And, of course, when I get back on-line and read this, it has expired. Oh well, if you don't find someone, fire me off an email.
--- In [email protected], "edlueze" <edlu...@...> wrote: > > Hi Steve - thanks for the tips! > > Sorry for not getting back to you sooner (just got back online). The > Rent-a-coder project should still be active - the link is: > > http://www.rentacoder.com/RentACoder/misc/BidRequests/ShowBidRequest.asp?lngBidRequestId=1153997 > > All the best, > > Ted. > > --- In [email protected], "valdhor" <valdhorlists@> wrote: > > > > Generally what I do is create separate sub classes (Of MDIWindow) for each > > of the windows I need. I try to treat each window as a completely self > > contained "module". Each window is managed by MDIManager. Of course, you > > have to treat events differently as they will be parented by systemManager. > > > > Each window will then contain one or more components (Either Flex or > > custom). Usually the main part of the window will contain a datagrid then > > the controlbar will contain next/previous links and an export data link. I > > have to include code to make these links invisible if the user minimizes > > the window otherwise they clutter the display. > > > > If i were you, I would create custom components containing the display > > objects you need and add them to a custom MDIWindow class. > > > > > > Best Regards > > > > > > > > Steve > > > > PS. I was going to bid on your request on rentacoder but couldn't find it. > > Have you pulled it? > > > > > > --- In [email protected], "edlueze" <edlueze@> wrote: > > > > > > Hi Steve / Everybody: > > > > > > The FlexMDI library has invigorated my enthusiasm to re-think the layout > > > of the application - it was getting obvious that the application was > > > getting far too cluttered anyway. It sounds like you've developed some > > > best practices and I'm eager to hear about them. > > > > > > The initial architecture called for the following types of windows (with > > > a small "w"). Note that these names defined in the architecture do not > > > match the Flex names as they pre-date the decision to even use Flex: > > > > > > - "Canvas" (large spanning windows - it would be great if the user could > > > toggle these take up the entire screen) > > > - "Windows" (wide-but-short windows designed to take the entire width of > > > the screen - they do not always need to be showing) > > > - "Property Window" (type of Window containing the properties of whatever > > > is selected in the Canvas) > > > - "Tabs" within the Property Window > > > - "Panels" (small floating windows with many purposes - it would be nice > > > to also be able to group and anchor these) > > > - "Dialog Boxes" (modal pop-ups) > > > - "Wizards" (fancy modal pop-ups) > > > > > > Do you have any ideas on how I can take this framework and adapt it to > > > FlexMDI or whatever else you like using? > > > > > > All the best, > > > > > > Ted. > > > > > > --- In [email protected], "valdhor" <valdhorlists@> wrote: > > > > > > > > It looks as though it is not a true window. > > > > > > > > I don't think I would have done it the way you have (But who am I to > > > > tell you how to do something ;-). > > > > > > > > Does it still happen with only one accordion item? > > > > > > > > Maybe it's a deferred instantiation problem in the tab navigator. What > > > > happens if you set all tabs to instantiate? > > > > > > > > I really don't think this is something we are going to be able to > > > > figure out in a forum and posting code is obviously out of the > > > > question. If I can free up some time I may respond to your ad. > > > > > > > > > > > > --- In [email protected], "edlueze" <edlueze@> wrote: > > > > > > > > > > Hi Steve - everybody: > > > > > > > > > > The FlexMDI definitely looks interesting! And I might take you up on > > > > > your offer to pay you to take a look at our bug. > > > > > > > > > > For everybody's benefit, I've put together a simple PowerPoint giving > > > > > an example of the problem. You can view the PPT here: > > > > > > > > > > http://www.onegen.com/download/FlexWindowNotLoadingBug001.ppt > > > > > > > > > > Thanks really for all your time and help - this one's got us baffled! > > > > > > > > > > Ted. > > > > > > > > > > --- In [email protected], "valdhor" <valdhorlists@> wrote: > > > > > > > > > > > > You may like to look at FlexMDI (Part of FlexLib) for your > > > > > > windowing requirements. I use an older version extensively and have > > > > > > not seen anything like what you are experiencing. > > > > > > > > > > > > > > > > > > Best Regards > > > > > > > > > > > > > > > > > > > > > > > > Steve > > > > > > > > > > > > PS. If you want to pay me to look at it I'll gladly do so ;-} > > > > > > > > > > > > > > > > > > --- In [email protected], "edlueze" <edlueze@> wrote: > > > > > > > > > > > > > > Please help me fix a windowing integration bug in our Adobe Flex > > > > > > > application. I'd love some free advice, but we really need the > > > > > > > problem solved so we've also posted the question on rent-a-coder. > > > > > > > > > > > > > > All of our Flex windows > > > > > > > (Application/Canvas/Panel/TitleWindow/etc) work fine individually > > > > > > > when running in a browser (Firefox and Internet Explorer). Now we > > > > > > > are trying to integrate all of the Windows together (less than > > > > > > > 100 in all) and we are observing some pretty bizarre behavior - > > > > > > > sometimes a window will load and other times the same window will > > > > > > > not load. But the problem is not intermittent - it happens very > > > > > > > often and we can usually force a window failure on demand. > > > > > > > > > > > > > > When the window does not load we observe the CPU usage in Task > > > > > > > Manager hitting the maximum, and we believe that the Flex > > > > > > > Application has fallen into an endless race condition in which it > > > > > > > is constantly resizing the components within the window (we > > > > > > > randomly pause the application and usually find the stack doing > > > > > > > something like this). When this happens the window's > > > > > > > creationComplete event doesn't trigger and we cannot insert a > > > > > > > breakpoint. > > > > > > > > > > > > > > We make heavy use of constraintRows and constraintColumns and we > > > > > > > are suspicious of them. We use a lot of Accordion's as well as > > > > > > > the custom HAccordion from Flexlib, but we've had trouble > > > > > > > positively identifying them as the culprit. Each window can get > > > > > > > pretty buried in a deep hierarchy of windows, but it is not > > > > > > > always the deepest windows that will fail to load. > > > > > > > > > > > > > > We've also implemented a suspicious custom component early on to > > > > > > > handle system-wide ToolTips. It is old code and does some funky > > > > > > > things, but when we commented it out the behavior didn't > > > > > > > dramatically improve. > > > > > > > > > > > > > > Our testing has been limited to the Debugger Flash Player, but > > > > > > > we've tried both debug and release versions of the application. > > > > > > > > > > > > > > I've combed Flexcoders and I've not come across anything like it. > > > > > > > I'd love to hear your ideas?! > > > > > > > > > > > > > > > > > > > > > > > > > > > >
