RE: lingo-l Lingo Gurus! Need some Input!
Can you set the colors on this sprite so that you can't see it? (leave it visible, just camouflage it. That's pricisely the kludge I'm running now. And it works, so I'm not going to fight it any longer. The QA Mode status text sprite is on-stage, but is shrunk to a height of only a few pixels (so that no text shows), and the scroll bar (which is evidently necessary) is just barely off-stage and not visible. At the point I need this, it doesn't show anything. I presume you are not doing this in your parsing, etc.? repeat with var = 1 to blah.count Well, Buzz...the parsing is long since done at the point of my problems. The parsing is a split second thing earlier in the start-up of the engine. There's just some DOM access going on in the object creation loop -- no parsing, per se, going on -- and the repeat block iterates to a maximum held within a property variable of the ItemManager object: property miNumItems miNumItems = goXML.getDocumentElement().getAttribute(NumItems) repeat with i = 1 to miNumItems ¯¯¯ Christopher Watson Sr. Software Engineer Interactive Web Media Lightspan, Inc. Tel 858.824.8457 Fax 858.824.8001 ___ [To remove yourself from this list, or to change to digest mode, go to http://www.penworks.com/lingo-l.cgi To post messages to the list, email [EMAIL PROTECTED] (Problems, email [EMAIL PROTECTED]). Lingo-L is for learning and helping with programming Lingo. Thanks!]
RE: lingo-l Lingo Gurus! Need some Input!
I'm bamboozled. Ditto. Have you tried throwing some put statements in there or commenting out specific chunks of code to try to determine what portion of the code is causing the slow-down? I've never seen an issue with instanciation of child objects, no matter how many levels deep. The latest version of assessment software that I wrote had a similar initialization routine that only took a second or two to initialize on low end machines. How bulky are the XML files that you are parsing? Could they be slowing things down? Where are you doing this intitalization (frame script, parent script, movie script)? Brian Romanko Lead Developer - Neo/SCI Corporation Member - Greater Rochester Macromedia User Group [To remove yourself from this list, or to change to digest mode, go to http://www.penworks.com/lingo-l.cgi To post messages to the list, email [EMAIL PROTECTED] (Problems, email [EMAIL PROTECTED]). Lingo-L is for learning and helping with programming Lingo. Thanks!]
RE: lingo-l Lingo Gurus! Need some Input!
Don't have a clue, but have you thought about leaving the text sprite there and not worrying about it? What happens if you have the text sprite, but don't write to it, or only write to it once in the whole routine? We ran a test that left the text sprite on-stage, and wrote into it in Release Mode just as we do in QA Mode. But as far as we can tell, the speed-up only occurs when that text sprite is on stage and visible. And we simply can't have that in Release Mode. It's not part of the design. Hoping that updateStage (or otherwise making visual changes to that which is on-stage during this initialization process) would kick the speed up, we tied the object creation loop into the progress bar that is already a part of the engine start-up. No dice. Even making that update once within each iteration of the loop does nothing to kick it up. So it's not tied to writing to the stage or updating it. It's looking more and more like it's all about the ancestry tree I'm creating. It must just be too complex for these slower, older machines. But that just seems weird. ¯¯¯ Christopher Watson Sr. Software Engineer Interactive Web Media Lightspan, Inc. Tel 858.824.8457 Fax 858.824.8001 ___ [To remove yourself from this list, or to change to digest mode, go to http://www.penworks.com/lingo-l.cgi To post messages to the list, email [EMAIL PROTECTED] (Problems, email [EMAIL PROTECTED]). Lingo-L is for learning and helping with programming Lingo. Thanks!]
RE: lingo-l Lingo Gurus! Need some Input!
Have you tried throwing some put statements in there or commenting out specific chunks of code to try to determine what portion of the code is causing the slow-down? Yes. See my reply to Jakob. How bulky are the XML files that you are parsing? Could they be slowing things down? The XML is really pretty small. Definitely not bulky. Nothing unique or special about it, either. And DOM-Lingo handles everything just fine. It's not the bottleneck. Where are you doing this intitalization (frame script, parent script, movie script)? The main Engine object is created in the startMovie handler. But it's a frame script that calls the Engine object to create the ItemManager object, whose new handler conatins the repeat loop that creates the list of Item objects, which ancestor themselves to Layout objects, which ancestor themselves to Template objects, which ancestor themselves to SampleLayout objects, which ancestor themselves to IntroLayout objects. Everything is a parent script. The one and only movie script contains the startMovie handler, and a few global utility functions. ¯¯¯ Christopher Watson Sr. Software Engineer Interactive Web Media Lightspan, Inc. Tel 858.824.8457 Fax 858.824.8001 ___ [To remove yourself from this list, or to change to digest mode, go to http://www.penworks.com/lingo-l.cgi To post messages to the list, email [EMAIL PROTECTED] (Problems, email [EMAIL PROTECTED]). Lingo-L is for learning and helping with programming Lingo. Thanks!]
RE: lingo-l Lingo Gurus! Need some Input!
Does the text-fix also work if the member is not on the stage, or not even in a sprite? No. The sprite needs to be on-stage and visible for the speed-up to occur. -If it only 'fixes' while on the stage, then it might have to do with the sort of 'mini-updatestage' i anticipate is happening when modifying staged text-members. Sounds right to me, but I can't put any other text sprites on the stage, because it's not in the design spec to do that. I wish I could, but I can't. -Does a regular 'updateStage' also introduce a fix? No. It doesn't. That was the first thing we tried. -Have you tried to eliminate conceptual parts of your test, ie: is it the number of objects alone that grinds, or is generically using ancestry, or is it specific to some of your complex code - say DOM-XML? Yes. I have strategically commented out and/or segregated chunks of code to other repeat loops in order to pinpoint where the slow-down is actually occuring, and it is indeed a result of my object creation and complex ancestry tree. I took out the calls to DOM-Lingo, and that definitely wasn't the problem, either. It's really looking more and more like a by-product of my complex architecture. ¯¯¯ Christopher Watson Sr. Software Engineer Interactive Web Media Lightspan, Inc. Tel 858.824.8457 Fax 858.824.8001 ___ [To remove yourself from this list, or to change to digest mode, go to http://www.penworks.com/lingo-l.cgi To post messages to the list, email [EMAIL PROTECTED] (Problems, email [EMAIL PROTECTED]). Lingo-L is for learning and helping with programming Lingo. Thanks!]
RE: lingo-l Lingo Gurus! Need some Input!
At 9:42 AM -0700 4/23/02, you wrote: Does the text-fix also work if the member is not on the stage, or not even in a sprite? No. The sprite needs to be on-stage and visible for the speed-up to occur. Can you set the colors on this sprite so that you can't see it? (leave it visible, just camouflage it. -If it only 'fixes' while on the stage, then it might have to do with the sort of 'mini-updatestage' i anticipate is happening when modifying staged text-members. Sounds right to me, but I can't put any other text sprites on the stage, because it's not in the design spec to do that. I wish I could, but I can't. -Does a regular 'updateStage' also introduce a fix? No. It doesn't. That was the first thing we tried. -Have you tried to eliminate conceptual parts of your test, ie: is it the number of objects alone that grinds, or is generically using ancestry, or is it specific to some of your complex code - say DOM-XML? I presume you are not doing this in your parsing, etc.? repeat with var = 1 to blah.count -Buzz Yes. I have strategically commented out and/or segregated chunks of code to other repeat loops in order to pinpoint where the slow-down is actually occuring, and it is indeed a result of my object creation and complex ancestry tree. I took out the calls to DOM-Lingo, and that definitely wasn't the problem, either. It's really looking more and more like a by-product of my complex architecture. ¯¯¯ Christopher Watson Sr. Software Engineer Interactive Web Media Lightspan, Inc. Tel 858.824.8457 Fax 858.824.8001 ___ [To remove yourself from this list, or to change to digest mode, go to http://www.penworks.com/lingo-l.cgi To post messages to the list, email [EMAIL PROTECTED] (Problems, email [EMAIL PROTECTED]). Lingo-L is for learning and helping with programming Lingo. Thanks!] [To remove yourself from this list, or to change to digest mode, go to http://www.penworks.com/lingo-l.cgi To post messages to the list, email [EMAIL PROTECTED] (Problems, email [EMAIL PROTECTED]). Lingo-L is for learning and helping with programming Lingo. Thanks!]
Re: lingo-l Lingo Gurus! Need some Input!
I'm bamboozled. If it's not memory, why is this object creation process bogging down? And morfe importantly, why does it NOT bog down if I'm writing strings to a text sprite? Don't have a clue, but have you thought about leaving the text sprite there and not worrying about it? What happens if you have the text sprite, but don't write to it, or only write to it once in the whole routine? [To remove yourself from this list, or to change to digest mode, go to http://www.penworks.com/lingo-l.cgi To post messages to the list, email [EMAIL PROTECTED] (Problems, email [EMAIL PROTECTED]). Lingo-L is for learning and helping with programming Lingo. Thanks!]
Re: lingo-l Lingo Gurus! Need some Input!
At 16:05 -0700 22/04/02, Watson, Christopher wrote: I'm bamboozled. Questions: -Does the text-fix also work if the member is not on the stage, or not even in a sprite? -If it only 'fixes' while on the stage, then it might have to do with the sort of 'mini-updatestage' i anticipate is happening when modifying staged text-members. -Does a regular 'updateStage' also introduce a fix? -Have you tried to eliminate conceptual parts of your test, ie: is it the number of objects alone that grinds, or is generically using ancestry, or is it specific to some of your complex code - say DOM-XML? I have done complex 'object-webs' on your targetplatform, and have not seen slowdowns near your range. Jakob [To remove yourself from this list, or to change to digest mode, go to http://www.penworks.com/lingo-l.cgi To post messages to the list, email [EMAIL PROTECTED] (Problems, email [EMAIL PROTECTED]). Lingo-L is for learning and helping with programming Lingo. Thanks!]