Right now, the flex_sdk Ant CI build is failing because a third-party download is unavailable. Once that comes back, flex_falcon will be built again and hopefully it will all work. Probably will take several hours to shake out so don't spend any time thinking about it until after you see the flex_sdk build pass again.
-Alex On 9/27/16, 5:08 PM, "Greg Dove" <greg.d...@gmail.com> wrote: >Alex, can you please let me know if this is related to something I did? >Is there something I need to fix? >I added some fixes this morning and checked that the maven build was >working locally. >But I don't know what I am doing that is different to the CI build. > > > >On Wed, Sep 28, 2016 at 12:40 PM, Alex Harui <aha...@adobe.com> wrote: > >> The CI Ant build now fails as well. I just saw the email on the list. >> The CI server keeps getting stuck on flex_sdk_test so Falcon hadn't run >>in >> a couple of days. The CI falcon build also runs the sdk.dependent.tests >> and flexjs.dependent.tests targets that don't run by default. >> >> -Alex >> >> On 9/27/16, 1:58 PM, "Christofer Dutz" <christofer.d...@c-ware.de> >>wrote: >> >> >Hi Greg, >> > >> > >> >Well investigating this had me worry a little ... not the problem >>itself, >> >but the Ant build seems to be reporting that all is fine and I could >> >confirm the method being listed as successful in the build report. >> > >> > >> >But even if I run the test in IntelliJ without Maven, the test fails >>and >> >from having a look at it, it should fail. So the question remains: why >> >didn't the Ant build fail? >> > >> > >> >Chris >> > >> >________________________________ >> >Von: Greg Dove <greg.d...@gmail.com> >> >Gesendet: Dienstag, 27. September 2016 22:46:27 >> >An: dev@flex.apache.org >> >Betreff: Re: [FlexJS] Problem with some test cases (@export was changed >> >to @expose)? >> > >> >I wonder if I had something cached somewhere that needed an extra >>'clean' >> >- >> >I am still learning this, I guess. >> > >> >On Wed, Sep 28, 2016 at 9:43 AM, Greg Dove <greg.d...@gmail.com> wrote: >> > >> >> I am working through these now with the maven build, I found a few >>areas >> >> that needed attention - sorry. >> >> Yes I had expected this to work the same across the two builds as >>well. >> >> >> >> >> >> >> >> On Wed, Sep 28, 2016 at 9:42 AM, Christofer Dutz < >> >> christofer.d...@c-ware.de> wrote: >> >> >> >>> Ok ... so I tracked down the problem to >> >>> >> >>> JSGoogDocEmitter >> >>> >> >>> there in line 344 in emitPublic, you seem to have changed the >>output. >> >>> >> >>> Should I adjust the testcase? I don't quite understand why the test >> >>> doesn't break in the Ant build, because the output should be wrong >> >>>there >> >>> too ... >> >>> >> >>> >> >>> Chris >> >>> >> >>> >> >>> ________________________________ >> >>> Von: Greg Dove <greg.d...@gmail.com> >> >>> Gesendet: Dienstag, 27. September 2016 19:53:45 >> >>> An: dev@flex.apache.org >> >>> Betreff: Re: [FlexJS] Problem with some test cases (@export was >>changed >> >>> to @expose)? >> >>> >> >>> This was me, and was needed for reflection support into static >>members. >> >>> >> >>> @expose is supposed to be deprecated, but currently seems to be the >> >>>only >> >>> option that works for statics. Josh discovered this a while back I >> >>>think, >> >>> with static accessors. >> >>> >> >>> I will check this test now, I thought I had them all updated. >> >>> >> >>> cheers >> >>> Greg >> >>> >> >>> >> >>> >> >>> On Wed, Sep 28, 2016 at 3:46 AM, Christofer Dutz < >> >>> christofer.d...@c-ware.de> >> >>> wrote: >> >>> >> >>> > Hi, >> >>> > >> >>> > >> >>> > it seems some recent changes broke things in the test-suite. I am >> >>> getting >> >>> > failures in TestFlexJSClass. >> >>> > >> >>> > For: >> >>> > >> >>> > >> >>> > org.apache.flex.compiler.internal.codegen.js.flexjs. >> >>> > TestFlexJSClass#testConstants >> >>> > >> >>> > >> >>> > and >> >>> > >> >>> > >> >>> > org.apache.flex.compiler.internal.codegen.js.flexjs. >> >>> > TestFlexJSClass#testMethods >> >>> > >> >>> > >> >>> > it seems @export was changed to @expose ... would be cool if >>someone >> >>>who >> >>> > knows what happened here could fix that. >> >>> > >> >>> > >> >>> > Chris >> >>> > >> >>> >> >> >> >> >> >>
