[flexcoders] Re: BlazeDS RemoteObject Performance
D'oh, you beat me to it! The only addition I can make is that this can usually be changed in the registry for IE, or in Firefox via the about:config page. I think the defaults in IE 8 and Firefox 3 are 8 connections, IE before 8 was indeed 2. --- In flexcoders@yahoogroups.com, Fotis Chatzinikos fotis.chatzini...@... wrote: Are these calls all coming from the same client? Most browsers will open two concurrent connections to the same domain not more. If this is the case its not a server limitation but a client limitation
[flexcoders] Re: Check to see if component is visible
Not quite what I did, but you are right, once I reset my loop to go all the way up to the application level I did find a level where visible was false. Thanks for you help. --- In flexcoders@yahoogroups.com, Alex Harui aha...@... wrote: If you did: mx:TabNavigator mx:ModuleLoader mx:ModuleLoader then the ModuleLoader for the tab that is not selected should have visible=false Alex Harui Flex SDK Developer Adobe Systems Inc.http://www.adobe.com/ Blog: http://blogs.adobe.com/aharui
[flexcoders] Check to see if component is visible
Sorry if this explained in detail somewhere, but if it is I haven't been able to find it. I've got modules embedded in tabs, if the module isn't visible I don't want it to react to certain eventsthe problem is, every component I've checked, including the module itself, the visible property is true. One suggestion I did find said that perhaps something in the parent chain would show as visible false...but this does not appear to be the case. Anyone have any suggestions? tia Mike
[flexcoders] Re: BlazeDS messaging over SSL
I was able to get it push working, via secure long poll, not streaming. Mike --- In flexcoders@yahoogroups.com, Pratima Rao pratima_j...@... wrote: Hi, I have an application that uses the data push technology of blazeDS to send data to a Flex Client event 5 seconds. The application works fine when I run it via HTTP with or without a proxy. When I run it via https the data push doesn't work anymore. I get the following error rootCause [IOErrorEvent type=ioError bubbles=false cancelable=false eventPhase=2 text=Error #2032: Stream Error. URL: https://localhost/admin/messagebroker/streamingamfsecure?command=openversion=1 Has anyone successfully got streaming to work over SSL? Thanks, Pratima
[flexcoders] Re: BlazeDS data push over SSL
We got it working, not with the actual streaming but via long poll. we had to set the poll length to under 5 minutes in order to placate Weblogic, but otherwise it works fine. --- In flexcoders@yahoogroups.com, pratima_jrao pratima_j...@... wrote: I have an application that uses the data push technology of blazeDS to send data to a Flex Client event 5 seconds. The application works fine when I run it via HTTP with or without a proxy. When I run it via https the data push doesn't work anymore. I get the following error rootCause [IOErrorEvent type=ioError bubbles=false cancelable=false eventPhase=2 text=Error #2032: Stream Error. URL: https://localhost/admin/messagebroker/streamingamfsecure?command=openversion=1 Has anyone successfully got streaming to work over SSL? Thanks, Pratima
[flexcoders] Re: Flex Builder Compiler Performance Benchmarking
Hm, as you said in a later post, I'd suspect the xml file, or perhaps other resources. I'm working on a fairly large project, some Java, a few graphics and css, as well as 70+ modules of various sizes and our build time with a little homegrown Maven plugin is coming in at under 8 minutes. Please let us know if you find out what the issue is. Good luck! Mike --- In flexcoders@yahoogroups.com, edlueze edlu...@... wrote: Hi Folks: Can somebody tell me what a reasonable compile-time for a Flex Application is? A clean build for my application will take about 2 hours, and an incremental compile after I've changed several files will often take an hour. If I've only changed one file then an incremental compile will usually take less than a minute.
[flexcoders] Re: Flex RegExp issues
We faced a similar problem where I'm working, everyone had to implement a filtering mechanism using user input, including some special characters. While we were comparing solutions it was noted that all but mine had troubles with the special characters.I was just using indexOf(string)-1 rather than the RegExp methods. Now, they kind of had me beat when we had to make it case-insensitive, they only had to add the i second parameter whereas I had to upper or lower case everything, so I had to type more characters there, but it was still way less than the escaping code they had to add. I guess long story short is, take a look at all of the tools in the box, and make sure you're using the right one for your use-case. --- In flexcoders@yahoogroups.com, Manu Dhanda manuraj.dha...@... wrote: The problem is: I am trying to filter an arraycollection(labels) against the text. Using a filter function and in there I am trying to use RegExp. More specifically, whenever I enter a text like 12.0, it returns me results with - character as well like 12-0 etc. So I want that whenever I enter . in my search, it should only return me results with . and NOT - and vice-versa. Thanks, Manu.
[flexcoders] Complicated form layout
I find it odd that if I layout a flex form as mx:HBox mx:Form mxFormItem label=A Label mx:TextInput/ /mx:FormItem /mx:Form mx:Form mxFormItem label=B Label mx:TextInput/ /mx:FormItem /mx:Form /mx:HBox I get the same behavior programmatically as the following, but a very different visual experience if I do it as mx:Form mx:HBox mx:FormItem label=A Label mx:TextInput/ /mx:FormItem mx:FormItem label=B Label mx:TextInput/ /mx:FormItem /mx:HBox /mx:Form Now to my wee poor brain, the 2 should be the same, and while they areroughlythe label alignment and all that is very different. Can anyone explain to me why? And by that I don't mean, If you put any container between the form declaration and the form items it'll whack out your stuff., that much I've figured out by this point, I mean, why the heck would they go to the trouble of making things whack out if a form item's immediate parent ISN'T a form, as long as its eventual parent is??? Ok, deep breath...sometimes you just have to let go. [:)] back to work
[flexcoders] Re: Complicated form layout
--- In flexcoders@yahoogroups.com, Tracy Spratt tspr...@... wrote: An HBox in a Form is a very different thing than a Form in an HBox. A container is going to layout its children according to is own rules. You can't expect it to alter those rules depending upon its parent. Tracy Spratt Lariat Services Assuming these are done with styles, is there any cascading in these styles? I would think that a child would conform to the styles of its parent. I mean they put the C in CSS for a reason. I've already noticed that any styles declared against, say a ComboBox, don't seem to apply to any OO descendants of combobox. So do we need to rename Flex stylesheets from css to just ss? I acknowledge the fact that this is indeed the way things are, but I feel that it more likely a naive implementation of cascading styles rather than a deliberate decision. Just like the problem with certain global resources in modules, like the resourcemanager, might go missing at odd moments unless you do certain things, it is the way it is, but that doesn't make it any less a bug, nor does it make it a thing that shouldn't be fixed. Ok, deep breath...thank you very much for your reply. I agree that it is most probably my misconception in this case that just because when I apply a style to xml, or html, or xhtml, a container sure does alter its rules based on those of its parent, again based on the little c in css, it is wrong of me to assume that mxml would react the same way, or that anyone would feel a need to either make it consistent with the rest of the world or document the difference anywhere. But you've been kind in making a reply to me, again, thank you.
[flexcoders] Re: Are flex events really thread safe?
No, there is no way to my knowledge that you could tell which response object would be actually correspond to which particular call, that would require a synchronous response and even IE 6 should be allowing you to execute up to 2 simultaneous calls. You probably should be using separate response methods. Actually I've never heard that the events in flex are supposed to be thread safe...only that they are asynchronous so they allow us to fake faster response and LOOK like things are multi-threaded. --- In flexcoders@yahoogroups.com, Shyam shyammohan.sugat...@... wrote: Consider this scenario. i have issued some 5 network calls form the flex client and my complete event Handler is one and the same for all these http calls. can i be 100% sure that two simultaneous controls will never execute the code in my event Handler ? Putting it another way: how does flex handles event synchronization? i have read up Ted Patrics article on the Elastic Racetrack (http://www.powersdk.com/ted/2005/07/flash-player-mental-model-elastic.php) but couldnt figure out the answer for the above :( Regards Shyam
[flexcoders] Re: Flex/BlazeDS/Hibernate - serialization question
We are still dealing with this problem on a daily basis on my current project, and there still doesn't seem to be a good answer, at least as far as BlazeDS goes. The way that web frameworks deal with the issue is lazy loading, so that the data access to load the child objects doesn't happen until you actually access those objects, but because we're serializing the objects when we send them acrossed the wire in BlazeDS, the entire object graph gets realized, leading to some HUGE amounts of data possible, depending on how complex your data is. Now supposedly they have fixed this, meaning enabled lazy loading, in GraniteDS, and many are hoping for a similar solution coming out of the Spring/BlazeDS integration project. I say supposedly because I haven't personally tested it, we're 7 months into an 8 month project, not a good time to be thinking about changing out such a major part of our project! Good luck! Mike --- In flexcoders@yahoogroups.com, max.bruchmann So now everytime we send a person object back to the blazeds/hibernate backend all the groups he belongs to are getting serialized as well. This means a lot of overhead gets send to backend when I only modify some primitve data like the name. Even more problematic would be if in addition the group objects themselves would have list of persons that belongs to them. I guess I could use the [Transient] Metatag to not send the collection fields, but then I could not modify the groups in a person object easily. So if there is any best practice for these kind of problems, please let me know. Thanks in advance Max
[flexcoders] Re: Interfaces not working the way I understand anyway
Not necessarily, but then the app has been a bit.different from day one. In it most of the modules are discreet bits of functionality where the Main App simply organizes and hosts the other modules, which may in turn host other modules that the main app knows absolutely nothing about. Except now it seems pretty clear that if the main module REALLY knows nothing about any discreet piece, that piece might go walkabout at any time. Important safety tipwe live and learn. thx for all the help everyone! Mike --- In flexcoders@yahoogroups.com, Josh McDonald dzn...@... wrote: Just the interfaces, but your modules should be using interfaces defined (and used) in the main app anyway, no? -Josh
[flexcoders] Re: Interfaces not working the way I understand anyway
--- In flexcoders@yahoogroups.com, Alex Harui aha...@... wrote: Sounds like an applicationDomain topology problem. See the modules presentation on my blog. Make sure the interface is linked into the main app. That got it! Thanks!! (maybe modules should be marked experimental or something)
[flexcoders] Re: Interfaces not working the way I understand anyway
--- In flexcoders@yahoogroups.com, mmormando m...@... wrote: --- In flexcoders@yahoogroups.com, Alex Harui aharui@ wrote: Sounds like an applicationDomain topology problem. See the modules presentation on my blog. Make sure the interface is linked into the main app. That got it! Thanks!! (maybe modules should be marked experimental or something) After seeing this, does it more or less mean that I need to statically link a copy of every stinking module I've got into my main module in order to be sure I won't run afoul of this problem every time I turn around? Not exactly sounding like an optimal solution.
[flexcoders] Interfaces not working the way I understand anyway
Ok, I've got a module implemented in an mxml file that implements an interface IMyInterface, so marked with the implements=com.my.interface.IMyInterface tag in the right place. So, after dynamically loading the module I try to either cast it to IMyInterface, or check it with is, and it says that it isn't an IMyInterface. In straight actionscript it seems to be working fine, and if I keep the instance as a raw object and try calling any of the methods that are defined in IMyInterface it works. And if I change IMyInterface and recompile without changing my mxml file I get the compiler error that MyInterfaceImplementation.mxml doesn't implement IMyInterface. Anyone know what's going on?? tia Mike
[flexcoders] Flex Builder 3.0.2 - Not a visual component
I'm building a LOT of modules, enough so that I've created my own mx:Module descendant, and everyone on the project is using that to base their own modules on. Everything worked fine until I upgraded to v3.0.2 of Flex Builder, now when I try to switch over to the visual design window, which worked in earlier version of Flex Builder 3, I'm getting an error message This component is based on 'MyCustomModuleType', which is not a visual component. Switch to source mode to edit it. Anybody else having similar problems? thx Mike
[flexcoders] Re: Parallel requests using RemoteObject with RTMP
It doesn't have anything to do with Flex, its IE. IE 6 and I believe 7, and Firefox 2 all default to only allow 2 downloads from a single subdomain. I got that from the folks at Yahoo that bring us the YSlow stuff. If you are using an Apache front end you could make your request to multiple subdomains, e.g. data1.yourdomain.com and data2.yourdomain.com would let you go to 4, conversely if you have control of all the computers hitting your service there are registry tweaks you can set, something along the lines of max_server_connections, that's from memory so don't take it as doctrine. I've also read that IE 8 beta defaults to something like 30. I'm using Firefox 3 and set it to 30, as well as allowed it to use http pipelining, up to 30 requests per socket, and things seem to just scream!! Good luck Mike