Unfortunately, FB does not call our code in a useful fashion during a "clean". That's something that may also be a limitation of FB. We'll see if I find time to dig into FB more and find a workaround. The workaround may be to always clean the bin folder when running in FB.
-Alex On 11/7/17, 5:58 AM, "Harbs" <[email protected]> wrote: >FB should probably delete the bin folder when performing clean, but it >could be that it only removes the swf bin-debug folder and not the js bin >folder. > >> On Nov 7, 2017, at 3:56 PM, Harbs <[email protected]> wrote: >> >> Glad you figured it out! >> >> “It’s a feature — not a bug”. ;-) >> >> The compiler does not replace files from swcs when doing subsequent >>compiles. This has two advantages: >> >> 1. Subsequent compiles are faster. >> 2. It allows you to manually change the js files included in swcs and >>recompile your apps. I’ve used this while trying to track down bugs. >> >> Our ant script has a clean before compile. When I use asconfigc to >>compile I just delete the entire bin folder when the contents of one of >>the swcs I’m using has changed. >> >> We should probably document this somewhere before we forget… >> >> HTH, >> Harbs >> >>> On Nov 7, 2017, at 3:31 PM, Idylog - Nicolas Granon >>><[email protected]> wrote: >>> >>> I have finally fixed the instantiation issue. >>> >>> It turns out that the compiler (when invoked from FB) does *not* >>> systematically replace "lib" files (js dependencies in the >>>org.apache... >>> subfolders) even if performing a "clean build". For some unknown >>>reason the >>> CurrencyFormatter js file was empty and was never replaced by a "good" >>> one... >>> >>> I'm not sure why... maybe there is some "lock" problem between FB and >>>the >>> compiler, maybe the compiler assumes that if a file is already there >>>it must >>> not be replaced... >>> I don't know for sure. It's not what is expected from a clean build, >>> anyway... >>> >>> I also noted that the compiler does not remove unused dependencies in >>>the >>> output dir (even after a clean build). >>> >>> But the fact is, I closed FB, manually cleaned the output dir, >>>relaunched FB >>> and rebuilt and *then* the problem disappeared. >>> >>> Nicolas Granon >>> >>> >>> >>> >>>> -----Message d'origine----- >>>> De : Harbs [mailto:[email protected]] >>>> Envoyé : lundi 6 novembre 2017 16:24 >>>> À : [email protected]; [email protected] >>>> Objet : Re: Currency formatter instanciation issue (was RE: >>>> ApplicationComplete event not firing) >>>> >>>> Can you post your console output from compiling the app and the >>>> contents of the index.html file? >>>> >>>>> On Nov 6, 2017, at 5:18 PM, Idylog - Nicolas Granon >>>> <[email protected]> wrote: >>>>> >>>>> FlashBuilder 4.7, Win7, Apache Royale 0.9 binaries : file name : >>>>> apache.royale-jsonly-0.9.0-bin.zip downloaded on 10/25/2017. >>>>> >>>>> Nicolas Granon >>>>> >>>>> >>>>> >>>>> >>>>>> -----Message d'origine----- >>>>>> De : Harbs [mailto:[email protected]] Envoyé : lundi 6 novembre >>>>>> 2017 16:10 À : [email protected] Objet : Re: Currency formatter >>>>>> instanciation issue (was RE: >>>>>> ApplicationComplete event not firing) >>>>>> >>>>>> Related: >>>>>> >>>>>> What are you using to build your app? >>>>>> >>>>>>> On Nov 6, 2017, at 5:01 PM, Yishay Weiss <[email protected]> >>>>>> wrote: >>>>>>> >>>>>>> Does >>>>>>> >>>>>>> >>>>>>> >>>>>>> %ROYALE_HOME%/frameworks/libs/Formatters.swc exist? >>>>>>> >>>>>>> >>>>>>> >>>>>>> ________________________________ >>>>>>> From: Idylog - Nicolas Granon <[email protected]> >>>>>>> Sent: Monday, November 6, 2017 4:27:22 PM >>>>>>> To: [email protected] >>>>>>> Subject: Currency formatter instanciation issue (was RE: >>>>>>> ApplicationComplete event not firing) >>>>>>> >>>>>>> No, declaring the formatter in a <fx:declaration/> tag doesn't work >>>>>>> either... >>>>>>> >>>>>>> Execution is stopped after preinitialize event fires. >>>>>>> Initialize and applicationComplete events never fire. (I presume >>>>>>> that a declaration tag is processed at initialize step ?). >>>>>>> >>>>>>> (Thank you for the link to the examples repo !) >>>>>>> >>>>>>> Nicolas Granon >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>>> -----Message d'origine----- >>>>>>>> De : Yishay Weiss [mailto:[email protected]] Envoyé : lundi 6 >>>>>>>> novembre 2017 14:53 À : [email protected]; [email protected] >>>>>>>> Objet : RE: ApplicationComplete event not firing >>>>>>>> >>>>>>>> Can’t see anything wrong with what you’ve given us. Does this [1] >>>>>> run >>>>>>>> for you? >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> BTW, feel free to post examples of problems (or successful runs) >>>> on >>>>>>>> that [1] repo. >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> [1] >>>>>>>> >>>>>> >>>> >>>>https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub >>>>.com%2Fyishayw%2FExamples%2Fblob%2FCurrencyFormatter%2FExamples.m&data= >>>>02%7C01%7C%7Cb6776b33f1954997f61008d525e79a61%7Cfa7b1b5a7b34438794aed2c >>>>178decee1%7C0%7C0%7C636456599011814830&sdata=TPosroiXbXs9s4QqNa0lgmQvUR >>>>RL9XEdGN6Wu8fA3qQ%3D&reserved=0 >>>>>>>> xm >>>>>>>> l >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> ________________________________ >>>>>>>> From: Idylog - Nicolas Granon <[email protected]> >>>>>>>> Sent: Monday, November 6, 2017 2:35:28 PM >>>>>>>> To: 'Harbs'; [email protected] >>>>>>>> Subject: RE: ApplicationComplete event not firing >>>>>>>> >>>>>>>> You are right ! It does fire...unless some "silent" error happens >>>>>>>> while building the initial view ! >>>>>>>> >>>>>>>> We found that there was an error in the listener for the inner >>>> view >>>>>>>> "initComplete" event. >>>>>>>> This "inner error" prevents the applicationComplete to fire. >>>>>>>> >>>>>>>> But we do not understand why this error happens : >>>>>>>> >>>>>>>> The "inner view" initComplete listener error is caused by the >>>>>>>> following statements : >>>>>>>> var cf:CurrencyFormatter; >>>>>>>> cf=new CurrencyFormatter(); >>>>>>>> >>>>>>>> It is the instantiation of the currency formatter which causes >>>> some >>>>>>>> problem (???). >>>>>>>> >>>>>>>> After adding a try/catch block, we get this message from the error >>>>>>>> object : >>>>>>>> org.apache.royale.html.accessories is undefined >>>>>>>> >>>>>>>> although, of course, we import it like this (import is generated >>>> by >>>>>>>> code assist) : >>>>>>>> import org.apache.royale.html.accessories.CurrencyFormatter; >>>>>>>> >>>>>>>> What are we doing wrong ? >>>>>>>> >>>>>>>> >>>>>>>> Nicolas Granon >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>>> -----Message d'origine----- >>>>>>>>> De : Harbs [mailto:[email protected]] Envoyé : lundi 6 >>>>>>>>> novembre >>>>>>>>> 2017 12:53 À : [email protected]; [email protected] Objet : >>>>>> Re: >>>>>>>>> ApplicationComplete event not firing >>>>>>>>> >>>>>>>>> We are using applicationComplete and it definitely fires. >>>>>>>>> >>>>>>>>> If you post some code maybe I’d have some input. >>>>>>>>> >>>>>>>>> The event type for most events in Royale is Event. Some events >>>>>>>>> have a payload and they have a different event type (such as >>>>>> ValueEvent), >>>>>>>> but >>>>>>>>> events in Royale are much more generalized than Flex. >>>>>>>>> >>>>>>>>> HTH, >>>>>>>>> Harbs >>>>>>>>> >>>>>>>>>> On Nov 6, 2017, at 12:01 PM, Idylog - Nicolas Granon >>>>>>>>> <[email protected]> wrote: >>>>>>>>>> >>>>>>>>>> (maybe this is not the right place for this kind of message ? >>>> let >>>>>>>> me >>>>>>>>>> know) >>>>>>>>>> >>>>>>>>>> When running our (almost) Hello world application (JS target) it >>>>>>>>> seems >>>>>>>>>> that "applicationComplete" event does not fire. >>>>>>>>>> >>>>>>>>>> "preinitialize" and "initialize" events both fire. >>>>>>>>>> >>>>>>>>>> The application "viewChanged" event does not fire on initial >>>>>>>>>> load, >>>>>>>>> but >>>>>>>>>> maybe it fires only when the is a view *change*, not when the >>>>>> first >>>>>>>>>> view in "initial view" loads ? >>>>>>>>>> >>>>>>>>>> Also, what is the class of these events ? Code assist shows the >>>>>>>>>> event name, but not its class (???). For now, I declare them as >>>>>>>>>> "Event" but I like to correctly type my arguments... >>>>>>>>>> >>>>>>>>>> Nicolas Granon >>>>>>>>>> >>>>>>>> >>>>>>> >>>>>>> >>>>> >>>>> >>> >>> >> >
