above, make just one javascript patch and avoid the copying. > > p.s. beer makes me talk to much;) > > > On Apr 11, 2010, at 5:21 PM, Alastair Leith wrote: > > Yes okay, so I can do this kind of iteration in the main function (often I > do it that way but sometimes in a function call) and still get cloning not > referencing right? > > Yeah an old screen shot saved me on my last project ;) One day I would like > to get to the bottom of that. I was able to write the JS in a fresh patch > carefully testing for parsing as I went and I got it functioning but the > exact same text was jamming in another JS patch no matter how much I retyped > it. I feel the JS patch is a little illogical at times (as in it gets > broken), if that's a good way to put it. > > Oops, yes recursive is what I meant but it seems to be a regressive move in > JS too. > > On 12 April 2010 01:12, Alastair Leith <[email protected]> wrote: > >> >> >> ---------- Forwarded message ---------- >> From: Alastair Leith <[email protected]> >> Date: 12 April 2010 01:06 >> Subject: Re: Javascript issue modifying a structure (again… :/) >> To: Christopher Wright <[email protected]> >> >> >> Yes that helps heaps. Thanks cwright. I kind of twigged to the referencing >> thing a few weeks back but didn't really extrapolate the general rule (I was >> getting some strange results!) and still not 100% sure I have yet. So… >> >> Does the function you wrote also assign individual item references back to >> the original structure? If they are *values* and not *references* put >> into *newObj*, is that distinction because of the call to another >> function or just the iterating through at the lowest child level of the >> structure? If items had sub-items would I need to iterate though sub-items >> (in more function calls ? regressive function calls are out in JS aren't >> they?)? >> >> So many questions! I'm really grateful for your help already though, >> thanks again. >> >> Alastair >> >> I guess the downside of a loosely typed language like JS when your trying >> to teach yourself scripting (that's me btw ;) ) is that things can get a >> little slippery and not generate meaningful error mesages... >> >> On 12 April 2010 00:49, Christopher Wright <[email protected]> wrote: >> >>> If anybody could identify my misconception here I be very grateful: >>>> >>> >>> >>> You're assuming that you can modify inputs -- you can't. >>> >>> You're assuming that JS copies by value -- it doesn't (it copies by >>> reference). >>> >>> So, when you assign glyph to an input: glyph = glyphs[cursor]; >>> glyph is a reference to the input -- trying to modify that will fail. >>> >>> You need a copy function, as follows: >>> >>> function copy(obj) >>> { >>> var newObj = new Object(); >>> for(var e in obj) >>> newObj[e] = obj[e]; >>> return newObj; >>> >>> } >>> >>> cursor = 5; >>> function (__structure Glyphs_Out, __boolean Same, __string Font, >>> __structure Glyph) >>> main (__structure Glyphs_In) >>> { >>> var result = new Object(); >>> var glyph = new Object(); >>> var glyphs = new Array(); >>> var font = new String(); >>> font = "LucidaGrande"; >>> if (!_testMode) >>> { >>> glyphs = Glyphs_In; >>> glyph = copy(glyphs[cursor]); >>> >>> >>> // This line works: >>> font = glyph.Font_Name; >>> glyph.Font_Name = font; >>> glyph.Y +=0.1; >>> // Can't modify inputs, so you'll have to copy some other >>> inputs too. >>> //glyphs[cursor]["Y"] +=0.1; >>> >>> >>> result.Font = font; >>> result.Glyph = glyph; >>> result.Glyphs_Out = glyphs; >>> return result >>> } >>> } >>> >>> Hope that helps! :) >>> >>> -- >>> [ christopher wright ] >>> [email protected] >>> http://kineme.net/ >>> >>> -- >>> [ christopher wright ] >>> [email protected] >>> http://kineme.net/ >>> >>> >> >> > _______________________________________________ > Do not post admin requests to the list. They will be ignored. > Quartzcomposer-dev mailing list ([email protected]) > Help/Unsubscribe/Update your Subscription: > > http://lists.apple.com/mailman/options/quartzcomposer-dev/fskolog%40gmail.com > > This email sent to [email protected] > > > lp, > fšk > > > >
_______________________________________________ Do not post admin requests to the list. They will be ignored. Quartzcomposer-dev mailing list ([email protected]) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/quartzcomposer-dev/archive%40mail-archive.com This email sent to [email protected]
