Skip to search.
flexcoders · RIA Development with Adobe Flex

Group Information

  • Members: 9017
  • Category: Development
  • Founded: Mar 17, 2004
  • Language: English
? Already a member? Sign in to Yahoo!

Yahoo! Groups Tips

Did you know...
Message search is now enhanced, find messages faster. Take it for a spin.

Messages

  Messages Help
Advanced
Flex compilation bug   Message List  
Reply Message #21926 of 165212 |
Re: [flexcoders] Re: Flex compilation bug

Thanks for the taking the time to detail out the issue.
Going to try the workaround mentioned and see if that helps.

- superabe


----- Original Message -----
From: "Eric Raymond" <ericraymondmaps@...>
To: <flexcoders@yahoogroups.com>
Sent: Tuesday, September 27, 2005 8:33 PM
Subject: [flexcoders] Re: Flex compilation bug


> It is likely you are experiencing the dreaded "NoChange" Flex compiler
> bug. It is largely silent and affected by incremental compilation.
>
> It is usually the result of some cyclic dependency between various
> source files ... and the reference is via actionscript (embedded in
> mxml or in a source .as file).
>
> If you configure Flex to leave the generated actionscript files, you
> can search for a listener with "NoChange" (I think?) in the .as files.
>
> If you are using incremental compilation, the order in which you
> change files can make this go away (and come back). (So one ugly work
> around is to delete and touch files in a certain order each time the
> problem occurs). The command line compiler always breaks (IMHO this
> is better) ... albeit silently.
>
> The best way to fix this is to use mxmlc to compile the files and make
> changes to your source until it is gone. COmment out code until the
> error disappears.
>
> This has been acknowledged by Macromedia and it has been suggested
> that it will be gone in 2.0 (translation: "You are going to have to
> deal with this yourself."). And because this is really a problem at
> the application level, it's not easy to send support a stripped down
> test case (and once you gotten to that point, you've basically have
> your solution).
>
> Good luck. This is not fun.
>
>
> P.S. Joe Berkovitz gave me the following excellent advice (this
> thread is in the archive):
>
> In short, turn on the keep-generated option in flex-config.xml, look in
> the <classname>-generated.as files for the string "NoChange". If you
> find it in the generated code for watcher setup, this is most likely
> the compiler bug. You could verify that it's the bug by touching the
> affected files and doing an incremental recompile by re-requesting the
> app, then comparing with the updated -generated.as files; they should
> be different.
>
> I worked around the problem by having the AS class which instantiates
> the MXML component X also reference an MXML component (call it
> XWrapper) which *includes* the component X by tag reference, binding
> the broken properties to any value at all.
>
> So, in your .as file (which presumably does createChild(X, ...) at some
> point) you'd do this:
>
> static var XWrapperRef = XWrapper; //
>
> and in XWrapper.mxml, the root components would be:
>
> <X property1="{undefined}" property2="{undefined}"/>
>
> where X.mxml is the file in which bindings to property1, etc. are
> broken.
>
> Note that XWRapper is only referenced as a class, it is never
> instantiated! That's enough to change the compiler behavior.
>
>
>
>
> --- In flexcoders@yahoogroups.com, "superabe" <superabe@g...> wrote:
>> Am facing a mystery bug that I can't seem to fathom. Hoping somebody
> on the
>> list know's of a solution.
>> The flex server is set to debug mode and the app uses the Cairngorm
>> Framework and the FAST toolkit.
>>
>> Basically what seems to happen is every time I restart my flex
> server it
>> fails to recognize updates to certain components/renderers.
>>
>>
>> For E.g.
>>
>> I have a number of custom components extending of the TitleWindow,
> that are
>> instantiated using the PopupManager.
>> These have certain components being used as "renderers" within a reeater
>>
>> <mx:Repeater id="categoriesList" dataProvider="{planDetail.categories}"
>>
> repeatEnd="initOptions(ModelLocator.cart.isPlanInCart(planDetail.id))">
>> <popupRenderer:OptionsRenderer id="optionsRenderer"
>> categoryName="{categoriesList.currentItem.categoryName}"
>> options="{categoriesList.currentItem.options}"
>> optionUpdate="updatePlanPrice()">
>> </popupRenderer:OptionsRenderer>
>>
>> The popupRenderer namespace is pointed to
>> "com.company.appName.view.productDisplay.renderers.*"
>>
>> When I restart my flex server, the repeater works fine and show "n"
> number
>> of OptionsRenderers, but fails to display the values for
> categoryName and
>> options.
>> Now if I open up the mxml file that contains the Repeater and add a
> dummy
>> line of code, like (Echo.debug()), save it and refresh the browser, the
>> values show up and everything is fine.
>> If I restart my flex server the same problem happens again.
>>
>> I initially thought this might be happening only for Popups or
> Repeaters,
>> but I have seen this also in cases where I have components that are not
>> being used in popups and are not in Repeaters
>>
>> Anybody have any ideas / suggestions?
>> Any help is appreciated as this seems to be a flex bug and am not
> sure how
>> to tackle this.
>>
>> Thanks,
>>
>> - superabe
>
>
>
>
>
> --
> Flexcoders Mailing List
> FAQ: http://groups.yahoo.com/group/flexcoders/files/flexcodersFAQ.txt
> Search Archives: http://www.mail-archive.com/flexcoders%40yahoogroups.com
> Yahoo! Groups Links
>
>
>
>
>
>




Wed Sep 28, 2005 2:24 pm

superabe@...
Send Email Send Email

Message #21926 of 165212 |
Expand Messages Author Sort by Date

Am facing a mystery bug that I can't seem to fathom. Hoping somebody on the list know's of a solution. The flex server is set to debug mode and the app uses...
superabe
superabe@... Send Email
Sep 27, 2005
9:24 pm

It is likely you are experiencing the dreaded "NoChange" Flex compiler bug. It is largely silent and affected by incremental compilation. It is usually the...
Eric Raymond Offline Send Email Sep 28, 2005
12:33 am

Thanks for the taking the time to detail out the issue. Going to try the workaround mentioned and see if that helps. - superabe ... From: "Eric Raymond"...
superabe
superabe@... Send Email
Sep 28, 2005
2:17 pm

Just checking to see if anyone has made any progress on this issue, because I'm experiencing it now. Any other methods of dealing with it that people have come...
fowleryj Offline Send Email Oct 6, 2005
8:04 pm
Advanced

Copyright © 2010 Yahoo! Inc. All rights reserved.
Privacy Policy - Terms of Service - Guidelines NEW - Help