Re: [flexcoders] Re: Flex alternatives
I very much enjoyed my time as a Flex developer, and wish the Apache project well. Some of the criticism of Adobe seems misplaced. The writing was on the wall when the most popular mobile platform (iOS) didn't allow browser plug-ins. I'm not painting Apple as a villain either. It's just business. Though Phonegap looks nice, AIR still seems pretty compelling for cross-platform mobile *app* development. It will be interesting to watch the level of investment in AIR (it is, after all, used by Adobe apps). Now that browser plug-ins are unpopular/impossible, and captive runtime is the way to go, it seems to me Adobe could open-source the AIR runtime, while still retaining control of the Flash Player browser plug-in. Wouldn't that further energize the Apache Flex project? The prospect of a cross-platform mobile app framework with the side-benefit of running on still-ubiquitous-on-desktop Flash Player? Aside: Remember OpenLaszlo? On Fri, Jan 13, 2012 at 6:28 AM, michael_reg...@dell.com wrote: ** You bring up some good points. I have been on a few calls between our company and with Adobe on this exact subject right after the announcements were made. We grilled them with questions, and though I won’t fully disclose many of their answers here, our development teams were confident enough to stick with Adobe Flex. We realize that long-term, technologies shift. I started out as a C++ developer doing low-level SCSI. Now doing UIs in Flex. In 5 years will I be doing HTML 5, ZK, Silverlight, some other new technology? Who knows. But for now, I’m actually excited to have a greater role in the direction Flex takes for now, and welcome any challenges making it Open Source may bring. ** ** The product we developed using Adobe Flex was ranked as one of the top 15 products for 2011 by CRN. Didn’t see any HTML 5 apps there. I think this says something about where Flex is, and it still holds some ground. The technology decision should be based on your projects, your long-term direction, and your talent pool. I’d recommend not jumping ship, just to jump ship. You never know where that ship may sail! ** ** *Michael J. Regert*** ** ** *From:* flexcoders@yahoogroups.com [mailto:flexcoders@yahoogroups.com] *On Behalf Of *Ron G *Sent:* Thursday, January 12, 2012 11:39 AM *To:* flexcoders@yahoogroups.com *Subject:* [flexcoders] Re: Flex alternatives ** ** Hi James, I certainly respect the decision of those who are sticking with Flex, but I would suggest that developers do so with the recognition that they may be developing with a technology that isn't going to be around that long. I could write at length about this, but, in a nutshell, here's why. On the one hand, you have an open-source project that is geared toward enterprise application development, but it is completely dependent on a proprietary runtime. That runtime is manufactured by a company who has stated its future is digital media and digital marketing, and that it believes the future of enterprise web application development is HTML5. It then begs the question, How long will they bloat their Flashplayer to support an open-source Flex community's enterprise web application development goals and wishes? To accommodate the Flex community, Adobe gets nothing in return for its expenditure of time and money in designing, developing, testing the features the Flex community requires now and in the future. It also means that, by supporting Flex in their runtime, the Flashplayer has an unnecessarily larger footprint than would otherwise be required. So, ask yourself if you truly believe Flex will be a supported product by Adobe in 5-10 years from now. I highly doubt it. On the other hand, I think if a developer uses Flash Pro to develop digital media for their applications, they can probably count on that being around indefinitely. But, not Flex. Ron --- In flexcoders@yahoogroups.com, James Ong yanlilei64@... wrote: Using ZK and Java is great. I'm still sticking to Flex for developing desktop applications and gaming. Of course, many will still using it for animations, there is no such thing as abandon, some developers are just over use Flash and end up hurting user experience than necessary. When it comes to web application, I stick to PHP but will definitely use Flex for mobile, desktop and components within the web browser. On Thu, Jan 12, 2012 at 10:16 AM, michael_regert@... wrote: ** Staying with Flex. Not looking elsewhere. ** ** *Michael* ** ** *From:* flexcoders@yahoogroups.com [mailto:flexcoders@yahoogroups.com] *On Behalf Of *Ron G *Sent:* Wednesday, January 11, 2012 8:15 PM *To:* flexcoders@yahoogroups.com *Subject:* [flexcoders] Re: Flex alternatives ** ** Yes, we have also abandoned Flex in favor of ZKoss. Since we are already a Java shop, on the
Re: [flexcoders] Re: Flex alternatives
I meant to say prospect of a cross-platform mobile app framework that doesn't depend on a proprietary runtime On Fri, Jan 13, 2012 at 9:17 AM, Richard Rodseth rrods...@gmail.com wrote: I very much enjoyed my time as a Flex developer, and wish the Apache project well. Some of the criticism of Adobe seems misplaced. The writing was on the wall when the most popular mobile platform (iOS) didn't allow browser plug-ins. I'm not painting Apple as a villain either. It's just business. Though Phonegap looks nice, AIR still seems pretty compelling for cross-platform mobile *app* development. It will be interesting to watch the level of investment in AIR (it is, after all, used by Adobe apps). Now that browser plug-ins are unpopular/impossible, and captive runtime is the way to go, it seems to me Adobe could open-source the AIR runtime, while still retaining control of the Flash Player browser plug-in. Wouldn't that further energize the Apache Flex project? The prospect of a cross-platform mobile app framework with the side-benefit of running on still-ubiquitous-on-desktop Flash Player? Aside: Remember OpenLaszlo? On Fri, Jan 13, 2012 at 6:28 AM, michael_reg...@dell.com wrote: ** You bring up some good points. I have been on a few calls between our company and with Adobe on this exact subject right after the announcements were made. We grilled them with questions, and though I won’t fully disclose many of their answers here, our development teams were confident enough to stick with Adobe Flex. We realize that long-term, technologies shift. I started out as a C++ developer doing low-level SCSI. Now doing UIs in Flex. In 5 years will I be doing HTML 5, ZK, Silverlight, some other new technology? Who knows. But for now, I’m actually excited to have a greater role in the direction Flex takes for now, and welcome any challenges making it Open Source may bring. ** ** The product we developed using Adobe Flex was ranked as one of the top 15 products for 2011 by CRN. Didn’t see any HTML 5 apps there. I think this says something about where Flex is, and it still holds some ground. The technology decision should be based on your projects, your long-term direction, and your talent pool. I’d recommend not jumping ship, just to jump ship. You never know where that ship may sail! ** ** *Michael J. Regert*** ** ** *From:* flexcoders@yahoogroups.com [mailto:flexcoders@yahoogroups.com] *On Behalf Of *Ron G *Sent:* Thursday, January 12, 2012 11:39 AM *To:* flexcoders@yahoogroups.com *Subject:* [flexcoders] Re: Flex alternatives ** ** Hi James, I certainly respect the decision of those who are sticking with Flex, but I would suggest that developers do so with the recognition that they may be developing with a technology that isn't going to be around that long. I could write at length about this, but, in a nutshell, here's why. On the one hand, you have an open-source project that is geared toward enterprise application development, but it is completely dependent on a proprietary runtime. That runtime is manufactured by a company who has stated its future is digital media and digital marketing, and that it believes the future of enterprise web application development is HTML5. It then begs the question, How long will they bloat their Flashplayer to support an open-source Flex community's enterprise web application development goals and wishes? To accommodate the Flex community, Adobe gets nothing in return for its expenditure of time and money in designing, developing, testing the features the Flex community requires now and in the future. It also means that, by supporting Flex in their runtime, the Flashplayer has an unnecessarily larger footprint than would otherwise be required. So, ask yourself if you truly believe Flex will be a supported product by Adobe in 5-10 years from now. I highly doubt it. On the other hand, I think if a developer uses Flash Pro to develop digital media for their applications, they can probably count on that being around indefinitely. But, not Flex. Ron --- In flexcoders@yahoogroups.com, James Ong yanlilei64@... wrote: Using ZK and Java is great. I'm still sticking to Flex for developing desktop applications and gaming. Of course, many will still using it for animations, there is no such thing as abandon, some developers are just over use Flash and end up hurting user experience than necessary. When it comes to web application, I stick to PHP but will definitely use Flex for mobile, desktop and components within the web browser. On Thu, Jan 12, 2012 at 10:16 AM, michael_regert@... wrote: ** Staying with Flex. Not looking elsewhere. ** ** *Michael* ** ** *From:* flexcoders@yahoogroups.com [mailto:flexcoders@yahoogroups.com] *On Behalf Of *Ron G *Sent:* Wednesday, January 11, 2012 8:15 PM
Re: [flexcoders] Re: Flex alternatives
Apache FlashKit ! On Fri, Jan 13, 2012 at 11:17 AM, Alex Harui aha...@adobe.com wrote: ** Open-sourcing AIR would be an awesome thing, but the Apache project has to assume it won’t happen. On 1/13/12 9:19 AM, Richard Rodseth rrods...@gmail.com wrote: I meant to say prospect of a cross-platform mobile app framework that doesn't depend on a proprietary runtime On Fri, Jan 13, 2012 at 9:17 AM, Richard Rodseth rrods...@gmail.com wrote: I very much enjoyed my time as a Flex developer, and wish the Apache project well. Some of the criticism of Adobe seems misplaced. The writing was on the wall when the most popular mobile platform (iOS) didn't allow browser plug-ins. I'm not painting Apple as a villain either. It's just business. Though Phonegap looks nice, AIR still seems pretty compelling for cross-platform mobile *app* development. It will be interesting to watch the level of investment in AIR (it is, after all, used by Adobe apps). Now that browser plug-ins are unpopular/impossible, and captive runtime is the way to go, it seems to me Adobe could open-source the AIR runtime, while still retaining control of the Flash Player browser plug-in. Wouldn't that further energize the Apache Flex project? The prospect of a cross-platform mobile app framework with the side-benefit of running on still-ubiquitous-on-desktop Flash Player? Aside: Remember OpenLaszlo? On Fri, Jan 13, 2012 at 6:28 AM, michael_reg...@dell.com wrote: You bring up some good points. I have been on a few calls between our company and with Adobe on this exact subject right after the announcements were made. We grilled them with questions, and though I won’t fully disclose many of their answers here, our development teams were confident enough to stick with Adobe Flex. We realize that long-term, technologies shift. I started out as a C++ developer doing low-level SCSI. Now doing UIs in Flex. In 5 years will I be doing HTML 5, ZK, Silverlight, some other new technology? Who knows. But for now, I’m actually excited to have a greater role in the direction Flex takes for now, and welcome any challenges making it Open Source may bring. The product we developed using Adobe Flex was ranked as one of the top 15 products for 2011 by CRN. Didn’t see any HTML 5 apps there. I think this says something about where Flex is, and it still holds some ground. The technology decision should be based on your projects, your long-term direction, and your talent pool. I’d recommend not jumping ship, just to jump ship. You never know where that ship may sail! *Michael J. Regert * *From:* flexcoders@yahoogroups.com [mailto:flexcoders@yahoogroups.comflexcoders@yahoogroups.com] *On Behalf Of *Ron G *Sent:* Thursday, January 12, 2012 11:39 AM *To:* flexcoders@yahoogroups.com *Subject:* [flexcoders] Re: Flex alternatives Hi James, I certainly respect the decision of those who are sticking with Flex, but I would suggest that developers do so with the recognition that they may be developing with a technology that isn't going to be around that long. I could write at length about this, but, in a nutshell, here's why. On the one hand, you have an open-source project that is geared toward enterprise application development, but it is completely dependent on a proprietary runtime. That runtime is manufactured by a company who has stated its future is digital media and digital marketing, and that it believes the future of enterprise web application development is HTML5. It then begs the question, How long will they bloat their Flashplayer to support an open-source Flex community's enterprise web application development goals and wishes? To accommodate the Flex community, Adobe gets nothing in return for its expenditure of time and money in designing, developing, testing the features the Flex community requires now and in the future. It also means that, by supporting Flex in their runtime, the Flashplayer has an unnecessarily larger footprint than would otherwise be required. So, ask yourself if you truly believe Flex will be a supported product by Adobe in 5-10 years from now. I highly doubt it. On the other hand, I think if a developer uses Flash Pro to develop digital media for their applications, they can probably count on that being around indefinitely. But, not Flex. Ron --- In flexcoders@yahoogroups.com mailto:flexcoders%40yahoogroups.comflexcoders%40yahoogroups.com , James Ong yanlilei64@... wrote: Using ZK and Java is great. I'm still sticking to Flex for developing desktop applications and gaming. Of course, many will still using it for animations, there is no such thing as abandon, some developers are just over use Flash and end up hurting user experience than necessary. When it comes to web application, I stick to PHP but will definitely use Flex for mobile, desktop and components within the web browser. On Thu
[flexcoders] Reading barcodes
Can anyone recommend a good strategy for reading barcodes on an AIR 3 mobile app? The as3 port of the zxing library does not appear to have been updated since 2009, but may work well, for all I know. But it occurs to me, that this could also be a good opportunity for a native extension. Any experiences or thoughts?
Re: [flexcoders] Reading barcodes
Yes, I clearly don't want to do this from scratch, hence the question. Any particular library you recommend? Any in AS3? On Tue, Sep 27, 2011 at 9:43 AM, James Ong yanlile...@gmail.com wrote: ** use ready made libraries from Sourceforge and you can get start faster. On Wed, Sep 28, 2011 at 12:37 AM, Richard Rodseth rrods...@gmail.comwrote: ** Can anyone recommend a good strategy for reading barcodes on an AIR 3 mobile app? The as3 port of the zxing library does not appear to have been updated since 2009, but may work well, for all I know. But it occurs to me, that this could also be a good opportunity for a native extension. Any experiences or thoughts?
Re: [flexcoders] Flex 3.5 to 4.1
So far I have compatibility mode turned on and the Halo theme added. Working through the errors. The latest is that I get a null pointer exception in HaloBorder Main Thread (Suspended: TypeError: Error #1009: Cannot access a property or method of a null object reference.) mx.skins.halo::HaloBorder/get borderMetrics mx.core::Container/get borderMetrics mx.core::Container/get viewMetrics mx.core::Container/get viewMetricsAndPadding mx.containers.utilityClasses::BoxLayout/ http://www.adobe.com/2006/flex/mx/internal::widthPadding mx.containers.utilityClasses::BoxLayout/measure mx.containers::Box/measure mx.core::UIComponent/measureSizes mx.core::UIComponent/validateSize mx.core::Container/validateSize mx.managers::LayoutManager/validateSize mx.managers::LayoutManager/doPhasedInstantiation mx.managers::LayoutManager/doPhasedInstantiationCallback This related question did not appear to get a response on the forums: http://forums.adobe.com/message/3149760 Thanks for any tips. On Mon, Oct 11, 2010 at 3:25 PM, Glenn Williams gl...@tinylion.co.ukwrote: [Attachment(s) #12b9d6686db7c478_TopText from Glenn Williams included below] hi I found it a simple process on a fairly large project we were working on. to just get it to compile you shouldn’t have any problems to then start adding spark. first change the application tag. mx: s: application that’s when the fun starts. the one thing I found useful to get your mx components looking how you expect, is to manually include the original halo css file. you can then use spark theme but still have any css of styling you’ve done on your mx components still work as you’d expect. I was about to start writing a list of ‘look outs’ but thought I’d have a look at your link first. it’s all in there, looks like a very good start anyway. our project was about 2500 source files and I think in total it was maybe a couple of days to have a spark application but, with all mx components and looking just like it was before the change. then it was just a matter of starting to replace some things. note. we did this as a test, the project was already complete and we used it to learn about the change from mx to s. it was quite simple (apart from needing to include the halo css default file by hand which caught us out. styles just weren’t doing what we expected, and as soon as we added the default.css from the halo directory all was well. One thing I will say – do it. Spark is in every way better. and once you have the luxury of starting a new project from scratch with the new spark architecture you’ll wonder how you ever managed without it. It’s a huge step forward in my opinion. it really is time to put flex 3 away and start enjoying the new framework. It’s wonderful to work with, and in all our tests just generally faster and more stable and uses less memory. In comparing the two there isn’t anything I miss about flex 3 when using flex 4. it’s worth the effort (well what little effort the change actually needs) if you have any trouble please feel free to email me off forum and we’ll help all we can. all the best glenn tiny*lion* development design *From:* flexcoders@yahoogroups.com [mailto:flexcod...@yahoogroups.com] *On Behalf Of *Richard Rodseth *Sent:* Monday, October 11, 2010 10:09 PM *To:* flexcoders@yahoogroups.com *Subject:* [flexcoders] Flex 3.5 to 4.1 Can anyone give me a sense for how big an effort it is to port from Flex 3.6 to Flex 4 SDK, without (at first) taking advantage of any of the new features. In other words, just getting everything compiling and running as well as it did before, Halo components only. This recent article is helpful, but I'd like some more experience reports. http://www.adobe.com/devnet/flex/articles/flex3_to_flex4_transitioning.html Thanks image001.jpg
[flexcoders] No openItems property on ADG?
Just when I had it working so nicely on Tree, bound to a presentation model. Is there anything equivalent for an ADG with a hierarchical data provider.
[flexcoders] Module-friendly modals
Someone on the Mate forums advocated a way to treat a modal popup as a state of a view, where the presentation model has a modalShown property, and the view observes it using bindSetter. This has the nice effect of avoiding the need to pass in Application.application to createPopUp, but it's obviously quite a bit more work to create the bubbling events and injections. And in the end you typically want to center the popup over the whole application, rather than some intermediate container. So I find myself wondering if there is a cleaner way than Application.application to refer to the root context. Follow the DisplayObject.parent chain until it's null?
Re: [flexcoders] Tree - itemClose and change
I was able to work around this issue by accessing selectedItem and updating my presentation model during the itemClose operation. But to answer your question, it looks like I get two valueCommits when closing a node. On Thu, Sep 2, 2010 at 10:58 PM, Alex Harui aha...@adobe.com wrote: I thought there was already a bug on that. Do you get a valueCommit? On 9/2/10 4:23 PM, Richard Rodseth rrods...@gmail.com wrote: It appears that when you close a node in a tree control, the selection (of a contained node) is lost, but no change event is fired. Does this sound like a bug? I'm trying to restore tree state (selection and open nodes) across a service call. I've got the open nodes working pretty well and the refresh use case also mostly works (I store the selection as guids from the objects in question and find them again after the refresh). But it seems I will have to handle openItem and closeItem, and clear the guid-based selection if the node is contained in the path being closed. -- Alex Harui Flex SDK Team Adobe System, Inc. http://blogs.adobe.com/aharui
Re: [flexcoders] Tree - itemClose and change
Thanks. I'll consider using valueCommit in this case. I'd only used it in the context of things like text fields before. On Fri, Sep 3, 2010 at 1:06 PM, Alex Harui aha...@adobe.com wrote: ValueCommit is supposedly for “any changes to selection”, the change event is supposedly only for user-initiated changes to selection. The binding system listens to both. The change of selection on close is currently seen as programmatic since the user didn’t directly do it, but I think there is already a bug that argues that change should fire as well. On 9/3/10 11:12 AM, Richard Rodseth rrods...@gmail.com wrote: I was able to work around this issue by accessing selectedItem and updating my presentation model during the itemClose operation. But to answer your question, it looks like I get two valueCommits when closing a node. On Thu, Sep 2, 2010 at 10:58 PM, Alex Harui aha...@adobe.com wrote: I thought there was already a bug on that. Do you get a valueCommit? On 9/2/10 4:23 PM, Richard Rodseth rrods...@gmail.com http://rrods...@gmail.com wrote: It appears that when you close a node in a tree control, the selection (of a contained node) is lost, but no change event is fired. Does this sound like a bug? I'm trying to restore tree state (selection and open nodes) across a service call. I've got the open nodes working pretty well and the refresh use case also mostly works (I store the selection as guids from the objects in question and find them again after the refresh). But it seems I will have to handle openItem and closeItem, and clear the guid-based selection if the node is contained in the path being closed. -- Alex Harui Flex SDK Team Adobe System, Inc. http://blogs.adobe.com/aharui
[flexcoders] Tree - itemClose and change
It appears that when you close a node in a tree control, the selection (of a contained node) is lost, but no change event is fired. Does this sound like a bug? I'm trying to restore tree state (selection and open nodes) across a service call. I've got the open nodes working pretty well and the refresh use case also mostly works (I store the selection as guids from the objects in question and find them again after the refresh). But it seems I will have to handle openItem and closeItem, and clear the guid-based selection if the node is contained in the path being closed.
[flexcoders] Flex 3.5 - 4 upgrade risk
My colleagues in QA are having better results with Selenium + FlexMonkium than with FlexMonkey + 3.5. But FlexMonkium apparently only supports Flex 4. I have been cautious/wary about proposing an upgrade to Flex 4 at this time, particularly since I encountered a ComboBox bug introduced in 3.5 when I upgraded to get fixes for charting bugs in 3.4. But a solid automation story would be quite compelling. So I guess I'm looking for reassurance, and also wondering a few things. Is there some bug fix in automation libraries that makes it more robust even if you're using Halo components? What's compatibility mode and why would I use it? Thanks.
[flexcoders] Image artifact
I'm trying to do a HDividedBox where one of the segments is collapsible. But
in the open state, an artifact of the expand icon image appears at top left
even though its container is invisible in that state, and it's normal
position is halfway down the screen.
Code is something like the following.
Note that the expand icon is outside the HDvidedBox, and the collapse icon
is inside.
Any ideas?
mx:HBox width=100% height=100% borderStyle=solid
mx:VBox visible={!(model.navigatorShown)}}
includeInLayout={!(model.navigatorShown)} height=100%
mx:Spacer height=100%/
mx:Image id=leftGrip
width=16 height=16
click=leftGrip_clickHandler(event)
source=@Embed('/assets/images/icon_open_right.png')/
mx:Spacer height=100%/
/mx:VBox
mx:HDividedBox id=dividedBox width=100% height=100%
mx:VBox width=200 height=100%
visible={model.navigatorShown} includeInLayout={model.navigatorShown}
!-- navigator plus collapse icon --
mx:HBox width=100% height=100%
mx:VBox width=100% height=100% borderStyle=solid
mx:Label text=Navigator/
/mx:VBox
mx:VBox visible={model.navigatorShown}}
includeInLayout={model.navigatorShown} height=100%
mx:Spacer height=100%/
mx:Image id=rightGrip width=16 height=16
click=rightGrip_clickHandler(event)
source=@Embed('/assets/images/icon_close_left.png')/
mx:Spacer height=100%/
/mx:VBox
/mx:HBox
/mx:VBox
mx:VBox width=80% height=100% borderStyle=solid
mx:Label text=RHS/
/mx:VBox
/mx:HDividedBox
/mx:HBox
[flexcoders] ComboBox fiasco
Back in ComboBox hell, and finally read this: https://bugs.adobe.com/jira/browse/SDK-25705 I've always had problems with comboboxes and binding, but didn't realize this particular bug was *introduced* in 3.5, rather than having been there all along. Of course I updated our application to use 3.5 because of bad bugs in charting. I'm glad bug fixes continue in the 3.x branch, but it's absolutely critical that there are no regressions. So while I hope 3.6 is imminent, let's hope it's well-tested. I trust the testing infrastructure and architecture has been improved for 4.x.
Re: [flexcoders] Re: Refresh list with itemRenderer/ComboBox
The workaround for ComboBox bug
https://bugs.adobe.com/jira/browse/SDK-25705
suggested by Ben Atkins worked for me. In fact, since I didn't have the
resize issues, I reduced it to the following.
I sent a separate nastygram (sorry) about regressions in the 3.x branch.
public class ComboBoxPatched extends ComboBox
{
public function ComboBoxPatched()
{
super();
}
//This addresses the problem where the list base wasn't reflecting
changes to the data provider
override public function set dataProvider(value:Object):void
{
super.dataProvider = value;
if (dropdown) {
dropdown.dataProvider = value;
}
}
}
On Thu, Jul 29, 2010 at 4:25 PM, Richard Rodseth rrods...@gmail.com wrote:
Yes, that was my plan, but I've since discovered that I misspoke and my
item renderer has a PopUpButton rather than a ComboBox (sorry!)
Looking into other possible mistakes in my binding.
On Thu, Jul 29, 2010 at 3:31 PM, Amy amyblankens...@bellsouth.net wrote:
--- In flexcoders@yahoogroups.com flexcoders%40yahoogroups.com, Richard
Rodseth rrods...@... wrote:
Hmm. I did a diff on the 3.x and 3.5 versions of ComboBase and ComboBox,
and
didn't see anything that seemed relevant - just some stuff about
tweening.
Richard, have you considered subclassing and just fixing it yourself
(assuming the part you need to fix isn't private, of course).
-Amy
Re: [flexcoders] Re: Refresh list with itemRenderer/ComboBox
Yes, that was my plan, but I've since discovered that I misspoke and my item renderer has a PopUpButton rather than a ComboBox (sorry!) Looking into other possible mistakes in my binding. On Thu, Jul 29, 2010 at 3:31 PM, Amy amyblankens...@bellsouth.net wrote: --- In flexcoders@yahoogroups.com flexcoders%40yahoogroups.com, Richard Rodseth rrods...@... wrote: Hmm. I did a diff on the 3.x and 3.5 versions of ComboBase and ComboBox, and didn't see anything that seemed relevant - just some stuff about tweening. Richard, have you considered subclassing and just fixing it yourself (assuming the part you need to fix isn't private, of course). -Amy
Re: [flexcoders] Refresh list with itemRenderer/ComboBox
True - it's given me many headaches. Any ideas for a workaround? A judiciously-placed callLater ? Some combination of invalidateXX ? Even something like this doesn't work: var source:Array = conditions.source; conditions.source = []; conditions.source = source; conditionList.invalidateList(); On Tue, Jul 27, 2010 at 10:12 PM, Alex Harui aha...@adobe.com wrote: Flex 3.5 has some issues with ComboBox updating. On 7/27/10 4:22 PM, Richard Rodseth rrods...@gmail.com wrote: I have two scenarios where I have to refresh rows of a list or datagrid because the contents are computed, and one of the inputs to the computation (some meta-information) has changed. In the first (a DataGrid) I just have a labelFunction for one of the columns, and calling refresh() on the ArrayCollection in the presentation model does the trick. The second is more complicated - the list has an inline item renderer which is really a mini-editor containing a combo box. I'm struggling to get the combobox label to update. Tried calling refresh() on the array collection, invalidateList() on the list, and even executeBindings() on the list, to no avail. Binding works in all other use cases, eg. inserting and deleting items from the list. Any special considerations with an inline item renderer or combobox when forcing a refresh? As an aside, this description from the LiveDocs for executeBindings is a bit odd. Executes the data bindings into this UIComponent object. Workaround for MXML container/bindings problem (177074): override Container.executeBindings() to prefer descriptor.document over parentDocument in the call to BindingManager.executeBindings(). This should always provide the correct behavior for instances created by descriptor, and will provide the original behavior for procedurally-created instances. (The bug may or may not appear in the latter case.) A more complete fix, guaranteeing correct behavior in both non-DI and reparented-component scenarios, is anticipated for updater 1. -- Alex Harui Flex SDK Team Adobe System, Inc. http://blogs.adobe.com/aharui
Re: [flexcoders] Refresh list with itemRenderer/ComboBox
Good to know. Of course, in the real world, we can't update on a whim, so any other suggestions are welcome. I may be able to get the bug deferred. On Wed, Jul 28, 2010 at 9:58 AM, Alex Harui aha...@adobe.com wrote: 3.6 is due out soon. It is fixed in that code base. I think there may be nightly build on the 3.x branch. On 7/28/10 9:23 AM, Richard Rodseth rrods...@gmail.com wrote: True - it's given me many headaches. Any ideas for a workaround? A judiciously-placed callLater ? Some combination of invalidateXX ? Even something like this doesn't work: var source:Array = conditions.source; conditions.source = []; conditions.source = source; conditionList.invalidateList(); On Tue, Jul 27, 2010 at 10:12 PM, Alex Harui aha...@adobe.com wrote: Flex 3.5 has some issues with ComboBox updating. On 7/27/10 4:22 PM, Richard Rodseth rrods...@gmail.com http://rrods...@gmail.com wrote: I have two scenarios where I have to refresh rows of a list or datagrid because the contents are computed, and one of the inputs to the computation (some meta-information) has changed. In the first (a DataGrid) I just have a labelFunction for one of the columns, and calling refresh() on the ArrayCollection in the presentation model does the trick. The second is more complicated - the list has an inline item renderer which is really a mini-editor containing a combo box. I'm struggling to get the combobox label to update. Tried calling refresh() on the array collection, invalidateList() on the list, and even executeBindings() on the list, to no avail. Binding works in all other use cases, eg. inserting and deleting items from the list. Any special considerations with an inline item renderer or combobox when forcing a refresh? As an aside, this description from the LiveDocs for executeBindings is a bit odd. Executes the data bindings into this UIComponent object. Workaround for MXML container/bindings problem (177074): override Container.executeBindings() to prefer descriptor.document over parentDocument in the call to BindingManager.executeBindings(). This should always provide the correct behavior for instances created by descriptor, and will provide the original behavior for procedurally-created instances. (The bug may or may not appear in the latter case.) A more complete fix, guaranteeing correct behavior in both non-DI and reparented-component scenarios, is anticipated for updater 1. -- Alex Harui Flex SDK Team Adobe System, Inc. http://blogs.adobe.com/aharui
Re: [flexcoders] Refresh list with itemRenderer/ComboBox
Hmm. I did a diff on the 3.x and 3.5 versions of ComboBase and ComboBox, and didn't see anything that seemed relevant - just some stuff about tweening. On Wed, Jul 28, 2010 at 10:09 AM, Richard Rodseth rrods...@gmail.comwrote: Good to know. Of course, in the real world, we can't update on a whim, so any other suggestions are welcome. I may be able to get the bug deferred. On Wed, Jul 28, 2010 at 9:58 AM, Alex Harui aha...@adobe.com wrote: 3.6 is due out soon. It is fixed in that code base. I think there may be nightly build on the 3.x branch. On 7/28/10 9:23 AM, Richard Rodseth rrods...@gmail.com wrote: True - it's given me many headaches. Any ideas for a workaround? A judiciously-placed callLater ? Some combination of invalidateXX ? Even something like this doesn't work: var source:Array = conditions.source; conditions.source = []; conditions.source = source; conditionList.invalidateList(); On Tue, Jul 27, 2010 at 10:12 PM, Alex Harui aha...@adobe.com wrote: Flex 3.5 has some issues with ComboBox updating. On 7/27/10 4:22 PM, Richard Rodseth rrods...@gmail.com http://rrods...@gmail.com wrote: I have two scenarios where I have to refresh rows of a list or datagrid because the contents are computed, and one of the inputs to the computation (some meta-information) has changed. In the first (a DataGrid) I just have a labelFunction for one of the columns, and calling refresh() on the ArrayCollection in the presentation model does the trick. The second is more complicated - the list has an inline item renderer which is really a mini-editor containing a combo box. I'm struggling to get the combobox label to update. Tried calling refresh() on the array collection, invalidateList() on the list, and even executeBindings() on the list, to no avail. Binding works in all other use cases, eg. inserting and deleting items from the list. Any special considerations with an inline item renderer or combobox when forcing a refresh? As an aside, this description from the LiveDocs for executeBindings is a bit odd. Executes the data bindings into this UIComponent object. Workaround for MXML container/bindings problem (177074): override Container.executeBindings() to prefer descriptor.document over parentDocument in the call to BindingManager.executeBindings(). This should always provide the correct behavior for instances created by descriptor, and will provide the original behavior for procedurally-created instances. (The bug may or may not appear in the latter case.) A more complete fix, guaranteeing correct behavior in both non-DI and reparented-component scenarios, is anticipated for updater 1. -- Alex Harui Flex SDK Team Adobe System, Inc. http://blogs.adobe.com/aharui
Re: [flexcoders] Tree click vs change
Thanks. This worked well. On Fri, Jul 2, 2010 at 2:43 PM, Alex Harui aha...@adobe.com wrote: Add a high priority mouseDown handler and record the selectedIndex then. On 7/2/10 2:34 PM, Richard Rodseth rrods...@gmail.com wrote: We have a datagrid that gets updated when the user clicks a node in a tree (or moves through it using arrow keys). We're handling the change event in the tree to issue the update request. A request has come in that clicking an already selected node should do a refresh. We tried having both change and click handlers, where the click handler would check if the selection was unchanged. But that didn't work because the change event fires first so the selection reported a click time is always the same. Any suggestions? -- Alex Harui Flex SDK Team Adobe System, Inc. http://blogs.adobe.com/aharui
[flexcoders] Refresh list with itemRenderer/ComboBox
I have two scenarios where I have to refresh rows of a list or datagrid because the contents are computed, and one of the inputs to the computation (some meta-information) has changed. In the first (a DataGrid) I just have a labelFunction for one of the columns, and calling refresh() on the ArrayCollection in the presentation model does the trick. The second is more complicated - the list has an inline item renderer which is really a mini-editor containing a combo box. I'm struggling to get the combobox label to update. Tried calling refresh() on the array collection, invalidateList() on the list, and even executeBindings() on the list, to no avail. Binding works in all other use cases, eg. inserting and deleting items from the list. Any special considerations with an inline item renderer or combobox when forcing a refresh? As an aside, this description from the LiveDocs for executeBindings is a bit odd. Executes the data bindings into this UIComponent object. Workaround for MXML container/bindings problem (177074): override Container.executeBindings() to prefer descriptor.document over parentDocument in the call to BindingManager.executeBindings(). This should always provide the correct behavior for instances created by descriptor, and will provide the original behavior for procedurally-created instances. (The bug may or may not appear in the latter case.) A more complete fix, guaranteeing correct behavior in both non-DI and reparented-component scenarios, is anticipated for updater 1.
Re: [flexcoders] Refresh list with itemRenderer/ComboBox
I'm not re-populating the array. Imagine it's a list of employees, and one
of the properties is a job id. When displaying the list, the job title is
looked up (eg. by a labelFunction on a DataGridColumn). Now, imagine the job
title is edited elsewhere in the application. When we detect this (eg. via a
listener) we can just call refresh() on the employee array, and the
itemrenderers will run again.
The second scenario I alluded to would be something like having a combobox
in the grid cell for changing the job type. Once again the job title needs
to be re-displayed if it is changed elsewhere, but in this case it is the
dynamically-computed label of the combobox.
On Tue, Jul 27, 2010 at 6:21 PM, Amanda Machutta amachu...@gmail.comwrote:
Hi Richard,
How are you refreshing the ArrayCollection? If you are doing something like
this:
oldArr = newArry;
then its not going to work because of the way binding works. If you haven't
done it already, try this:
oldArr.removeAll();
for(var i:int=0; inewArr.length; i++){
oldArr.addItem(newArr.getItemAt(i));
}
Good luck!
Amanda
On Tue, Jul 27, 2010 at 7:22 PM, Richard Rodseth rrods...@gmail.comwrote:
I have two scenarios where I have to refresh rows of a list or
datagrid because the contents are computed, and one of the inputs to
the computation (some meta-information) has changed.
In the first (a DataGrid) I just have a labelFunction for one of the
columns, and calling refresh() on the ArrayCollection in the
presentation model does the trick.
The second is more complicated - the list has an inline item renderer
which is really a mini-editor containing a combo box. I'm struggling
to get the combobox label to update. Tried calling refresh() on the
array collection, invalidateList() on the list, and even
executeBindings() on the list, to no avail. Binding works in all other
use cases, eg. inserting and deleting items from the list.
Any special considerations with an inline item renderer or combobox
when forcing a refresh?
As an aside, this description from the LiveDocs for executeBindings is
a bit odd.
Executes the data bindings into this UIComponent object. Workaround
for MXML container/bindings problem (177074): override
Container.executeBindings() to prefer descriptor.document over
parentDocument in the call to BindingManager.executeBindings(). This
should always provide the correct behavior for instances created by
descriptor, and will provide the original behavior for
procedurally-created instances. (The bug may or may not appear in the
latter case.) A more complete fix, guaranteeing correct behavior in
both non-DI and reparented-component scenarios, is anticipated for
updater 1.
--
´¨) __o
.·´ .·´¨)¸.·´¨) _'\ .
(¸.·´ (¸.·´ (¸.·´¨¨ Amanda ¨¨( * ) ( )
Re: [flexcoders] Value Object --- Object Proxy --- huh???
Not sure it's worth the trouble, but did you have some particular technique in mind? I think I might have asked about this before, but if MXML had a way to do constructor injection it could also help frameworks like Mate so they don't need tags with attributes called arguments. On Fri, Jul 23, 2010 at 3:46 PM, Alex Harui aha...@adobe.com wrote: You could write your set functions to be write-once. On 7/23/10 11:35 AM, Richard Rodseth rrods...@gmail.com wrote: Getting a little off-topic, but in case anyone's interested: http://www.developer.com/lang/other/article.php/3874551/Clojure-Immutability-at-the-Language-Level.htm One Flex-specific wrinkle is that if you define a class with a 1 argument constructor and getters only, you can't instantiate it in MXML. And in AS3, you can't have mutliple constructors. I'm not sure, but I imagine some of the serialization libraries for JSON etc might have issues too if you don't have the zero-argument constructor. On Fri, Jul 23, 2010 at 11:25 AM, Richard Rodseth rrods...@gmail.com wrote: Yes, obviously. But not the same thing. If you write your VO’s to have only get functions, they are effectively immutable. On 7/23/10 8:56 AM, Richard Rodseth rrods...@gmail.com http://rrods...@gmail.com wrote: Interesting. I haven't used that one either. The docs emphasize the addEventListener capability, which I guess distinguishes this from plain old Object, which can also hold arbitrary properties. http://livedocs.adobe.com/flex/3/langref/mx/utils/ObjectProxy.html As an aside, I would love to have language support for designating VOs as immutable. On Thu, Jul 22, 2010 at 6:11 PM, Alex Harui aha...@adobe.com http://aha...@adobe.com wrote: ObjectProxy is a generic VO. It is dynamic and can hold any properties and values. And it is slow. Best to create custom Vos that are strongly typed. On 7/22/10 2:52 PM, mark.embrey mark.c.emb...@gmail.com http://mark.c.emb...@gmail.com http://mark.c.emb...@gmail.com wrote: I'm trying to wrap my feeble brain around these concepts. Are they the same? -similar? -completely different? Your feedback will be appreciated by this greenhorn. Mark -- Alex Harui Flex SDK Team Adobe System, Inc. http://blogs.adobe.com/aharui
Re: [flexcoders] Value Object --- Object Proxy --- huh???
Interesting. I haven't used that one either. The docs emphasize the addEventListener capability, which I guess distinguishes this from plain old Object, which can also hold arbitrary properties. http://livedocs.adobe.com/flex/3/langref/mx/utils/ObjectProxy.html As an aside, I would love to have language support for designating VOs as immutable. On Thu, Jul 22, 2010 at 6:11 PM, Alex Harui aha...@adobe.com wrote: ObjectProxy is a generic VO. It is dynamic and can hold any properties and values. And it is slow. Best to create custom Vos that are strongly typed. On 7/22/10 2:52 PM, mark.embrey mark.c.emb...@gmail.com wrote: I'm trying to wrap my feeble brain around these concepts. Are they the same? -similar? -completely different? Your feedback will be appreciated by this greenhorn. Mark -- Alex Harui Flex SDK Team Adobe System, Inc. http://blogs.adobe.com/aharui
Re: [flexcoders] Value Object --- Object Proxy --- huh???
Yes, obviously. But not the same thing. If you write your VO’s to have only get functions, they are effectively immutable. On 7/23/10 8:56 AM, Richard Rodseth rrods...@gmail.com wrote: Interesting. I haven't used that one either. The docs emphasize the addEventListener capability, which I guess distinguishes this from plain old Object, which can also hold arbitrary properties. http://livedocs.adobe.com/flex/3/langref/mx/utils/ObjectProxy.html As an aside, I would love to have language support for designating VOs as immutable. On Thu, Jul 22, 2010 at 6:11 PM, Alex Harui aha...@adobe.com wrote: ObjectProxy is a generic VO. It is dynamic and can hold any properties and values. And it is slow. Best to create custom Vos that are strongly typed. On 7/22/10 2:52 PM, mark.embrey mark.c.emb...@gmail.com http://mark.c.emb...@gmail.com wrote: I'm trying to wrap my feeble brain around these concepts. Are they the same? -similar? -completely different? Your feedback will be appreciated by this greenhorn. Mark -- Alex Harui Flex SDK Team Adobe System, Inc. http://blogs.adobe.com/aharui
Re: [flexcoders] Value Object --- Object Proxy --- huh???
Getting a little off-topic, but in case anyone's interested: http://www.developer.com/lang/other/article.php/3874551/Clojure-Immutability-at-the-Language-Level.htm One Flex-specific wrinkle is that if you define a class with a 1 argument constructor and getters only, you can't instantiate it in MXML. And in AS3, you can't have mutliple constructors. I'm not sure, but I imagine some of the serialization libraries for JSON etc might have issues too if you don't have the zero-argument constructor. On Fri, Jul 23, 2010 at 11:25 AM, Richard Rodseth rrods...@gmail.comwrote: Yes, obviously. But not the same thing. If you write your VO’s to have only get functions, they are effectively immutable. On 7/23/10 8:56 AM, Richard Rodseth rrods...@gmail.com wrote: Interesting. I haven't used that one either. The docs emphasize the addEventListener capability, which I guess distinguishes this from plain old Object, which can also hold arbitrary properties. http://livedocs.adobe.com/flex/3/langref/mx/utils/ObjectProxy.html As an aside, I would love to have language support for designating VOs as immutable. On Thu, Jul 22, 2010 at 6:11 PM, Alex Harui aha...@adobe.com wrote: ObjectProxy is a generic VO. It is dynamic and can hold any properties and values. And it is slow. Best to create custom Vos that are strongly typed. On 7/22/10 2:52 PM, mark.embrey mark.c.emb...@gmail.com http://mark.c.emb...@gmail.com wrote: I'm trying to wrap my feeble brain around these concepts. Are they the same? -similar? -completely different? Your feedback will be appreciated by this greenhorn. Mark -- Alex Harui Flex SDK Team Adobe System, Inc. http://blogs.adobe.com/aharui
[flexcoders] Button with tooltip but no overskin
Is there a simple way to set up a button so that it's appearance doesn't change on rollOver, but the tooltip does appear? I am simulating a menu bar of sorts using a horizontal layout of PopUpButtons and Buttons. The Buttons are just filler that need to follow the style of the others, but have no associated action. But some have meaningful labels and tooltips. In other words, I need to create a dead button with the same style as others, and a tooltip.
[flexcoders] Re: Button with tooltip but no overskin
I should have noted that this is for Flex 3. I created a subclass of ButtonSkin which overrides name() to return upSkin. If anyone has a more elegant solution please let me know. On Thu, Jul 22, 2010 at 12:35 PM, Richard Rodseth rrods...@gmail.com wrote: Is there a simple way to set up a button so that it's appearance doesn't change on rollOver, but the tooltip does appear? I am simulating a menu bar of sorts using a horizontal layout of PopUpButtons and Buttons. The Buttons are just filler that need to follow the style of the others, but have no associated action. But some have meaningful labels and tooltips. In other words, I need to create a dead button with the same style as others, and a tooltip.
Re: [flexcoders] Value Object --- Object Proxy --- huh???
Hi Mark In the fairly large applications I've worked on, I've not had occasion to use Proxy, if you mean this: http://livedocs.adobe.com/flex/3/langref/flash/utils/Proxy.html Basically it lets you create an object that poses as something else, possibly supplementing it. For example, in the example given, they create a ProxyArray which actually derives from Proxy, but (dynamically) implements the properties, operators and methods of Array, adding sum() and clear() methods. Others may disagree, and I'd love to hear of good examples, but I would actually think twice before using this. You're sacrificing type safety and possibly performance. On the other hand, Value Objects (closely related to DTOs [Data Transfer Objects], and to be contrasted with Entity objects), are bread and butter things. http://c2.com/cgi/wiki?ValueObject Their identity is based on their state rather than on their object identity. This way, you can have multiple copies of the same conceptual value object i.e. two Date instances with the same day, month etc. represent the same date, whereas var employee:Employee might refer to different Employee instances at different times, in which case we're talking about different Employees. In a Flex app, VOs are typically created to represent messages. Your API (whether AMF, JSON or XML) will return flattened data which is then converted into an object instance tree structure (not to be confused with a class inheritance tree, if you're new to OOP). If you are creating these VO classes yourself (as opposed to using ones whose source is generated by tools provided by LiveCycle, WebOrb etc) you should resist the impulse to add business logic to them. They'll typically have nothing more than a few public properties. So you might have both an EmployeeVO and an Employee class, where EmployeeVO is a Value Object or DTO, and Employee is an entity. Hope this helps. I'm sure others will weigh in. - Richard P.S. It's perhaps more appropriate to server-side work, but I love Eric Evans' book on Domain-Driven-Design http://domaindrivendesign.org/ On Thu, Jul 22, 2010 at 2:52 PM, mark.embrey mark.c.emb...@gmail.comwrote: I'm trying to wrap my feeble brain around these concepts. Are they the same? -similar? -completely different? Your feedback will be appreciated by this greenhorn. Mark
[flexcoders] DateTimeAxis labels
My stakeholders almost always seem to want something for DateTimeAxis labels that I can't give them out of the box. Example: I have a weeks worth of data in 15-minute increments, and would like the labels at 12AM and 12PM. Or if only one per day, at 12PM rather than 12AM. But if I set labelUnits to hours, I get one label per day at 12AM. Anyone discovered any tricks?
Re: [flexcoders] DateTimeAxis labels
Before embarking on potentially time-consuming or risky customization, we usually have to weigh other priorities. But thanks, I'll keep that in mind, and look into it when time permits. On Wed, Jul 14, 2010 at 3:30 PM, Brendan Meutzner bmeutz...@gmail.comwrote: Have you looked at using a custom IAxisRenderer for the series in which you limit the labels which are shown? You would be able see the labels which are to be displayed and show/hide them accordingly. I believe that you'll find the setter for gutters has the labels which are added automatically. Brendan On Wed, Jul 14, 2010 at 4:56 PM, Richard Rodseth rrods...@gmail.comwrote: My stakeholders almost always seem to want something for DateTimeAxis labels that I can't give them out of the box. Example: I have a weeks worth of data in 15-minute increments, and would like the labels at 12AM and 12PM. Or if only one per day, at 12PM rather than 12AM. But if I set labelUnits to hours, I get one label per day at 12AM. Anyone discovered any tricks?
[flexcoders] Flashbuilder 4.0.1 debugger issue?
A colleague is experiencing a problem where our app hangs the browser when run in the debugger, but not when launched in non-debug mode (from Flashbuilder). He's the only one on the team who started Flex work since the release of 4.0.1 so I'm wondering if there are any known problems that could explain this.
[flexcoders] Re: Flashbuilder 4.0.1 debugger issue?
He's also on Flash Player 10.1 and after deleting shared objects used by our splash screen, the app is locking up earlier (i.e. before dismissing the splash screen. Are there some changes I should know about relating to 10.1, security sandboxes, Shared Objects and the 4.0.1 debugger? On Mon, Jul 12, 2010 at 2:41 PM, Richard Rodseth rrods...@gmail.com wrote: A colleague is experiencing a problem where our app hangs the browser when run in the debugger, but not when launched in non-debug mode (from Flashbuilder). He's the only one on the team who started Flex work since the release of 4.0.1 so I'm wondering if there are any known problems that could explain this.
[flexcoders] Flashbuilder hang loading workbench
This is happening again this morning, and I don't remember how I got past it before. Flashbuilder hangs at the start up splash screen (before workspace selection even) while displaying Loading workbench. OS X spinning beachball. Must force quit. Launching with -clean didn't help. Anyone else?
[flexcoders] Re: Flashbuilder hang loading workbench
Deleting the .snap file mentioned here did the trick: http://www.michaellabieniec.com/?p=51 On Fri, Jul 9, 2010 at 10:08 AM, Richard Rodseth rrods...@gmail.com wrote: This is happening again this morning, and I don't remember how I got past it before. Flashbuilder hangs at the start up splash screen (before workspace selection even) while displaying Loading workbench. OS X spinning beachball. Must force quit. Launching with -clean didn't help. Anyone else?
[flexcoders] DateTimeAxis - day + hour:minute
I have a Flex 3 LineChart which displays data for one day or one week, in 15-minute increments. In the week version I need the axis labels to include day name (not specific date) + hour:minute, or perhaps someway to indicate the day only on the 0:00 hour:minute label. Thanks for any suggestions.
[flexcoders] Mate: listener injector vs manager/PM binding
Design question for any other Mate users out there. Will cross-post to Mate forum. I use Mate with the Manager + Presentation Model pattern. i.e. presentation model dispatches bubbling events. Mate event map invokes services calls, sets result on manager. Properties are injected from manager to presentation model. I find it very convenient to have a data status property in the presentation model for each service call, with values NOT_CURRENT, LOADING, LOADED and ERROR. Then the view can display appropriate status and progress indicators based on this property. One approach (a) is to have the event map set the status on the manager and have the map inject that status into the presentation model. Another (b) is to set the status in the presentation model just before dispatching the event, and when receiving the result (i.e. in a custom setter). I like the directness of (b), but it doesn't cover the error case unless the manager also has something like a faultevent property that also gets injected into the presentation model. A third approach (c) would use listener injectors. The map would inject a listener into the presentation model and would announce a special failure event. Thoughts?
[flexcoders] Tree click vs change
We have a datagrid that gets updated when the user clicks a node in a tree (or moves through it using arrow keys). We're handling the change event in the tree to issue the update request. A request has come in that clicking an already selected node should do a refresh. We tried having both change and click handlers, where the click handler would check if the selection was unchanged. But that didn't work because the change event fires first so the selection reported a click time is always the same. Any suggestions?
Re: [flexcoders] Re: Mate: listener injector vs manager/PM binding
The map shouldn't interact with the PM (other than building it with ObjectBuilder and injecting it into the view). I'm leaning towards a). That way it's centralized and manager properties can potentially be injected into multiple presentation models. On Fri, Jul 2, 2010 at 1:56 PM, turbo_vb timh...@aol.com wrote: I usually use (b). For the error, you can call a public method in the PM from the faultHandler in the map; as one alternative. --- In flexcoders@yahoogroups.com flexcoders%40yahoogroups.com, Richard Rodseth rrods...@... wrote: Design question for any other Mate users out there. Will cross-post to Mate forum. I use Mate with the Manager + Presentation Model pattern. i.e. presentation model dispatches bubbling events. Mate event map invokes services calls, sets result on manager. Properties are injected from manager to presentation model. I find it very convenient to have a data status property in the presentation model for each service call, with values NOT_CURRENT, LOADING, LOADED and ERROR. Then the view can display appropriate status and progress indicators based on this property. One approach (a) is to have the event map set the status on the manager and have the map inject that status into the presentation model. Another (b) is to set the status in the presentation model just before dispatching the event, and when receiving the result (i.e. in a custom setter). I like the directness of (b), but it doesn't cover the error case unless the manager also has something like a faultevent property that also gets injected into the presentation model. A third approach (c) would use listener injectors. The map would inject a listener into the presentation model and would announce a special failure event. Thoughts?
Re: [flexcoders] Re: Mate: listener injector vs manager/PM binding
Haven't looked at Swiz yet, but as I mentioned, I can certainly announce an event in the faultHandler and inject a listener for it in the PM. On Fri, Jul 2, 2010 at 3:36 PM, turbo_vb timh...@aol.com wrote: I agree, but it can; since the map knows about the PM anyway. This is where Swiz has an advantage with the Mediate metadata. In that case the PM method can just listens for an event; which can be dispatched on fault. -TH --- In flexcoders@yahoogroups.com flexcoders%40yahoogroups.com, Richard Rodseth rrods...@... wrote: The map shouldn't interact with the PM (other than building it with ObjectBuilder and injecting it into the view). I'm leaning towards a). That way it's centralized and manager properties can potentially be injected into multiple presentation models. On Fri, Jul 2, 2010 at 1:56 PM, turbo_vb timh...@... wrote: I usually use (b). For the error, you can call a public method in the PM from the faultHandler in the map; as one alternative. --- In flexcoders@yahoogroups.com flexcoders%40yahoogroups.comflexcoders% 40yahoogroups.com, Richard Rodseth rrodseth@ wrote: Design question for any other Mate users out there. Will cross-post to Mate forum. I use Mate with the Manager + Presentation Model pattern. i.e. presentation model dispatches bubbling events. Mate event map invokes services calls, sets result on manager. Properties are injected from manager to presentation model. I find it very convenient to have a data status property in the presentation model for each service call, with values NOT_CURRENT, LOADING, LOADED and ERROR. Then the view can display appropriate status and progress indicators based on this property. One approach (a) is to have the event map set the status on the manager and have the map inject that status into the presentation model. Another (b) is to set the status in the presentation model just before dispatching the event, and when receiving the result (i.e. in a custom setter). I like the directness of (b), but it doesn't cover the error case unless the manager also has something like a faultevent property that also gets injected into the presentation model. A third approach (c) would use listener injectors. The map would inject a listener into the presentation model and would announce a special failure event. Thoughts?
Re: [flexcoders] Bug in Flex Chart in Flash Builder?
Funny you should mention this. I just encountered an issue where the axis would scale to 100 after a resize (in my case the chart is being re-parented in a different layout with more space). I worked around it by computing the maximum myself in my presentation model, and binding the axis maximum property to my computed value. On Thu, Jun 10, 2010 at 8:16 AM, Raymond Brown silenttr...@yahoo.comwrote: I have created a simple chart in flex, if I resize the window by dragging the corner edge everything works. If I try to click a button and change the graph screen width and height via a state the Y scale of the graph resets to 0 through 100. When my data is between 0 and 3 obviously this doesn't work. The only way to fix it is reload the entire page which stinks for user interaction. I have asked this question before on this list with no answer - am I the only person hitting this or rather using flex charts and changing their size via a state? i.e. ...width.state1=100% width.state2=250... I capture the resize event and tried to invalidate the display list and refresh the data provider but nothing works. -r
Re: [flexcoders] Blocking
I was able to make an adjustment to the surrounding code, so this is no longer an issue for me, but thanks for the information. On Thu, Jun 3, 2010 at 3:55 AM, Oleg Sivokon olegsivo...@gmail.com wrote: You can load it with JavaScript synchronously, ExternalInterface is also synchronous, however, it is possible that you will run out of time and the application will crash because of that. Because, even though you can try-catch timeout error, if the function that produced the error won't exit, you will get another error that you cannot try-catch.
[flexcoders] Blocking
I know this sounds terrible, but is there a way to wrap an asynchronous call in one which blocks while waiting for it to complete? I'm loading XML via URLLoader to configure a mock API implementation and am limited in what I can change in the main application architecture. So I need a way to block while loading the configuration file. I've tried setting a _loadState variable and looping while it's set to loading, but that didn't seem to do the trick.
Re: [flexcoders] Blocking
Thanks. I guess it makes sense now that I think about it. On Wed, Jun 2, 2010 at 9:05 PM, Alex Harui aha...@adobe.com wrote: I don’t know of any way to block for async network calls. On 6/2/10 6:59 PM, Richard Rodseth rrods...@gmail.com wrote: I know this sounds terrible, but is there a way to wrap an asynchronous call in one which blocks while waiting for it to complete? I'm loading XML via URLLoader to configure a mock API implementation and am limited in what I can change in the main application architecture. So I need a way to block while loading the configuration file. I've tried setting a _loadState variable and looping while it's set to loading, but that didn't seem to do the trick. -- Alex Harui Flex SDK Team Adobe System, Inc. http://blogs.adobe.com/aharui
Re: [flexcoders] AS3/MXML Display list
I'm not an old-school Flash developer, so I'm sure you'll get better-informed replies, but I would just note that if all you care about is back-end connection, HttpService, Binding etc. can be all instantiated in AS3, with no markup required. Also, this in the script block is going to refer to a class which descends from UIComponent which descends from DisplayObject. On Mon, May 31, 2010 at 4:52 PM, Clark Stevenson a.scots...@gmail.comwrote: Hey all. Can someone please help me understand how to work with the display list? Im an AS3 developer and very fond of the freedom i have with AS3. MXML has a great deal of useful components for connection to the backend and so on. These components could be handy for me and thus i need to use the MXML framework. So you start a new MXML Project and under the Script tag you add this code: fx:Script ![CDATA[ private var circle:Sprite; circle = new Sprite(); circle.graphics.beginFill(0x00); circle.graphics.drawCircle(100, 100, 100); circle.graphics.endFill(); addChild(circle); ]] /fx:Script Everything from the begin fill is an error. Can someone explain to me why Sprite() is AS3, why its available to the Script tags but why its not really supported? I just feel that MXML is screwing up my understanding of the player. Main.as is a game and it extends Sprite So if MXML was AS3, id simply add an instance of var main:Main = new Main(); --- this is my entire game. So you think, well Main extends Canvas but the Script tags on Canvas dont understand addChild either Is what i am trying to do just impossible? I know that MXML is a time saver and so on, but to be honest, i hate markup and i just cant find a way to go back what what i know. This is something which has bugged me for years... Id love to understand flex more, but after investing so many years in AS3 development, its just another AS2 AS3 jump to start rewriting every single project in MXML and to be honest, i cant be bothered right now. If i cant use MXML i cant use them and that great, but theres trillions of forum posts on google with a trillion views on this subject and none of them really agree. Cheers
Re: [flexcoders] E4X wildcard/dynamic query
Thanks. I'll explore those techniques, but if that's what it takes, I think
I'd be inclined to just return xml.nodes and iterate in code.
Or perhaps
result = xml.node.(matches(???))
assuming there's a way to refer to the current node in place of ???
I was hoping there was something like
q = { a : 1 b:* }
result = xml.node.(@a == q.a @b == q.b)
I suppose another approach would be to ensure that each node has all the
candidate attributes, with some of them having a value any, but that would
be a pity.
On Sat, May 29, 2010 at 3:08 AM, Oleg Sivokon olegsivo...@gmail.com wrote:
var xml : XML =
nodes
node a=1/
node a=1 b=2/
/nodes;
xml.no...@*.(/a|b/.test(localName()) doSomething(valueOf()));
. . .
private function doSomething(xml:XML):void { trace(xml); }
Like this?
Re: [flexcoders] E4X wildcard/dynamic query
Thanks Oleg. I got something similar working. The requirement is just to
get the most specific result set with a variable number of supplied query
parameters. The following seems to do the trick, but I'm still surprised if
there isn't something more native.
private function evaluate(data:XML, query:Object):XMLList {
trace(Evaluate + query.company + + query.location);
var result:XMLList = data.node.(match(valueOf(), query) );
return result;
}
private function match(node:XML, query:Object):Boolean {
if (query.company no...@company != query.company)
return false;
if (query.location no...@location != query.location)
return false;
return true;
}
var data:XML = nodes
node company=google
google
/node
node company=google location=seattle
google seattle
/node
node company=google location=san jose
google san jose
/node
node company=adobe location=san jose
adobe san jose
/node
node company=adobe location=seattle
adobe seattle
/node
/nodes
var q1:Object = { company : google, location: null };
var q2:Object = { company : google , location: seattle };
var q3:Object = { company : null , location: seattle };
var result1:XMLList = evaluate(data, q1);
trace(result1);
var result2:XMLList = evaluate(data, q2);
trace(result2);
var result3:XMLList = evaluate(data, q3);
trace(result3);
Output:
Evaluate google null
node company=googlegoogle/node
node company=google location=seattlegoogle seattle/node
node company=google location=san josegoogle san jose/node
Evaluate google seattle
google seattle
Evaluate null seattle
node company=google location=seattlegoogle seattle/node
node company=adobe location=seattleadobe seattle/node
On Sat, May 29, 2010 at 9:02 AM, Oleg Sivokon olegsivo...@gmail.com wrote:
Yes, valueOf() in the context of filter operator will return you the
reference to the node currently processed. E4X, although uses special
opcodes, from the AS3 programmer's perspective is the same part of the
language as functions or variables are :) So, when you iterate over XML
using E4X operators you are essentially doing that in code. If you want,
it's not like XPath or regular expressions, think of LinQ as a better
analogue.
So, if the requirement is to check if all fields of an object have
corresponding attributes, I would do it like so:
var xml : XML =
nodes
item foo=12345/
item foo=34567 bar=qwerty/
/nodes;
var pattern:Object = { foo : true, bar : true };
xml.item.(this.check(valueOf(), pattern)).(trace(valueOf().toXMLString()));
. . .
private function check(node:XML, pattern:Object):Boolean
{
for (var p:String in pattern)
{
if (!(@ + p in node)) return false;
}
return true;
}
Best.
Oleg
Re: [flexcoders] E4X wildcard/dynamic query
Not sure where you got the idea that I'm validating. I'm actually loading a node to use for configuration information for a mock api (where the mock api has parameters which may or may not have non-null values). If the configurer chooses to specify a configuration for a specific company and location I want to pick that node when the api user provides a company and location as inputs. The node with no attributes would be the default configuration. I can certainly see that I could require that the nodes all have the same attributes with some specifying a null value, but I think the approach we've arrived at makes for a cleaner configuration file. With a database I can say select where company=google or I can say select where company=google and location = seattle The column values can be nullable so I don't have to specify a value for each column in each row. Anyway, I'm happy with the solution, I mean no disrespect for E4X, and perhaps my usage is atypical. Thanks for the help! On Sat, May 29, 2010 at 11:04 AM, Oleg Sivokon olegsivo...@gmail.comwrote: Well, normally you would know what parameters are required for the node to have. So, you would just write an E4X for that specific set of requirements and if the requirements change you would replace the entire block. Just as you would replace CSS in HTML template. That is I kind of fail to see the point in making validation dynamic and construct it at runtime instead of hardcoding it. What I mean is, since you know that you have to look for @company and @location attributes - why would you provide them dynamically and try to construct a validating expression instead of having that expression hardcoded? As per my experience the hardcoded validation always works better... - less errors, easier to read code. Compare this to constructing DTD or XSL dynamically - usually it kind of defeats the purpose of having those tools :) Well, I think this all comes from the inherent to XML lack of typing and structure. It is usually to generic for the data it serializes. There are different solutions for transferring data preserving typing and stucture, which are also more compact and often times easier to use - AMF and Protobuf would be the most popular probably.
[flexcoders] E4X wildcard/dynamic query
Does attribute matching in E4X include some kind of wildcard value? I'd like to be able to do something like: nodes.node.(attribute( a) == query.a attribute(b) == query.b) nodes node a=1 /node node a=1 b=2 /node nodes and have both nodes match if query.b is null.
Re: [flexcoders] Re: Security sandbox violation
That's how I started I out and will most likely end up. But getting the error while debugging (which has since gone away) and reading the responses here indicated it shouldn't work. On Wed, May 26, 2010 at 6:06 AM, valdhor valdhorli...@embarqmail.comwrote: Why not just keep networking on and access the local file with a local URL? mx:HttpService url=test.xml ... --- In flexcoders@yahoogroups.com flexcoders%40yahoogroups.com, Richard Rodseth rrods...@... wrote: Thanks. That's good to know about. It seems, though, that one has to choose between allowing local access or network access. Can anyone tell me if it's possible to create a single SWF (plus companion wrapper html and config xml that can either be hosted on a web server, or unzipped and run locally? Or would the end user have to use that page you linked to? On Tue, May 25, 2010 at 6:18 AM, valdhor valdhorli...@...wrote: What I do for debugging purposes is to set my debug folder as a trusted source in http://www.macromedia.com/support/documentation/en/flashplayer/help/settings_manager04.html To do this... Click Edit Locations-Add Location Click Browse for Folder Browse to your bin-debug folder This allows you to debug locally - ie. from a file://location_of_bin-debug_HTML_file. If you have a local server running (Which I do) the you can set your run/debug settings to open the html file on your local server. ie. http://127.0.0.1/myproject/bin-debug/myproject.html. In this case, you would need a crossdomain.xml file. --- In flexcoders@yahoogroups.com flexcoders%40yahoogroups.comflexcoders% 40yahoogroups.com, Richard Rodseth rrodseth@ wrote: Not sure I follow that page, but it appears that the -use-network=false compiler flag may be helpful for this scenario. If only FB allowed you to set different compiler options for different applications within a project. On Mon, May 24, 2010 at 4:11 PM, Oleg Sivokon olegsivokon@ wrote: No, it is different. If the SWF is embedded in HTML page, it requires that you specify the trusted locations on your machine, so it could load content: http://macromedia.com/support/documentation/en/flashplayer/help/settings_manager04.html Read carefully the text in gray under the picture. But, if you launch it in standalone player, then, I reckon, it should be able to access local files.
Re: [flexcoders] Re: Security sandbox violation
I'm not sure we're all talking about the same thing, but there must be zillions of SWFs accessing external XML configurations without any user intervention. I haven't tested on a web server yet, but this -use-network=false seems appropriate. On Tue, May 25, 2010 at 8:49 AM, Amy amyblankens...@bellsouth.net wrote: --- In flexcoders@yahoogroups.com flexcoders%40yahoogroups.com, Oleg Sivokon olegsivo...@... wrote: No, it is different. If the SWF is embedded in HTML page, it requires that you specify the trusted locations on your machine, so it could load content: http://macromedia.com/support/documentation/en/flashplayer/help/settings_manager04.html Read carefully the text in gray under the picture. But, if you launch it in standalone player, then, I reckon, it should be able to access local files. No, I am pretty sure you (and all your users) will need to use the Settings Manager page in order to trust the directory with your XML in it. Apparently, you can provide a config file to get around this: http://help.adobe.com/en_US/ActionScript/3.0_ProgrammingAS3/WS5b3ccc516d4fbf351e63e3d118a9b90204-7c95.html, but the directory where this is supposed to live doesn't even exist on my computer, and there is no clue of where to get one to use as a starting point. Good luck! Amy
Re: [flexcoders] Re: Security sandbox violation
I'm obviously thoroughly confused. Presumably -use-network must be true in order to run the folder from the server, because the swf will be local once it's trying to read the XML config file sitting beside the SWF on the server. I'm hoping I don't need a different SWF for the second scenario (unzipping a folder and opening the wrapper html locally. I just tried again, launching the debugger, even without setting -use-network to false, and this time it worked. On Tue, May 25, 2010 at 11:08 AM, Richard Rodseth rrods...@gmail.comwrote: I'm not sure we're all talking about the same thing, but there must be zillions of SWFs accessing external XML configurations without any user intervention. I haven't tested on a web server yet, but this -use-network=false seems appropriate. On Tue, May 25, 2010 at 8:49 AM, Amy amyblankens...@bellsouth.net wrote: --- In flexcoders@yahoogroups.com flexcoders%40yahoogroups.com, Oleg Sivokon olegsivo...@... wrote: No, it is different. If the SWF is embedded in HTML page, it requires that you specify the trusted locations on your machine, so it could load content: http://macromedia.com/support/documentation/en/flashplayer/help/settings_manager04.html Read carefully the text in gray under the picture. But, if you launch it in standalone player, then, I reckon, it should be able to access local files. No, I am pretty sure you (and all your users) will need to use the Settings Manager page in order to trust the directory with your XML in it. Apparently, you can provide a config file to get around this: http://help.adobe.com/en_US/ActionScript/3.0_ProgrammingAS3/WS5b3ccc516d4fbf351e63e3d118a9b90204-7c95.html, but the directory where this is supposed to live doesn't even exist on my computer, and there is no clue of where to get one to use as a starting point. Good luck! Amy
Re: [flexcoders] Re: Security sandbox violation
Thanks. That's good to know about. It seems, though, that one has to choose between allowing local access or network access. Can anyone tell me if it's possible to create a single SWF (plus companion wrapper html and config xml that can either be hosted on a web server, or unzipped and run locally? Or would the end user have to use that page you linked to? On Tue, May 25, 2010 at 6:18 AM, valdhor valdhorli...@embarqmail.comwrote: What I do for debugging purposes is to set my debug folder as a trusted source in http://www.macromedia.com/support/documentation/en/flashplayer/help/settings_manager04.html To do this... Click Edit Locations-Add Location Click Browse for Folder Browse to your bin-debug folder This allows you to debug locally - ie. from a file://location_of_bin-debug_HTML_file. If you have a local server running (Which I do) the you can set your run/debug settings to open the html file on your local server. ie. http://127.0.0.1/myproject/bin-debug/myproject.html. In this case, you would need a crossdomain.xml file. --- In flexcoders@yahoogroups.com flexcoders%40yahoogroups.com, Richard Rodseth rrods...@... wrote: Not sure I follow that page, but it appears that the -use-network=false compiler flag may be helpful for this scenario. If only FB allowed you to set different compiler options for different applications within a project. On Mon, May 24, 2010 at 4:11 PM, Oleg Sivokon olegsivo...@... wrote: No, it is different. If the SWF is embedded in HTML page, it requires that you specify the trusted locations on your machine, so it could load content: http://macromedia.com/support/documentation/en/flashplayer/help/settings_manager04.html Read carefully the text in gray under the picture. But, if you launch it in standalone player, then, I reckon, it should be able to access local files.
Re: [flexcoders] Re: Security sandbox violation
Thanks. That sounds like a no for supporting my two use cases, unless the customer adds a trusted location or opens the SWF in the standalone player. On Tue, May 25, 2010 at 12:45 PM, Oleg Sivokon olegsivo...@gmail.comwrote: OK, it's like this: SWF launched in standalone player can interact with the system AND load remote content, just like the browser can do the same thing. SWF launched in browser plugin can only access permitted folders on your hard disc. To allow access to the folder you should either use the web interface (see my previous link), or read from the link Ami posted and try to figure out how to do that on Mac (you are not required to use web interface to establish local trust relationships between SWF and specific locations in file system, it's just that, I simply don't know how to do that on Mac...) There's yet another option for launching a SWF - that's how I usually do that - from the local HTTP server, in such case it acts more like as if it was deployed to the actual server and all the following rules apply. I would definitely recommend the later way of debugging as it is the most similar to the live situation. The loading of content other than SWF into another SWF is governed by policy files aka crossdomain.xml. You need these files if you are loading content from domains other than SWF origin. The loading of other SWFs by SWF is governed by two things: ApplicationDomain provided to the Loader when loading SWF with LoaderContext - by default it won't allow crosscripting. Security.allowDomain() doesn't define how classes are loaded into application domain, however, should prevent security errors related to crosscripting. Well, yes, it is complicated...
Re: [flexcoders] Re: Security sandbox violation
Sorry :) That seems to happen a lot with my topics. Use case 1: Customer (a sales engineer) accesses a demo that is hosted on the web like any other Flex app. This demo makes no api calls, but does need to access an XML config file that lives in a data folder beside the SWF Use case 2: Sales engineer has a copy of the entire directory (swf, wrapper html, config file) and runs it locally I think your second point confirms that to support both use cases with one version of the application, the customer would have to add a trusted location so that the app could get to the config. That may be fine, I would have to check. Since this is all read-only stuff, I would have hoped I could do it with one SWF and no user-intervention, but I imagine there are good reasons for these restrictions. On Tue, May 25, 2010 at 3:26 PM, Alex Harui aha...@adobe.com wrote: I got lost trying to follow this topic. Whether it is browser or standalone shouldn’t matter. What matters is the url of the SWF (file://or http://), the trust file settings on the computer, the use-network flag, and how you access the external resources. If you build a SWF to run locally then it cannot access the network. If you build a SWF to access the network, it cannot access the local files unless trusted. Even if trusted, you may run into issues using certain file paths. Be sure to use relative paths to subfolders of the trusted folder. On 5/25/10 2:42 PM, Richard Rodseth rrods...@gmail.com wrote: Thanks. That sounds like a no for supporting my two use cases, unless the customer adds a trusted location or opens the SWF in the standalone player. On Tue, May 25, 2010 at 12:45 PM, Oleg Sivokon olegsivo...@gmail.com wrote: OK, it's like this: SWF launched in standalone player can interact with the system AND load remote content, just like the browser can do the same thing. SWF launched in browser plugin can only access permitted folders on your hard disc. To allow access to the folder you should either use the web interface (see my previous link), or read from the link Ami posted and try to figure out how to do that on Mac (you are not required to use web interface to establish local trust relationships between SWF and specific locations in file system, it's just that, I simply don't know how to do that on Mac...) There's yet another option for launching a SWF - that's how I usually do that - from the local HTTP server, in such case it acts more like as if it was deployed to the actual server and all the following rules apply. I would definitely recommend the later way of debugging as it is the most similar to the live situation. The loading of content other than SWF into another SWF is governed by policy files aka crossdomain.xml. You need these files if you are loading content from domains other than SWF origin. The loading of other SWFs by SWF is governed by two things: ApplicationDomain provided to the Loader when loading SWF with LoaderContext - by default it won't allow crosscripting. Security.allowDomain() doesn't define how classes are loaded into application domain, however, should prevent security errors related to crosscripting. Well, yes, it is complicated... -- Alex Harui Flex SDK Team Adobe System, Inc. http://blogs.adobe.com/aharui
[flexcoders] Security sandbox violation
I'm creating a demo swf that won't attach to a server and may be hosted, or launched from the desktop. I wish to externalize some data in a separate xml or text file that will live beside the SWF rather than being embedded in it (actually I may need both). What's the best approach here? In using HTTPService to read the data file (which I placed in html-template so it got copied to bin-debug), I got a security sandbox error running under the Flexbuilder debugger.
Re: [flexcoders] Security sandbox violation
Yes, that's what I'm seeing with both URLLoader and HTTPService. So I guess my question becomes: How do you debug this stuff in FB ? On Mon, May 24, 2010 at 12:18 PM, Oleg Sivokon olegsivo...@gmail.comwrote: Please correct me if I'm wrong, but I think that if you launch it in standalone player, you shouldn't get the security error, well, maybe for system or read-only files, or, maybe due to other system restrictions, but generally it should just work...
Re: [flexcoders] Security sandbox violation
Potentially both. The client wishes to host it on a web server (with no app server/db access) but it could also get run from the desktop, unless we provide the wrapper as part of our solution. The browser version is required. On Mon, May 24, 2010 at 12:48 PM, Oleg Sivokon olegsivo...@gmail.comwrote: Sorry, are you running it in standalone player or in browser? There shouldn't be a problem if you run it in standalone player.
Re: [flexcoders] Security sandbox violation
And both seem to work when I'm launching from the desktop, but not from the debugger. On Mon, May 24, 2010 at 1:19 PM, Richard Rodseth rrods...@gmail.com wrote: Potentially both. The client wishes to host it on a web server (with no app server/db access) but it could also get run from the desktop, unless we provide the wrapper as part of our solution. The browser version is required. On Mon, May 24, 2010 at 12:48 PM, Oleg Sivokon olegsivo...@gmail.comwrote: Sorry, are you running it in standalone player or in browser? There shouldn't be a problem if you run it in standalone player.
Re: [flexcoders] Security sandbox violation
It's possible I have multiple players on my system, and FB is using a different one. By launch, I just meant using the Run/Launch button within FB as opposed to finding the wrapper html in the Finder and opening it. On Mon, May 24, 2010 at 2:04 PM, Oleg Sivokon olegsivo...@gmail.com wrote: Not sure I get it, how do you launch that from debugger? There are several options available in FB (and few more in SDK and in alternative debuggers), you can launch in browser, and if you have debug player installed there, it will start the debugger client in FB, or, you can launch in standalone debugger player, it's hard to tell which player exactly FB will use, but most often it will pick the one from your SDK, and that will be a debug player, so that again it will connect back to the debugger client in FB. One more option to debug is by launching fdb commandline debugger... On the other hand, if you don't launch that in debug player, how do you know there aren't errors?
Re: [flexcoders] Security sandbox violation
Standard behavior in FB is to launch in browser (the html-wrapper folder gets copied to bin-debug with template items filled in. I'm guessing you aren't an FB user - I'm not away of any other way to launch from within FB. On Mon, May 24, 2010 at 3:14 PM, Oleg Sivokon olegsivo...@gmail.com wrote: When you click launch it may either launch in browser or in standalone player. Which one does it launch?
Re: [flexcoders] Security sandbox violation
Uh oh. So there appears to be a Safari/Firefox difference. I was launching Safari when opening the html file in the Finder, and Firefox in Eclipse. Whichever way I launch (fb or Finder), I get the security error in Firefox, and not in Safari. Sorry for the confusion. So now I'm back to wondering if my approach is incorrect, whether I need a crossdomain.xml etc etc. To recap: I just need a folder with a swf, wrapper html and external data file. Might be hosted on the web, or e-mailed to someone. I'll read up more about security, but if anyone has tips about this scenario, I'd appreciate them. On Mon, May 24, 2010 at 3:37 PM, Oleg Sivokon olegsivo...@gmail.com wrote: Well, I am, but I usually compile with Ant and launch the player from the build script. So, I really didn't remember which are the default settings. Well, that's not really related. In your case you'd need to configure the builder to launch whatever it compiles in the run configuration follow these steps: http://livedocs.adobe.com/flex/3/html/help.html?content=compile_debug_4.html#149153 they shouldn't be very different if you have FB4. Best. Oleg
Re: [flexcoders] Security sandbox violation
Not sure I follow that page, but it appears that the -use-network=false compiler flag may be helpful for this scenario. If only FB allowed you to set different compiler options for different applications within a project. On Mon, May 24, 2010 at 4:11 PM, Oleg Sivokon olegsivo...@gmail.com wrote: No, it is different. If the SWF is embedded in HTML page, it requires that you specify the trusted locations on your machine, so it could load content: http://macromedia.com/support/documentation/en/flashplayer/help/settings_manager04.html Read carefully the text in gray under the picture. But, if you launch it in standalone player, then, I reckon, it should be able to access local files.
[flexcoders] TabNavigator
Is there an event dispatched by TabNavigator when the *user* switches tabs? change is dispatched when the selected change happens programmatically (eg. because selectedChild is bound).
Re: [flexcoders] Re: TabNavigator
You might mean click, in which case yes. It is dispatched when clicking with the content area of the navigator, but not on the button bar. And it wouldn't cover keyboard navigation. On Fri, May 21, 2010 at 1:06 PM, turbo_vb timh...@aol.com wrote: You try itemClick? -TH --- In flexcoders@yahoogroups.com flexcoders%40yahoogroups.com, Richard Rodseth rrods...@... wrote: Is there an event dispatched by TabNavigator when the *user* switches tabs? change is dispatched when the selected change happens programmatically (eg. because selectedChild is bound).
Re: [flexcoders] Script timeout binding selection in Tree
I think I already mentioned it was called twice. I can double-check
tomorrow.
On Mon, May 17, 2010 at 9:35 PM, Alex Harui aha...@adobe.com wrote:
Set a breakpoint on setSelectionDataLoop. See how often it gets called.
On 5/17/10 4:49 PM, Richard Rodseth rrods...@gmail.com wrote:
Thanks for trying. No, the presentation model has an array collection whose
.source gets changed after an HTTPService call returns (although I should
mention that I am using my mock service layer at the moment). I added a a
collection change listener to the collection reference.
I have set a breakpoint in setSelectionDataLoop(). I hit it twice (the
parameters appear the same - index 0, items array containing the object to
select, useFind=false). Then the debugger stopped with the finally{} block
in Binding.watcherFired highlighted and the stack crawl below. Below this
stack crawl I've included the last few lines of the one that shows up in the
console.
Main Thread (Suspended: Error: Error #1502: A script has executed for
longer than the default timeout period of 15 seconds.)
mx.binding::Binding/watcherFired
mx.binding::Watcher/notifyListeners
mx.binding::PropertyWatcher/eventHandler
flash.events::EventDispatcher/dispatchEventFunction [no source]
flash.events::EventDispatcher/dispatchEvent [no source]
com.companyname.admin.deviceshome.ui.presentation::DevicesHome/changeSelectedGroupIds
com.companyname.admin.deviceshome.ui.presentation::DevicesHome/groupsChangedListener
flash.events::EventDispatcher/dispatchEventFunction [no source]
flash.events::EventDispatcher/dispatchEvent [no source]
mx.collections::ListCollectionView/dispatchEvent
mx.collections::ListCollectionView/
http://www.adobe.com/2006/flex/mx/internal::reset
mx.collections::ListCollectionView/set list
mx.collections::ArrayCollection/set source
anonymous
mx.rpc::Responder/result
anonymous
Array$/_forEach [no source]
Array/http://adobe.com/AS3/2006/builtin::forEach [no source]
anonymous
flash.utils::Timer/_timerDispatch [no source]
flash.utils::Timer/tick [no source]
Console:
Error: Error #1502: A script has executed for longer than the default
timeout period of 15 seconds.
at mx.controls.treeClasses::HierarchicalViewCursor/moveNext()
at
mx.controls.listClasses::ListBase/setSelectionDataLoop()[C:\autobuild\3.5.0\frameworks\projects\framework\src\mx\controls\listClasses\ListBase.as:6455]
On Mon, May 17, 2010 at 4:09 PM, Oleg Sivokon olegsivo...@gmail.com
wrote:
I wish I could help more... well here's another guess:
- if you are using XML / XMLList, and have used insertBefore / isertAfter
from the same XML / XMLList, then it is a known bug.
- if you dispatch collectionChange on your own you might have messed
something in the event's properties, which then will make the collection run
in a loop thinking it's being updated over and over again. It is slow, but
it isn't that slow to run out of script timeout limit :)
Anyways, bindings are the first suspect in such case, but if that's not...
ah, and, why won't you debug it? the debugger should be able to look into
the framework sources. If, for whatever reason it doesn't, copy the Tree and
other classes involved into your project and try debugging this way.
Best.
Oleg
--
Alex Harui
Flex SDK Team
Adobe System, Inc.
http://blogs.adobe.com/aharui
Re: [flexcoders] Script timeout binding selection in Tree
Two or three times. Then I get the spinning beach ball.
On Mon, May 17, 2010 at 11:12 PM, Richard Rodseth rrods...@gmail.comwrote:
I think I already mentioned it was called twice. I can double-check
tomorrow.
On Mon, May 17, 2010 at 9:35 PM, Alex Harui aha...@adobe.com wrote:
Set a breakpoint on setSelectionDataLoop. See how often it gets called.
On 5/17/10 4:49 PM, Richard Rodseth rrods...@gmail.com wrote:
Thanks for trying. No, the presentation model has an array collection
whose .source gets changed after an HTTPService call returns (although I
should mention that I am using my mock service layer at the moment). I added
a a collection change listener to the collection reference.
I have set a breakpoint in setSelectionDataLoop(). I hit it twice (the
parameters appear the same - index 0, items array containing the object to
select, useFind=false). Then the debugger stopped with the finally{} block
in Binding.watcherFired highlighted and the stack crawl below. Below this
stack crawl I've included the last few lines of the one that shows up in the
console.
Main Thread (Suspended: Error: Error #1502: A script has executed for
longer than the default timeout period of 15 seconds.)
mx.binding::Binding/watcherFired
mx.binding::Watcher/notifyListeners
mx.binding::PropertyWatcher/eventHandler
flash.events::EventDispatcher/dispatchEventFunction [no source]
flash.events::EventDispatcher/dispatchEvent [no source]
com.companyname.admin.deviceshome.ui.presentation::DevicesHome/changeSelectedGroupIds
com.companyname.admin.deviceshome.ui.presentation::DevicesHome/groupsChangedListener
flash.events::EventDispatcher/dispatchEventFunction [no source]
flash.events::EventDispatcher/dispatchEvent [no source]
mx.collections::ListCollectionView/dispatchEvent
mx.collections::ListCollectionView/
http://www.adobe.com/2006/flex/mx/internal::reset
mx.collections::ListCollectionView/set list
mx.collections::ArrayCollection/set source
anonymous
mx.rpc::Responder/result
anonymous
Array$/_forEach [no source]
Array/http://adobe.com/AS3/2006/builtin::forEach [no source]
anonymous
flash.utils::Timer/_timerDispatch [no source]
flash.utils::Timer/tick [no source]
Console:
Error: Error #1502: A script has executed for longer than the default
timeout period of 15 seconds.
at mx.controls.treeClasses::HierarchicalViewCursor/moveNext()
at
mx.controls.listClasses::ListBase/setSelectionDataLoop()[C:\autobuild\3.5.0\frameworks\projects\framework\src\mx\controls\listClasses\ListBase.as:6455]
On Mon, May 17, 2010 at 4:09 PM, Oleg Sivokon olegsivo...@gmail.com
wrote:
I wish I could help more... well here's another guess:
- if you are using XML / XMLList, and have used insertBefore / isertAfter
from the same XML / XMLList, then it is a known bug.
- if you dispatch collectionChange on your own you might have messed
something in the event's properties, which then will make the collection run
in a loop thinking it's being updated over and over again. It is slow, but
it isn't that slow to run out of script timeout limit :)
Anyways, bindings are the first suspect in such case, but if that's not...
ah, and, why won't you debug it? the debugger should be able to look into
the framework sources. If, for whatever reason it doesn't, copy the Tree and
other classes involved into your project and try debugging this way.
Best.
Oleg
--
Alex Harui
Flex SDK Team
Adobe System, Inc.
http://blogs.adobe.com/aharui
Re: [flexcoders] Script timeout binding selection in Tree
That sounds about right. moveNext() does appear in the stack sometimes, and
I am changing the data provider contents.
I may have to put the restoreselection code somewhere other than in a
collection change handler.
On Tue, May 18, 2010 at 9:16 AM, Alex Harui aha...@adobe.com wrote:
If that’s true, on the second invocation, step through it. It might be
hung up in moveNext() not advancing to the next item, maybe because you’ve
run out of items.
On 5/17/10 11:12 PM, Richard Rodseth rrods...@gmail.com wrote:
I think I already mentioned it was called twice. I can double-check
tomorrow.
On Mon, May 17, 2010 at 9:35 PM, Alex Harui aha...@adobe.com wrote:
Set a breakpoint on setSelectionDataLoop. See how often it gets called.
On 5/17/10 4:49 PM, Richard Rodseth rrods...@gmail.com
http://rrods...@gmail.com wrote:
Thanks for trying. No, the presentation model has an array collection whose
.source gets changed after an HTTPService call returns (although I should
mention that I am using my mock service layer at the moment). I added a a
collection change listener to the collection reference.
I have set a breakpoint in setSelectionDataLoop(). I hit it twice (the
parameters appear the same - index 0, items array containing the object to
select, useFind=false). Then the debugger stopped with the finally{} block
in Binding.watcherFired highlighted and the stack crawl below. Below this
stack crawl I've included the last few lines of the one that shows up in the
console.
Main Thread (Suspended: Error: Error #1502: A script has executed for
longer than the default timeout period of 15 seconds.)
mx.binding::Binding/watcherFired
mx.binding::Watcher/notifyListeners
mx.binding::PropertyWatcher/eventHandler
flash.events::EventDispatcher/dispatchEventFunction [no source]
flash.events::EventDispatcher/dispatchEvent [no source]
com.companyname.admin.deviceshome.ui.presentation::DevicesHome/changeSelectedGroupIds
com.companyname.admin.deviceshome.ui.presentation::DevicesHome/groupsChangedListener
flash.events::EventDispatcher/dispatchEventFunction [no source]
flash.events::EventDispatcher/dispatchEvent [no source]
mx.collections::ListCollectionView/dispatchEvent
mx.collections::ListCollectionView/
http://www.adobe.com/2006/flex/mx/internal::reset
mx.collections::ListCollectionView/set list
mx.collections::ArrayCollection/set source
anonymous
mx.rpc::Responder/result
anonymous
Array$/_forEach [no source]
Array/http://adobe.com/AS3/2006/builtin::forEach [no source]
anonymous
flash.utils::Timer/_timerDispatch [no source]
flash.utils::Timer/tick [no source]
Console:
Error: Error #1502: A script has executed for longer than the default
timeout period of 15 seconds.
at mx.controls.treeClasses::HierarchicalViewCursor/moveNext()
at
mx.controls.listClasses::ListBase/setSelectionDataLoop()[C:\autobuild\3.5.0\frameworks\projects\framework\src\mx\controls\listClasses\ListBase.as:6455]
On Mon, May 17, 2010 at 4:09 PM, Oleg Sivokon olegsivo...@gmail.com
http://olegsivo...@gmail.com wrote:
I wish I could help more... well here's another guess:
- if you are using XML / XMLList, and have used insertBefore / isertAfter
from the same XML / XMLList, then it is a known bug.
- if you dispatch collectionChange on your own you might have messed
something in the event's properties, which then will make the collection run
in a loop thinking it's being updated over and over again. It is slow, but
it isn't that slow to run out of script timeout limit :)
Anyways, bindings are the first suspect in such case, but if that's not...
ah, and, why won't you debug it? the debugger should be able to look into
the framework sources. If, for whatever reason it doesn't, copy the Tree and
other classes involved into your project and try debugging this way.
Best.
Oleg
--
Alex Harui
Flex SDK Team
Adobe System, Inc.
http://blogs.adobe.com/aharui
Re: [flexcoders] Script timeout binding selection in Tree
I've attempted to create a one-file test case, and while it doesn't exhibit
the script timeout, the selection restoration does not work as part of the
collection listener.
Is it significant that selectedItems does not show up in code completion
while selectedIndices does?
?xml version=1.0 encoding=utf-8?
mx:Application xmlns:mx=http://www.adobe.com/2006/mxml;
layout=vertical minWidth=955 minHeight=600
creationComplete=application1_creationCompleteHandler(event)
mx:Script
![CDATA[
import mx.collections.ArrayCollection;
import mx.events.CollectionEvent;
import mx.events.CollectionEventKind;
import mx.events.FlexEvent;
import mx.events.ListEvent;
[Bindable]
public var items:ArrayCollection = new ArrayCollection();
[Bindable(event=selectionChanged)]
public function get selectedItems():Array {
return _selectedItems;
}
private var _selectedItems:Array = [];
private var _selectedIds:Array = [];
private function selectItems(items:Array):void {
_selectedItems = items;
_selectedIds = this.toIds(items);
dispatchEvent(new Event(selectionChanged));
}
protected function
application1_creationCompleteHandler(event:FlexEvent):void
{
items.addEventListener(CollectionEvent.COLLECTION_CHANGE,
itemsListener, false, 0, true);
}
private function itemsListener(event:CollectionEvent):void {
if (event.kind == CollectionEventKind.RESET) {
this.restoreSelection();
}
}
private function refresh():void {
this.items.source = [ { uid:1, name:Root, children: [ {
uid:2, name:Child } ] } ];
}
private function toIds(items:Array):Array {
var result:Array = new Array();
for each (var item:Object in items) {
result.push(item.uid);
}
return result;
}
private function toItems(ids:Array):Array {
var root:Object = this.items ? this.items.getItemAt(0) :
null;
var result:Array = new Array();
for each (var id:String in ids) {
var item:Object = find(root, id);
if (item) {
result.push(item);
}
}
return result;
}
private function restoreSelection():void {
_selectedItems = toItems(_selectedIds);
dispatchEvent(new Event(selectionChanged));
}
private function find(start:Object, uid:String):Object {
if (start.uid == uid)
return start;
for each (var child:Object in start.children) {
var found:Object = this.find(child, uid);
if (found)
return found;
}
return null;
}
protected function tree1_changeHandler(event:ListEvent):void
{
this.selectItems(theTree.selectedItems);
}
protected function button1_clickHandler(event:MouseEvent):void
{
this.refresh();
}
]]
/mx:Script
mx:Tree id=theTree
dataProvider={this.items}
selectedItems={this.selectedItems}
labelField=name
allowMultipleSelection=true
change=tree1_changeHandler(event)/
mx:Button label=Refresh click=refresh()/
mx:Button label=Restore Selection click=restoreSelection()/
/mx:Application
On Tue, May 18, 2010 at 9:23 AM, Richard Rodseth rrods...@gmail.com wrote:
That sounds about right. moveNext() does appear in the stack sometimes, and
I am changing the data provider contents.
I may have to put the restoreselection code somewhere other than in a
collection change handler.
On Tue, May 18, 2010 at 9:16 AM, Alex Harui aha...@adobe.com wrote:
If that’s true, on the second invocation, step through it. It might be
hung up in moveNext() not advancing to the next item, maybe because you’ve
run out of items.
On 5/17/10 11:12 PM, Richard Rodseth rrods...@gmail.com wrote:
I think I already mentioned it was called twice. I can double-check
tomorrow.
On Mon, May 17, 2010 at 9:35 PM, Alex Harui aha...@adobe.com wrote:
Set a breakpoint on setSelectionDataLoop. See how often it gets called.
On 5/17/10 4:49 PM, Richard Rodseth rrods...@gmail.com
http://rrods...@gmail.com wrote:
Thanks for trying. No, the presentation model has an array collection
whose .source gets changed after an HTTPService call
Re: [flexcoders] Script timeout binding selection in Tree
My bad (at least in this example). The selectedItems getter is incorrect.
On Tue, May 18, 2010 at 12:15 PM, Richard Rodseth rrods...@gmail.comwrote:
I've attempted to create a one-file test case, and while it doesn't exhibit
the script timeout, the selection restoration does not work as part of the
collection listener.
Is it significant that selectedItems does not show up in code completion
while selectedIndices does?
?xml version=1.0 encoding=utf-8?
mx:Application xmlns:mx=http://www.adobe.com/2006/mxml;
layout=vertical minWidth=955 minHeight=600
creationComplete=application1_creationCompleteHandler(event)
mx:Script
![CDATA[
import mx.collections.ArrayCollection;
import mx.events.CollectionEvent;
import mx.events.CollectionEventKind;
import mx.events.FlexEvent;
import mx.events.ListEvent;
[Bindable]
public var items:ArrayCollection = new ArrayCollection();
[Bindable(event=selectionChanged)]
public function get selectedItems():Array {
return _selectedItems;
}
private var _selectedItems:Array = [];
private var _selectedIds:Array = [];
private function selectItems(items:Array):void {
_selectedItems = items;
_selectedIds = this.toIds(items);
dispatchEvent(new Event(selectionChanged));
}
protected function
application1_creationCompleteHandler(event:FlexEvent):void
{
items.addEventListener(CollectionEvent.COLLECTION_CHANGE,
itemsListener, false, 0, true);
}
private function itemsListener(event:CollectionEvent):void {
if (event.kind == CollectionEventKind.RESET) {
this.restoreSelection();
}
}
private function refresh():void {
this.items.source = [ { uid:1, name:Root, children: [ {
uid:2, name:Child } ] } ];
}
private function toIds(items:Array):Array {
var result:Array = new Array();
for each (var item:Object in items) {
result.push(item.uid);
}
return result;
}
private function toItems(ids:Array):Array {
var root:Object = this.items ? this.items.getItemAt(0) :
null;
var result:Array = new Array();
for each (var id:String in ids) {
var item:Object = find(root, id);
if (item) {
result.push(item);
}
}
return result;
}
private function restoreSelection():void {
_selectedItems = toItems(_selectedIds);
dispatchEvent(new Event(selectionChanged));
}
private function find(start:Object, uid:String):Object {
if (start.uid == uid)
return start;
for each (var child:Object in start.children) {
var found:Object = this.find(child, uid);
if (found)
return found;
}
return null;
}
protected function tree1_changeHandler(event:ListEvent):void
{
this.selectItems(theTree.selectedItems);
}
protected function button1_clickHandler(event:MouseEvent):void
{
this.refresh();
}
]]
/mx:Script
mx:Tree id=theTree
dataProvider={this.items}
selectedItems={this.selectedItems}
labelField=name
allowMultipleSelection=true
change=tree1_changeHandler(event)/
mx:Button label=Refresh click=refresh()/
mx:Button label=Restore Selection click=restoreSelection()/
/mx:Application
On Tue, May 18, 2010 at 9:23 AM, Richard Rodseth rrods...@gmail.comwrote:
That sounds about right. moveNext() does appear in the stack sometimes,
and I am changing the data provider contents.
I may have to put the restoreselection code somewhere other than in a
collection change handler.
On Tue, May 18, 2010 at 9:16 AM, Alex Harui aha...@adobe.com wrote:
If that’s true, on the second invocation, step through it. It might be
hung up in moveNext() not advancing to the next item, maybe because you’ve
run out of items.
On 5/17/10 11:12 PM, Richard Rodseth rrods...@gmail.com wrote:
I think I already mentioned it was called twice. I can double-check
tomorrow.
On Mon, May 17, 2010 at 9:35 PM, Alex Harui aha...@adobe.com wrote:
Set a breakpoint on setSelectionDataLoop. See how
Re: [flexcoders] Script timeout binding selection in Tree
Here's a corrected test case, where only the persistent uids are stored, and
selectedItems is computed. The Restore Selection button works, the
collection event listener doesn't. It would be great if someone could give
this a whirl and confirm what I see. Not sure how this relates to the script
timeout, but I really need a reliable way to move the selection into the
model and bind to it.
?xml version=1.0 encoding=utf-8?
mx:Application xmlns:mx=http://www.adobe.com/2006/mxml;
layout=vertical minWidth=955 minHeight=600
creationComplete=application1_creationCompleteHandler(event)
mx:Script
![CDATA[
import mx.collections.ArrayCollection;
import mx.events.CollectionEvent;
import mx.events.CollectionEventKind;
import mx.events.FlexEvent;
import mx.events.ListEvent;
[Bindable]
public var items:ArrayCollection = new ArrayCollection();
[Bindable(event=selectionChanged)]
public function get selectedItems():Array {
return toItems(_selectedIds);
}
private var _selectedIds:Array = [];
private function selectItems(items:Array):void {
_selectedIds = this.toIds(items);
dispatchEvent(new Event(selectionChanged));
}
protected function
application1_creationCompleteHandler(event:FlexEvent):void
{
items.addEventListener(CollectionEvent.COLLECTION_CHANGE,
itemsListener, false, 0, true);
}
private function itemsListener(event:CollectionEvent):void {
if (event.kind == CollectionEventKind.RESET) {
this.restoreSelection();
}
}
private function refresh():void {
this.items.source = [ { uid:1, name:Root, children: [ {
uid:2, name:Child } ] } ];
}
private function toIds(items:Array):Array {
var result:Array = new Array();
for each (var item:Object in items) {
result.push(item.uid);
}
return result;
}
private function toItems(ids:Array):Array {
var root:Object = this.items ? this.items.getItemAt(0) :
null;
var result:Array = new Array();
for each (var id:String in ids) {
var item:Object = find(root, id);
if (item) {
result.push(item);
}
}
return result;
}
private function toIndices(ids:Array):Array {
var root:Object = this.items ? this.items.getItemAt(0) :
null;
var result:Array = new Array();
for each (var id:String in ids) {
var item:Object = find(root, id);
if (item) {
var index:int = this.items.getItemIndex(item);
result.push(index);
}
}
return result;
}
private function restoreSelection():void {
//_selectedItems = toItems(_selectedIds);
dispatchEvent(new Event(selectionChanged));
}
private function find(start:Object, uid:String):Object {
if (start.uid == uid)
return start;
for each (var child:Object in start.children) {
var found:Object = this.find(child, uid);
if (found)
return found;
}
return null;
}
protected function tree1_changeHandler(event:ListEvent):void
{
this.selectItems(theTree.selectedItems);
}
protected function button1_clickHandler(event:MouseEvent):void
{
this.refresh();
}
]]
/mx:Script
mx:Tree id=theTree
dataProvider={this.items}
selectedItems={this.selectedItems}
labelField=name
allowMultipleSelection=true
change=tree1_changeHandler(event)/
mx:Button label=Refresh click=refresh()/
mx:Button label=Restore Selection click=restoreSelection()/
/mx:Application
On Tue, May 18, 2010 at 12:24 PM, Richard Rodseth rrods...@gmail.comwrote:
My bad (at least in this example). The selectedItems getter is incorrect.
On Tue, May 18, 2010 at 12:15 PM, Richard Rodseth rrods...@gmail.comwrote:
I've attempted to create a one-file test case, and while it doesn't
exhibit the script timeout, the selection restoration does not work as part
of the collection listener.
Is it significant that selectedItems does not show up in code completion
while selectedIndices does
Re: [flexcoders] Script timeout binding selection in Tree
Thanks for testing. At this point, I consider this a bug, that the tree does not display its selection, when it is updated via binding of selectedItems, where the property bound to is updated in response to a RESET of the data provider collection. My workaround might interest some people. It is inspired by the pop-ups-with-presentation-manager pattern described by Theo Iconara Hultberg here: http://mate.asfusion.com/forums/topic.php?id=591 So, using BindingUtils.bindSetter, I have a function in the view which observes a status property in the presentation model. In the past, I have also used the Observe tag from Adobe consulting for this sort of thing. The observing function does the selection updating (though tree.selectedItems is still bound as well). The status property in the presentation model is updated when the collection is reset. In order to get this to work, I *still* had to wrap the view-based selection updating in a callLater(), which I hate, but such is life. On Tue, May 18, 2010 at 6:20 PM, Oleg Sivokon olegsivo...@gmail.com wrote: If I rewrite it to SDK 4 (it's basically the same tree anyway), the collection event listener is called... well, I cannot see anything extraordinary in what it does to be honest.
[flexcoders] Script timeout binding selection in Tree
I'm getting the following error: Error: Error #1502: A script has executed for longer than the default timeout period of 15 seconds. at ?() at mx.controls.treeClasses::HierarchicalViewCursor/moveNext()[C:\autobuild\3.5.0\frameworks\projects\framework\src\mx\controls\treeClasses\HierarchicalViewCursor.as:449] at mx.controls.listClasses::ListBase/setSelectionDataLoop()[C:\autobuild\3.5.0\frameworks\projects\framework\src\mx\controls\listClasses\ListBase.as:6455] at mx.controls.listClasses::ListBase/commitSelectedItems()[C:\autobuild\3.5.0\frameworks\projects\framework\src\mx\controls\listClasses\ListBase.as:6360] at mx.controls.listClasses::ListBase/set selectedItems()[C:\autobuild\3.5.0\frameworks\projects\framework\src\mx\controls\listClasses\ListBase.as:2938] I'm binding the selectedItems property of the tree to a model property. I'm also trying to preserve the selection when the contents of the data provider changes, so I have a COLLECTION_CHANGE listener which builds a new selection array. The combination of these two is somehow causing this error. Any ideas what the cause might be, or how best to debug?
Re: [flexcoders] Script timeout binding selection in Tree
Thanks. Without the binding the timeout does not occur, since that's what causes the rather complex method ListBase.setSelectionDataLoop to be called. I suspect that it might have to do with looking for a selected object that is not part of the data provider collection, but that's only a temporary situation, and adding a selectedItemsCompareFunction did not help. Still digging... On Mon, May 17, 2010 at 1:04 AM, Oleg Sivokon olegsivo...@gmail.com wrote: First of I'd remove all the bindings, because binding may be silencing errors ans than may be running in loop not knowing what to update first. I also think that you only copied the top of the error stack trace, could you tell if there was something looking like recursion, i.e. you would see repetitive blocks of stack frames. Best. Oleg
Re: [flexcoders] Script timeout binding selection in Tree
I'm still stumped by this. Wish I had more information to add. The use case is that the user is navigating (via a button and a bookmarking mechanism I built) to a different screen, resulting in a) a change of active tab b) a refresh of the tree in that tab via asynchronous API call c) selection of a specific item in that tree Also, when on that screen, just clicking on different tabs should refresh the data but not lose the selection. So yes, I am more-or-less-simultaneously trying to update the contents of the list and ensure the selection is not lost. The TabNavigator's selected child, the tree's data provider and selectedItems are all bound to presentation model properties. The data provider has a CollectionChange listener to try to restore the selectedItems property when the tree contents are changed. Selection restoration works fine if I don't do the data refresh. Navigation state change works fine if I don't do the selection restoration. The two together, not so much. On Mon, May 17, 2010 at 11:09 AM, Richard Rodseth rrods...@gmail.comwrote: Thanks. Without the binding the timeout does not occur, since that's what causes the rather complex method ListBase.setSelectionDataLoop to be called. I suspect that it might have to do with looking for a selected object that is not part of the data provider collection, but that's only a temporary situation, and adding a selectedItemsCompareFunction did not help. Still digging... On Mon, May 17, 2010 at 1:04 AM, Oleg Sivokon olegsivo...@gmail.comwrote: First of I'd remove all the bindings, because binding may be silencing errors ans than may be running in loop not knowing what to update first. I also think that you only copied the top of the error stack trace, could you tell if there was something looking like recursion, i.e. you would see repetitive blocks of stack frames. Best. Oleg
Re: [flexcoders] Script timeout binding selection in Tree
Thanks for trying. No, the presentation model has an array collection whose
.source gets changed after an HTTPService call returns (although I should
mention that I am using my mock service layer at the moment). I added a a
collection change listener to the collection reference.
I have set a breakpoint in setSelectionDataLoop(). I hit it twice (the
parameters appear the same - index 0, items array containing the object to
select, useFind=false). Then the debugger stopped with the finally{} block
in Binding.watcherFired highlighted and the stack crawl below. Below this
stack crawl I've included the last few lines of the one that shows up in the
console.
Main Thread (Suspended: Error: Error #1502: A script has executed for longer
than the default timeout period of 15 seconds.)
mx.binding::Binding/watcherFired
mx.binding::Watcher/notifyListeners
mx.binding::PropertyWatcher/eventHandler
flash.events::EventDispatcher/dispatchEventFunction [no source]
flash.events::EventDispatcher/dispatchEvent [no source]
com.companyname.admin.deviceshome.ui.presentation::DevicesHome/changeSelectedGroupIds
com.companyname.admin.deviceshome.ui.presentation::DevicesHome/groupsChangedListener
flash.events::EventDispatcher/dispatchEventFunction [no source]
flash.events::EventDispatcher/dispatchEvent [no source]
mx.collections::ListCollectionView/dispatchEvent
mx.collections::ListCollectionView/
http://www.adobe.com/2006/flex/mx/internal::reset
mx.collections::ListCollectionView/set list
mx.collections::ArrayCollection/set source
anonymous
mx.rpc::Responder/result
anonymous
Array$/_forEach [no source]
Array/http://adobe.com/AS3/2006/builtin::forEach [no source]
anonymous
flash.utils::Timer/_timerDispatch [no source]
flash.utils::Timer/tick [no source]
Console:
Error: Error #1502: A script has executed for longer than the default
timeout period of 15 seconds.
at mx.controls.treeClasses::HierarchicalViewCursor/moveNext()
at
mx.controls.listClasses::ListBase/setSelectionDataLoop()[C:\autobuild\3.5.0\frameworks\projects\framework\src\mx\controls\listClasses\ListBase.as:6455]
On Mon, May 17, 2010 at 4:09 PM, Oleg Sivokon olegsivo...@gmail.com wrote:
I wish I could help more... well here's another guess:
- if you are using XML / XMLList, and have used insertBefore / isertAfter
from the same XML / XMLList, then it is a known bug.
- if you dispatch collectionChange on your own you might have messed
something in the event's properties, which then will make the collection run
in a loop thinking it's being updated over and over again. It is slow, but
it isn't that slow to run out of script timeout limit :)
Anyways, bindings are the first suspect in such case, but if that's not...
ah, and, why won't you debug it? the debugger should be able to look into
the framework sources. If, for whatever reason it doesn't, copy the Tree and
other classes involved into your project and try debugging this way.
Best.
Oleg
[flexcoders] Re: Folder-style dragOver highlight on tree
I'm now aware of dropIndicatorSkin and found an example that customized the skin to do this: g.beginFill(0xa1bde2, 0.5); g.drawRect(-5, -1, w, 23); which is obviously a bit, er, hard-coded. The location and height for the skin appear to be set for drawing a line between nodes. Recall that I just want to highlight the node itself, as though it were being selected. Tree.showDropFeedback and ListBase.showDropFeedback are quite large methods (with side-effects) to be overriding. Is that my best option, or can I have my skin access the tree node? Recommendations or examples appreciated.I feel sure someone must have used a tree to represent a folder hierarchy. Thanks. On Tue, May 11, 2010 at 3:40 PM, Richard Rodseth rrods...@gmail.com wrote: I'm maintaining some code which has a dragEnter handler which sets selectedIndex in order to highlight the destination tree node Can anyone point me at some sample code to do this, or a public method to highlight the item without selecting it? The showDropFeedback method draws a line between items, which is not the desired effect. Thanks.
Re: [flexcoders] Re: Folder-style dragOver highlight on tree
Could you elaborate a bit, please? The component is already used with a custom renderer. Are you suggesting that the dragEnter method should just do the highlighting directly? Makes sense, I suppose. On Wed, May 12, 2010 at 10:24 AM, Alex Harui aha...@adobe.com wrote: I would use a custom renderer. On 5/12/10 9:50 AM, Richard Rodseth rrods...@gmail.com wrote: I'm now aware of dropIndicatorSkin and found an example that customized the skin to do this: g.beginFill(0xa1bde2, 0.5); g.drawRect(-5, -1, w, 23); which is obviously a bit, er, hard-coded. The location and height for the skin appear to be set for drawing a line between nodes. Recall that I just want to highlight the node itself, as though it were being selected. Tree.showDropFeedback and ListBase.showDropFeedback are quite large methods (with side-effects) to be overriding. Is that my best option, or can I have my skin access the tree node? Recommendations or examples appreciated.I feel sure someone must have used a tree to represent a folder hierarchy. Thanks. On Tue, May 11, 2010 at 3:40 PM, Richard Rodseth rrods...@gmail.com mailto:rrodseth%40gmail.com rrodseth%40gmail.com wrote: I'm maintaining some code which has a dragEnter handler which sets selectedIndex in order to highlight the destination tree node Can anyone point me at some sample code to do this, or a public method to highlight the item without selecting it? The showDropFeedback method draws a line between items, which is not the desired effect. Thanks. -- Alex Harui Flex SDK Team Adobe System, Inc. http://blogs.adobe.com/aharui
Re: [flexcoders] Re: Folder-style dragOver highlight on tree
No, this is dragging items from a datagrid to a tree. On Wed, May 12, 2010 at 10:43 AM, Baz li...@thinkloop.com wrote: Doesn't the node already highlight because you are rolling over it?
Re: [flexcoders] Re: Folder-style dragOver highlight on tree
Though I strive to avoid subclassing the standard components, the code I
inherited already has a Tree subclass, so I added the following:
public function highlightItem(item:IListItemRenderer,
highlighted:Boolean):void {
this.drawItem(item, false, highlighted);
}
And call it from the renderer's dragEnter handler:
MyTreeSubclass(this.listData.owner).highlightItem(this, true);
Seems to be working OK.
On Wed, May 12, 2010 at 10:56 AM, Richard Rodseth rrods...@gmail.comwrote:
No, this is dragging items from a datagrid to a tree.
On Wed, May 12, 2010 at 10:43 AM, Baz li...@thinkloop.com wrote:
Doesn't the node already highlight because you are rolling over it?
Re: [flexcoders] Re: Folder-style dragOver highlight on tree
And of course I had to add to my renderers dragExit:
MyTreeSubclass(this.listData.
owner).highlightItem(this,false );
On Wed, May 12, 2010 at 11:10 AM, Richard Rodseth rrods...@gmail.comwrote:
Though I strive to avoid subclassing the standard components, the code I
inherited already has a Tree subclass, so I added the following:
public function highlightItem(item:IListItemRenderer,
highlighted:Boolean):void {
this.drawItem(item, false, highlighted);
}
And call it from the renderer's dragEnter handler:
MyTreeSubclass(this.listData.owner).highlightItem(this, true);
Seems to be working OK.
On Wed, May 12, 2010 at 10:56 AM, Richard Rodseth rrods...@gmail.comwrote:
No, this is dragging items from a datagrid to a tree.
On Wed, May 12, 2010 at 10:43 AM, Baz li...@thinkloop.com wrote:
Doesn't the node already highlight because you are rolling over it?
[flexcoders] Folder-style dragOver highlight on tree
I'm maintaining some code which has a dragEnter handler which sets selectedIndex in order to highlight the destination tree node Can anyone point me at some sample code to do this, or a public method to highlight the item without selecting it? The showDropFeedback method draws a line between items, which is not the desired effect. Thanks.
Re: [flexcoders] Re: mvc pattern for flex
I like Mate, but the lack of type safety in the event map *is* an issue. But
I don't find dealing with imports an issue. When entering something like:
MethodInvoker generator={ReportsManager} method=receiveData
I just delete the last two characters of ReportsManager and press
Control-Space.
On Thu, May 6, 2010 at 8:11 AM, Amy amyblankens...@bellsouth.net wrote:
--- In flexcoders@yahoogroups.com flexcoders%40yahoogroups.com, Wally
Kolcz wko...@... wrote:
I love and am a big advocate for Mate. Simple to learn, lots of good
built in tools, not MVC code in your views or business logic so if you
want to change down the road, just pop it out and handle the events.
My limited experience with Mate is that it does the opposite of what
whoever wrote the code completion stuff in Flex Builder was expecting, so
you wind up typing a lot of import statements and stuff that you could
normally let the editor handle. It seems like this would eat up any time you
might save by using this architecture.
[flexcoders] Overhead of curly braces :)
Something I've always wondered.
If I wish to use a constant, rather than a literal
eg
view:DailyView day={Days.MONDAY} /
does the compiler know that this is not a bindable property, or does
it add unnessary listeners?
Re: [flexcoders] Bindable metadata tag event name
Either I'm not following, or you have something else wrong.
If a model object (or the MXML component itself) has a property with custom
getter and setter as I showed, you can most certainly say
mx:Text text = {model.txt}/
or
mx:Text text = {this.txt} /
Of course I'm referring to one-way binding - I have not yet played with
binding in Flex 4.
On Sat, May 1, 2010 at 10:21 PM, sasuke uzumaki.naruto...@gmail.com wrote:
Hi,
Yes, even I've done that. But my main gripe is that when you have a
Bindable
tag with your custom event name, is it not possible to use it directly with
any custom/built-in flex control?
[Bindable(myEvent)]
private var txt:String = HI;
mx:Text text={txt} / // is this not possible if I use custom events?
I've always used plain bindable tags but I'd like to know whether I can
achieve the same auto-updation functionality when it comes to using
bindable
tags with custom events.
Regards,
sasuke
Richard Rodseth wrote:
I've never looked at the implementation. When I do this, I create a
custom
getter and setter:
[Bindable(fooChanged)]
public function get foo():Boolean {
return _foo;
}
public function set foo(value:Boolean):void {
if (value != _foo) {
_foo = value;
dispatchEvent(new Event(fooChanged))
}
}
Note that you can also have computed properties based on multiple
properties:
[Bindable(fooChanged)]
[Bindable(barChanged)]
public function get foobar():Boolean {
return _foo _bar;
}
On Sat, May 1, 2010 at 12:54 PM, sasuke
uzumaki.naruto...@gmail.comuzumaki.naruto.is%40gmail.com
wrote:
Hi folks,
The Bindable metadata tag allows an optional event name to be specified
which would be event name that would be dispatched when the property in
consideration changes. I've read somewhere that specifying a distinct
event
name for each bindable property reduces the performance overhead (less
work
for BindingUtils or so it seems, I can't seem to find the actual
implementation detail). But this doesn't work with bindable properties
when
used with flex controls. For e.g. this *works*
mx:Script
[Bindable]
private var txt:String = HI;
private function onCreationCompleteDone(event:Event):void
{
txt = Changed text!;
}
/mx:Script
mx:Text text={txt} /
This doesn't (notice the event name specified):
mx:Script
[Bindable(myEvent)]
private var txt:String = HI;
private function onCreationCompleteDone(event:Event):void
{
txt = Changed text!; // Never called!
}
/mx:Script
mx:Text text={txt} /
I did try analyzing the autogenerated AS files; the ones generated when
I
use simple [Bindable] is pretty easy to understand since it generates
plain
getter and setter for the `txt` property but when I use
[Bindable(someEvent)], the code generated is a bit cryptic with the
stuff
with Watchers and all that.
Would someone be kind enough to shed some light on this matter? How this
all
this Bindable stuff work under the hood?
TIA,
sasuke
--
View this message in context:
http://old.nabble.com/Bindable-metadata-tag---event-name-tp28422913p28422913.html
Sent from the FlexCoders mailing list archive at Nabble.com.
--
View this message in context:
http://old.nabble.com/Bindable-metadata-tag---event-name-tp28422913p28424902.html
Sent from the FlexCoders mailing list archive at Nabble.com.
Re: [flexcoders] Bindable metadata tag event name
I've never looked at the implementation. When I do this, I create a custom
getter and setter:
[Bindable(fooChanged)]
public function get foo():Boolean {
return _foo;
}
public function set foo(value:Boolean):void {
if (value != _foo) {
_foo = value;
dispatchEvent(new Event(fooChanged))
}
}
Note that you can also have computed properties based on multiple
properties:
[Bindable(fooChanged)]
[Bindable(barChanged)]
public function get foobar():Boolean {
return _foo _bar;
}
On Sat, May 1, 2010 at 12:54 PM, sasuke uzumaki.naruto...@gmail.com wrote:
Hi folks,
The Bindable metadata tag allows an optional event name to be specified
which would be event name that would be dispatched when the property in
consideration changes. I've read somewhere that specifying a distinct event
name for each bindable property reduces the performance overhead (less work
for BindingUtils or so it seems, I can't seem to find the actual
implementation detail). But this doesn't work with bindable properties when
used with flex controls. For e.g. this *works*
mx:Script
[Bindable]
private var txt:String = HI;
private function onCreationCompleteDone(event:Event):void
{
txt = Changed text!;
}
/mx:Script
mx:Text text={txt} /
This doesn't (notice the event name specified):
mx:Script
[Bindable(myEvent)]
private var txt:String = HI;
private function onCreationCompleteDone(event:Event):void
{
txt = Changed text!; // Never called!
}
/mx:Script
mx:Text text={txt} /
I did try analyzing the autogenerated AS files; the ones generated when I
use simple [Bindable] is pretty easy to understand since it generates plain
getter and setter for the `txt` property but when I use
[Bindable(someEvent)], the code generated is a bit cryptic with the stuff
with Watchers and all that.
Would someone be kind enough to shed some light on this matter? How this
all
this Bindable stuff work under the hood?
TIA,
sasuke
--
View this message in context:
http://old.nabble.com/Bindable-metadata-tag---event-name-tp28422913p28422913.html
Sent from the FlexCoders mailing list archive at Nabble.com.
Re: [flexcoders] Re: Binding and square brackets
without introducing an ArrayCollection
The data is an array and always changes atomically. Wrapping it in an
ArrayCollection is certainly one option.
On Fri, Apr 30, 2010 at 8:15 AM, Amy amyblankens...@bellsouth.net wrote:
--- In flexcoders@yahoogroups.com flexcoders%40yahoogroups.com, Richard
Rodseth rrods...@... wrote:
I'm interested in getting rid of Data binding will not be able to
detect changes when using square bracket operator without introducing
an ArrayCollection.
The view displays data for the seven days of the week. Not using a
repeater (though one could).
I invite opinions on the relative merits (performance/elegance) of the
following:
A) Split the data in the presentation model when the combined data comes
in
c:DailyView day=Monday data={model.mondayData}
c:DailyView day=Tuesday data={model.tuesdayData}
Obviously only an option because there are only seven days.
B) Use a function in the view
private function dailyData(combinedDailyData:Array, dayIndex:int) {
return model.combinedData[day]; // could also use a dictionary
}
c:DailyView day=Monday data={this.dailyData(model.stuff, 0)}/
C) A bindable non-getter in the presentation model
c:DailyView day=Monday data={model.dataForDay(0)}/
In the model:
[Bindable(event=combinedDataChanged)]
public function dataForDay(day:int):Array {
return combinedData[day];
}
Obviously the setter for combinedData dispatches combinedDataChanged.
I actually wasn't aware, or had forgotten, that C) is possible on a
non-getter function until I tried it just now. I actually like this
option because it encapsulates the storage used (array, dictionary,
whatever). I think I just answered my own question.
Had you considered just using getItemAt() on your ArrayCollection?
[flexcoders] Flashbuilder hang loading workbench
Not a good way to end the week. FlashBuilder hangs loading the workbench. Spinning beach ball, must Force Quit. Is there some cache I can delete to get up and running again? I saw a similar post on the flex forums from a few days ago, but no one has responded.
[flexcoders] Binding and square brackets
I'm interested in getting rid of Data binding will not be able to
detect changes when using square bracket operator without introducing
an ArrayCollection.
The view displays data for the seven days of the week. Not using a
repeater (though one could).
I invite opinions on the relative merits (performance/elegance) of the
following:
A) Split the data in the presentation model when the combined data comes in
c:DailyView day=Monday data={model.mondayData}
c:DailyView day=Tuesday data={model.tuesdayData}
Obviously only an option because there are only seven days.
B) Use a function in the view
private function dailyData(combinedDailyData:Array, dayIndex:int) {
return model.combinedData[day]; // could also use a dictionary
}
c:DailyView day=Monday data={this.dailyData(model.stuff, 0)}/
C) A bindable non-getter in the presentation model
c:DailyView day=Monday data={model.dataForDay(0)}/
In the model:
[Bindable(event=combinedDataChanged)]
public function dataForDay(day:int):Array {
return combinedData[day];
}
Obviously the setter for combinedData dispatches combinedDataChanged.
I actually wasn't aware, or had forgotten, that C) is possible on a
non-getter function until I tried it just now. I actually like this
option because it encapsulates the storage used (array, dictionary,
whatever). I think I just answered my own question.
[flexcoders] Selection highlighting in Flashbuilder 4
I just started using FB4 (standalone) for real work, and the selection highlighting in the editor is driving me nuts. Is there some preference I need to change? Basically, I get no highlighting when dragging over text, until I release the mouse. I'm on OS X.
[flexcoders] State of the art JSON
I've used AS3CoreLib's JSON encoding/decoding in the past, but some code I'm maintaining now has custom JSON handling. Note: we use ISO 8601 for Dates. Looking at the Flashbuilder 4 data wizards (briefy), I see JSON support and references to something call fiber, but am unclear whether this is tied to LiveCycle, and whether it's Flex 4 specific. So I guess I'm wondering what the best-practice is these days for JSON support a) With Flex 3 b) With Flex 4 Thanks.
Re: [flexcoders] Drag and drop vs scrolling
Thanks. I was able to fix the highlighting issue by modifying some code that used listData.rowIndex to use calculateDropIndex() instead. Autoscroll doesn't work, but I can live with that. On Tue, Apr 20, 2010 at 11:41 PM, Alex Harui aha...@adobe.com wrote: Some folks have tried making the renderer a drop target in order to add different meaning to dropping in folders. In theory it should work, but I think I’d work from the Tree instead, and then I would expect autoscroll to work. On 4/20/10 3:31 PM, Richard Rodseth rrods...@gmail.com wrote: I inherited some code which supports drag and drop from a list to a node of a tree. Unfortunately it does not work when the tree is scrolled. Not only does the tree not autoscroll, but even if you scroll the target item into view before beginning the drag, incorrect nodes highlight. I didn't see any code that's doing coordinate transformations, but did notice that there is a custom TreeItemRenderer, which contains the listeners for dragEnter etc. I'm guessing this is the root of the problem, since renderers are recycled. I've no idea why things were done this way, and am wondering if this is a legitimate idiom, or if I need to start over with listeners defined at the Tree level. Also, is autoscroll supported if you do things the right way? -- Alex Harui Flex SDK Team Adobe System, Inc. http://blogs.adobe.com/aharui
[flexcoders] Drag and drop vs scrolling
I inherited some code which supports drag and drop from a list to a node of a tree. Unfortunately it does not work when the tree is scrolled. Not only does the tree not autoscroll, but even if you scroll the target item into view before beginning the drag, incorrect nodes highlight. I didn't see any code that's doing coordinate transformations, but did notice that there is a custom TreeItemRenderer, which contains the listeners for dragEnter etc. I'm guessing this is the root of the problem, since renderers are recycled. I've no idea why things were done this way, and am wondering if this is a legitimate idiom, or if I need to start over with listeners defined at the Tree level. Also, is autoscroll supported if you do things the right way?
Re: [flexcoders] Exceeding Flash's Capabilities?
How many items are you (and the Excel user) talking about? What sort of Flex charts are you using? Are you using AMF, XML or JSON to retrieve the data? On Wed, Apr 14, 2010 at 3:48 PM, trefalgar trefal...@yahoo.com wrote: The majority of my flex work revolves around graphing data. If for some reason a user picks a lot of statistics to graph, over an even longer period of time, the amount of data that comes back to the Flash application overwhelms flash, consuming all of the boxes resources and regularly crashing the browser. Is there anything in the developers bag of tricks to help with this kind of problem, or is the sheer amount of data too much for the application/SDK/solution to handle? I understand I can put a limit on the amount of data returned, but as the user states, I can pull it out of the database, via Excel, and it works. Tref
Re: [flexcoders] Re: Property values based on Multiple Conditions
I did a little Googling for you:
http://twopointoh.vox.com/library/post/boolean-operations-in-mxml.html
Depending on what you are building, the Presentation Model pattern may be
overkill, but it's worth knowing about.
On Tue, Apr 13, 2010 at 12:28 AM, Doug doug_pie...@ymail.com wrote:
Awesome stuff guys!!
The amp; thing didn't work for me when I tried that... I thought just
maybe, but nah. This was part of the source of my question, How do you do a
logical AND?
The function method that Richard posted is something I am going to try as I
think that's exactly what I'm looking for. I would like to avoid setting up
an event listener, but if that's the only way...
Thanks again for the help thus far... if anybody has anything else, please
let me know!
--- In flexcoders@yahoogroups.com flexcoders%40yahoogroups.com, Richard
Rodseth rrods...@... wrote:
Rick is correct. You can also use a local binding function like this:
s:Button enabled={enableButton(showMe, hideMe)}/
where enableButton is a function in your Script block that takes two
boolean
parameters.
Or, if you have a presentation model, it could have a computed property
buttonEnabled
s:Button enabled={model.buttonEnabled}/
Excerpt of presentation model:
[Bindable(event=showMeChanged)]
[Bindable(event=hideMeChanged)]
public function get buttonEnabled():Boolean {
return showMe hideMe;
}
public function set showMe(value:Boolean):void {
if (value != this.showMe) {
this.showMe = value;
dispatchEvent(new Event(showMeChanged));
}
}
Hope this helps.
On Mon, Apr 12, 2010 at 4:31 PM, Rick Genter rick.gen...@... wrote:
On Apr 12, 2010, at 10:31 AM, Doug wrote:
I seem to have little success with this and would like to know other
ways
of accomplishing what I want to do.
I'm trying to enable/disable a button based on two pieces of
criteria. It
would seem that a logical OR or a logical AND would bridge these much
like
an IF statement, however, I'm struggling to make it work 100%.
Consider a button:
s:Button enabled={showMe || hideMe}/
This button's enabled property is based upon the binding of the
variables
for showMe and hideMe. If showMe or hideMe is true then the button
is
enabled. This works fine when using a logical OR. However, the
following -
using a logical AND - does not:
s:Button enabled={showMe hideMe}/
How do you do a logical AND?? Also, if I wanted to set showMe and
hideMe
based on several lines of code how would I best do it? How would you
bind a
function output to a display object? I only see a way to do it based
upon
event:
[Bind (event: some event here]
private function myFunction():void{}
Thanks in advance!!
I haven't tried this, but I suspect that you are being bitten by XML.
Try
this:
s:Button enabled={showMe amp;amp; hideMe}/
--
Rick Genter
rick.gen...@... rick.genter%40gmail.com
[flexcoders] Variable-height popup
In the context of: mx:PopUpButton mx:popup mx:VBox /mx:VBox /mx:popup /mx:PopUpButton Is there a simple way to give the popup a height of, say, 50% of the current application height, along with a min and max height?
Re: [flexcoders] Property values based on Multiple Conditions
Rick is correct. You can also use a local binding function like this:
s:Button enabled={enableButton(showMe, hideMe)}/
where enableButton is a function in your Script block that takes two boolean
parameters.
Or, if you have a presentation model, it could have a computed property
buttonEnabled
s:Button enabled={model.buttonEnabled}/
Excerpt of presentation model:
[Bindable(event=showMeChanged)]
[Bindable(event=hideMeChanged)]
public function get buttonEnabled():Boolean {
return showMe hideMe;
}
public function set showMe(value:Boolean):void {
if (value != this.showMe) {
this.showMe = value;
dispatchEvent(new Event(showMeChanged));
}
}
Hope this helps.
On Mon, Apr 12, 2010 at 4:31 PM, Rick Genter rick.gen...@gmail.com wrote:
On Apr 12, 2010, at 10:31 AM, Doug wrote:
I seem to have little success with this and would like to know other ways
of accomplishing what I want to do.
I'm trying to enable/disable a button based on two pieces of criteria. It
would seem that a logical OR or a logical AND would bridge these much like
an IF statement, however, I'm struggling to make it work 100%.
Consider a button:
s:Button enabled={showMe || hideMe}/
This button's enabled property is based upon the binding of the variables
for showMe and hideMe. If showMe or hideMe is true then the button is
enabled. This works fine when using a logical OR. However, the following -
using a logical AND - does not:
s:Button enabled={showMe hideMe}/
How do you do a logical AND?? Also, if I wanted to set showMe and hideMe
based on several lines of code how would I best do it? How would you bind a
function output to a display object? I only see a way to do it based upon
event:
[Bind (event: some event here]
private function myFunction():void{}
Thanks in advance!!
I haven't tried this, but I suspect that you are being bitten by XML. Try
this:
s:Button enabled={showMe amp;amp; hideMe}/
--
Rick Genter
rick.gen...@gmail.com rick.genter%40gmail.com
[flexcoders] Flash Builder Subversive
I'm referring to the Subversion Eclipse plug-in. Has anyone installed it in Flash Builder? I found a blog post about Subclipse, but not Subversive. I *really* wish Flash Builder had subversion support built-in...
Re: [flexcoders] Flash Builder Subversive
I use Flexbuilder 3 plug-in the same way, but always seem to have trouble with installation, so I'm thinking of switching to the the standalone version this time. I just installed Subclipse - maybe I'll give it a try. http://blogs.adobe.com/jasonsj/2010/03/installing_subclipse_in_flash_builder_4.html On Thu, Mar 25, 2010 at 10:40 AM, Julien Nicoulaud primo...@gmail.comwrote: Of course, Flash Builder is just an Eclipse plugin. I use it with Subversive and many other Eclipse plugins without any conflict. 2010/3/25 Richard Rodseth rrods...@gmail.com I'm referring to the Subversion Eclipse plug-in. Has anyone installed it in Flash Builder? I found a blog post about Subclipse, but not Subversive. I *really* wish Flash Builder had subversion support built-in...
Re: [flexcoders] Re: Framework Choice
I've been pretty happy with Mate and the Presentation Model pattern. My main issue is the lack of type-safety in the event maps, so at some point I should check out Swiz, Parsley etc. On Wed, Mar 24, 2010 at 10:00 AM, Oleg Sivokon olegsivo...@gmail.comwrote: If you know there are Cairngorm and Mate, you can be sure, you can write one of your own. And if you used either one of those even briefly, you can write a better one. This said, my choice would be none. I'm writing that myself, and, you just cannot imagine my happiness - two days ago I finally removed the last Cairngorm piece from the big project I inherited long ago and had a hell of a time refactoring. Don't make yourself a victim of other short-witted programmers, unless you absolutely must. Your own errors are much easier to fix, especially, if you work alone or with a small team. Best. Oleg
