It would be a separate compile as you noted but what other choice is there if we don't have access to the FB sources?
>>You definitely wouldn't get Falcon-based code hinting, and the data >>structures of the old Code Model would be taking up memory. Not sure why we wouldn't get Falcon-based code hinting. FB uses Falcon for code hinting today. Not planning to change that. On 9/4/13 1:20 PM, "Gordon Smith" <gosm...@adobe.com> wrote: >> The current FB can't use Falcon > >By making a new flex-compiler-oem.jar based on Falcon, you could make FB >use Falcon to compile. But I don't think you'd get memory-resident data >structures, which was one of the main points of Falcon... I think with >the flex-compiler-oem.jar approach you'd be starting over with each >compilation, parsing files, parsing SWCs, building symbol tables. >(Darrell, am I wrong here?) > >You definitely wouldn't get Falcon-based code hinting, and the data >structures of the old Code Model would be taking up memory. > >The point of Falcon was to have the code hinting and the compilation >share data structures that are resident in memory. > >- Gordon > >-----Original Message----- >From: Gordon Smith [mailto:gosm...@adobe.com] >Sent: Wednesday, September 04, 2013 10:14 AM >To: dev@flex.apache.org; Erik de Bruin >Subject: RE: [Falcon] BasicTests compiles and runs > >> FLEX_COMPILER_HOME > >I thought Alex wanted to start including Falcon as part of the SDK. If >that's the case, the Falcon JARs belong in the same lib directory as all >the other JARs, including those for the old compiler, and there wouldn't >be different "homes" for the two compilers. But there still needs to be a >switch to tell the launch scripts which JAR to use. > >> But I think getting FB to use Falcon ... > >The current FB can't use Falcon because we had to rename every package >from com.adobe to org.apache when we donated Falcon. > >- Gordon > >-----Original Message----- >From: Darrell Loverin [mailto:dlove...@adobe.com] >Sent: Wednesday, September 04, 2013 8:59 AM >To: dev@flex.apache.org; Erik de Bruin >Subject: Re: [Falcon] BasicTests compiles and runs > >I agree that an environment variable is a good way to break the flex sdk >and compiler location dependencies. I'd suggest something like >FLEX_COMPILER_HOME rather than a codename like "FALCON". > >It would be pretty quick to see if this would work: > >1. Command line - flex scripts defer to an alternate compiler's scripts >(falcon) using an environment variable. >2. Ant scripts - find new location of flexTasks.jar from environment >variable. > >But I think getting FB to use Falcon to create SWFs will more difficult >since we don't have access to the FB source. I believe that FB is not >only dependent on flex-compiler-oem.jar (oem api) but it cheats is also >dependent on mxmlc.jar. An approach I think could work is to create a >modified version of flex-compiler-oem.jar that satisfies FB and calls >falcon jars. You could think of the oem api having two major parts. >The first part is where FB calls in to convert its settings into mxmlc >configuration settings. The second part sets the configuration and then >compiles. We can probably keep the first part mostly in place. We modify >the second part to get the configuration options as a command line string >and use that to kick off the falcon build. > > >-Darrell > > >On 9/3/13 7:02 PM, "Gordon Smith" <gosm...@adobe.com> wrote: > >>Ah. Then maybe set a FALCON environment variable to true and have the >>launch scripts check it? >> >>- Gordon >> >>-----Original Message----- >>From: Alex Harui [mailto:aha...@adobe.com] >>Sent: Tuesday, September 03, 2013 3:58 PM >>To: dev@flex.apache.org; Erik de Bruin >>Subject: Re: [Falcon] BasicTests compiles and runs >> >>Yeah, but I think FB only knows how to find and launch mxmlc. And >>folks with ant scripts, etc, should be able to switch easily. >> >>The shell script and batch files could check some environment variable, >>but I think some other places load up mxmlc.jar and start talking to it. >> >>-Alex >> >>On 9/3/13 3:52 PM, "Gordon Smith" <gosm...@adobe.com> wrote: >> >>>The simplest thing is just to have dual launch scripts: fmxmlc >>>(Falcon) vs. mxmlc (old compiler), fcompc vs. compc, etc. >>> >>>- Gordon >>> >>>-----Original Message----- >>>From: Alex Harui [mailto:aha...@adobe.com] >>>Sent: Tuesday, September 03, 2013 2:45 PM >>>To: dev@flex.apache.org; Erik de Bruin >>>Subject: Re: [Falcon] BasicTests compiles and runs >>> >>>Excellent. I'll try it out now. >>> >>>Now that you've seen how we manage the source code in Git for Apache >>>Flex, what do you think we should do about packaging up Falcon to work >>>with an Apache Flex SDK? There would need to be some sort of "switch" >>>to allow someone to go back to the old MXMLC if they need to. >>> >>>Thanks, >>>-Alex >>> >>>On 9/3/13 2:37 PM, "Darrell Loverin" <dlove...@adobe.com> wrote: >>> >>>>Added a patch for the implicit import issue: >>>>https://issues.apache.org/jira/browse/FLEX-33701 >>>> >>>> >>>>-Darrell >>>> >>>>On 9/3/13 4:08 PM, "Darrell Loverin" <dlove...@adobe.com> wrote: >>>> >>>>>The below args almost work for me from eclipse. There are three >>>>>undefined classes because "flash.filters.*" is not auto included. >>>>>I'm close to have the fix for that. >>>>> >>>>>args: >>>>>/Users/dloverin/git-src/ApacheFlex/source/flex-sdk/frameworks/tests/ >>>>>b >>>>>a >>>>>sic >>>>>T >>>>>e >>>>>sts/BasicTests.mxml -debug >>>>>-library-path+=/Users/dloverin/git-src/ApacheFlex/source/flex-sdk/mu >>>>>-library-path+s >>>>>-library-path+t >>>>>-library-path+ell >>>>>a >>>>>/ >>>>>mustella.swc >>>>> >>>>>Vmargs: >>>>>-Xmx1024m -Dsun.io.useCanonCaches=false >>>>>-Dflexlib=/Users/dloverin/git-src/ApacheFlex/source/flex-sdk/framewo >>>>>r >>>>>k >>>>>s >>>>>-Dflexcompiler=/Users/dloverin/git-src/ApacheFlex/source/falcon/comp >>>>>i >>>>>l >>>>>er >>>>> >>>>> >>>>>-Darrell >>>>> >>>>> >>>>>On 9/3/13 2:49 PM, "Alex Harui" <aha...@adobe.com> wrote: >>>>> >>>>>>I'll try when I get back to my computer. >>>>>> >>>>>> >>>>>> >>>>>> >>>>>>Sent via the PANTECH Discover, an AT&T 4G LTE smartphone. >>>>>> >>>>>>Erik de Bruin <e...@ixsoftware.nl> wrote: >>>>>> >>>>>> >>>>>>I did an 'ant wipe' which seems to delete 'generated'. >>>>>> >>>>>>EdB >>>>>> >>>>>> >>>>>> >>>>>>On Tue, Sep 3, 2013 at 7:50 PM, Alex Harui <aha...@adobe.com> wrote: >>>>>>> I am using a launch configuration in eclipse. Maybe the jars in >>>>>>>generated are not updated. >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> Sent via the PANTECH Discover, an AT&T 4G LTE smartphone. >>>>>>> >>>>>>> Erik de Bruin <e...@ixsoftware.nl> wrote: >>>>>>> >>>>>>> >>>>>>> I am, last commit showing is "more fixes for BasicTests..." >>>>>>> (6b5dbbcbbda894b7c7b2998af0fe4776862769cd) >>>>>>> >>>>>>> Falcon builds fine, all tests pass. I've added the >>>>>>> "generated/dist/sdk/bin" directory to my PATH and cd to the >>>>>>> basicTests dir before running "mxmlc basicTests.mxml". >>>>>>> >>>>>>> The start of the terminal output is: >>>>>>> >>>>>>> Using Flex SDK: /Applications/Adobe Flash Builder 4.7/sdks/4.10.0 >>>>>>>(nightly) >>>>>>> Loading configuration: /Applications/Adobe Flash Builder >>>>>>> 4.7/sdks/4.10.0 (nightly)/frameworks/flex-config.xml >>>>>>> Loading configuration: >>>>>>> >>>>>>>/Users/erik/Documents/ApacheFlex/git/flex-sdk/frameworks/tests/bas >>>>>>>i >>>>>>>c >>>>>>>Tes >>>>>>>t >>>>>>>s >>>>>>>/BasicTests-config.xml >>>>>>> >>>>>>> not adding bindable variable trait for disclosureGroup in Qname: >>>>>>> Group::{PackageNs:"spark.components"} >>>>>>> not adding bindable variable trait for editor in Qname: >>>>>>> TextInput::{PackageNs:"spark.components"} >>>>>>> >>>>>>>/Users/erik/Documents/ApacheFlex/git/flex-sdk/frameworks/tests/bas >>>>>>>i >>>>>>>c >>>>>>>Tes >>>>>>>t >>>>>>>s >>>>>>>/dmv/scripts/ADGTestScript.mxml >>>>>>> Error: Internal error in syntax tree request subsystem, when >>>>>>>generating code for: >>>>>>> >>>>>>>/Users/erik/Documents/ApacheFlex/git/flex-sdk/frameworks/tests/bas >>>>>>>i >>>>>>>c >>>>>>>Tes >>>>>>>t >>>>>>>s >>>>>>>/dmv/scripts/ADGTestScript.mxml: >>>>>>> java.lang.ArrayIndexOutOfBoundsException: 0 at >>>>>>>org.apache.flex.compiler.internal.units.MXMLCompilationUnit.handle >>>>>>>S >>>>>>>y >>>>>>>nta >>>>>>>x >>>>>>>T >>>>>>>reeRequest(MXMLCompilationUnit.java:89) >>>>>>> at >>>>>>>org.apache.flex.compiler.internal.units.CompilationUnitBase.proces >>>>>>>s >>>>>>>S >>>>>>>ynt >>>>>>>a >>>>>>>x >>>>>>>TreeRequest(CompilationUnitBase.java:829) >>>>>>> at >>>>>>>org.apache.flex.compiler.internal.units.CompilationUnitBase.access >>>>>>>$ >>>>>>>1 >>>>>>>00( >>>>>>>C >>>>>>>o >>>>>>>mpilationUnitBase.java:107) >>>>>>> at >>>>>>>org.apache.flex.compiler.internal.units.CompilationUnitBase$2$1.ca >>>>>>>l >>>>>>>l >>>>>>>(Co >>>>>>>m >>>>>>>p >>>>>>>ilationUnitBase.java:259) >>>>>>> at >>>>>>>org.apache.flex.compiler.internal.units.CompilationUnitBase$2$1.ca >>>>>>>l >>>>>>>l >>>>>>>(Co >>>>>>>m >>>>>>>p >>>>>>>ilationUnitBase.java:255) >>>>>>> at >>>>>>>org.apache.flex.compiler.internal.units.requests.RequestMaker$1.ca >>>>>>>l >>>>>>>l >>>>>>>(Re >>>>>>>q >>>>>>>u >>>>>>>estMaker.java:228) >>>>>>> at >>>>>>>org.apache.flex.compiler.internal.units.requests.RequestMaker$1.ca >>>>>>>l >>>>>>>l >>>>>>>(Re >>>>>>>q >>>>>>>u >>>>>>>estMaker.java:222) >>>>>>> at >>>>>>>java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303) >>>>>>> at java.util.concurrent.FutureTask.run(FutureTask.java:138) >>>>>>> at >>>>>>>java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolE >>>>>>>x >>>>>>>e >>>>>>>cut >>>>>>>o >>>>>>>r >>>>>>>.java:895) >>>>>>> at >>>>>>>java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecu >>>>>>>t >>>>>>>o >>>>>>>r.j >>>>>>>a >>>>>>>v >>>>>>>a:918) >>>>>>> at java.lang.Thread.run(Thread.java:680) >>>>>>> >>>>>>> This "Error: Internal error..." bit is then repeated - with minor >>>>>>> variations - the kazillion times mentioned, all in a split second. >>>>>>> >>>>>>> EdB >>>>>>> >>>>>>> >>>>>>> >>>>>>> On Tue, Sep 3, 2013 at 7:39 PM, Alex Harui <aha...@adobe.com> >>>>>>>wrote: >>>>>>>> Are you on the develop branch? >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> Sent via the PANTECH Discover, an AT&T 4G LTE smartphone. >>>>>>>> >>>>>>>> Erik de Bruin <e...@ixsoftware.nl> wrote: >>>>>>>> >>>>>>>> >>>>>>>> I just tried to comile basicTests with Falcon, but even after >>>>>>>>some tweaking and rebuilding I get a kazillion errors like this >>>>>>>>one: >>>>>>>> >>>>>>>> Error: Internal error in ABC generator subsystem, when >>>>>>>>generating code >>>>>>>> for: >>>>>>>>/Users/erik/Documents/ApacheFlex/git/flex-sdk/frameworks/tests/ba >>>>>>>>s >>>>>>>>i >>>>>>>>cTe >>>>>>>>s >>>>>>>>t >>>>>>>>s/halo/scripts/ComboBoxTestScript.mxml: >>>>>>>> java.lang.NullPointerException >>>>>>>> >>>>>>>> Any ideas what I'm doing wrong? >>>>>>>> >>>>>>>> EdB >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> On Sat, Aug 31, 2013 at 9:02 AM, Alex Harui <aha...@adobe.com> >>>>>>>>wrote: >>>>>>>>> I finally got Falcon to compile and run BasicTests successfully >>>>>>>>>using the "old codegen". BasicTests is the test that runs when >>>>>>>>>you run "ant checkintests". Next task is to get it to run with >>>>>>>>>the mxml.children-as-data flag. After that, I'll be trying to >>>>>>>>>get my internal customer's app to run. >>>>>>>>> Far off is getting BasicTests to run without Flash because we'd >>>>>>>>>need to have JS versions of all of these components, but getting >>>>>>>>>this far means that Falcon's MXML handling is in decent shape. >>>>>>>>> >>>>>>>>> Thanks to Darrell Loverin and Gordon Smith for code and advice. >>>>>>>>> >>>>>>>>> -Alex >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> -- >>>>>>>> Ix Multimedia Software >>>>>>>> >>>>>>>> Jan Luykenstraat 27 >>>>>>>> 3521 VB Utrecht >>>>>>>> >>>>>>>> T. 06-51952295 >>>>>>>> I. www.ixsoftware.nl >>>>>>> >>>>>>> >>>>>>> >>>>>>> -- >>>>>>> Ix Multimedia Software >>>>>>> >>>>>>> Jan Luykenstraat 27 >>>>>>> 3521 VB Utrecht >>>>>>> >>>>>>> T. 06-51952295 >>>>>>> I. www.ixsoftware.nl >>>>>> >>>>>> >>>>>> >>>>>>-- >>>>>>Ix Multimedia Software >>>>>> >>>>>>Jan Luykenstraat 27 >>>>>>3521 VB Utrecht >>>>>> >>>>>>T. 06-51952295 >>>>>>I. www.ixsoftware.nl >>>>> >>>> >>> >> >
