Okay, the best I can tell so fa is that release step 13 appears to be failing for the same reason that regular nightly builds of royale-compiler and royale-asjs_jsonly have been failing recently. They're all failing at the same ASDateTests.
Some interesting console output that mentions the current environment variables in a few different scenarios: Last successful build of royale-compiler (815): [junit] environment property - FLEX_HOME = null [junit] environment property - PLAYERGLOBAL_HOME = null [junit] environment property - PLAYERGLOBAL_VERSION = 11.1 [junit] environment property - TLF_HOME = null [junit] environment property - AIR_HOME = null [junit] environment property - FLASHPLAYER_DEBUGGER = null [junit] environment property - ASJS_HOME = C:\jenkins\workspace\royale-asjs [junit] environment property - GOOG_HOME = null Current failing build of royale-compiler (886): [junit] environment property - FLEX_HOME = null [junit] environment property - PLAYERGLOBAL_HOME = null [junit] environment property - PLAYERGLOBAL_VERSION = 11.1 [junit] environment property - TLF_HOME = null [junit] environment property - AIR_HOME = null [junit] environment property - FLASHPLAYER_DEBUGGER = C:\adobe\flash\11.7\flashplayerdebugger.exe [junit] environment property - ASJS_HOME = C:\jenkins\workspace\royale-asjs_jsonly [junit] environment property - GOOG_HOME = null Release step 13 [junit] environment property - FLEX_HOME = null [junit] environment property - PLAYERGLOBAL_HOME = null [junit] environment property - PLAYERGLOBAL_VERSION = 11.1 [junit] environment property - TLF_HOME = null [junit] environment property - AIR_HOME = null [junit] environment property - FLASHPLAYER_DEBUGGER = C:\adobe\flash\11.7\flashplayerdebugger.exe [junit] environment property - ASJS_HOME = C:\jenkins\workspace\Royale_Release_Step_013\artifacts\royale-asjs [junit] environment property - GOOG_HOME = null The last successful build of royale-compiler did not have FLASHPLAYER_DEBUGGER defined, so it was not running these tests in Flash Player (only compiling them). The current failing build has FLASHPLAYER_DEBUGGER defined (and so does release step 13), so it is trying to run them, and failing. Now, I'm not saying that FLASHPLAYER_DEBUGGER needs to be cleared to fix this. It might be okay for FLASHPLAYER_DEBUGGER to be defined, and maybe we should actually prefer it to be defined so that the tests get run on the CI server (because it seems like they were only compiled previously, without running). However, if FLASHPLAYER_DEBUGGER is defined, then PLAYERGLOBAL_HOME also needs to be defined. It needs to be able to build a valid SWF (with Sprite as the base class) to run successfully in Flash Player, and that's why the test run fails. I am not super familiar with how the CI server is configured. However, I would guess that playerglobal.swc exists somewhere in C:\adobe\, since that's where flashplayerdebugger.exe can be found. I hope this helps! -- Josh Tynjala Bowler Hat LLC <https://bowlerhat.dev> On Mon, Jun 28, 2021 at 10:03 AM Josh Tynjala <joshtynj...@bowlerhat.dev> wrote: > ASDateTests extends ASFeatureTestsBase. In ASFeatureTestsBase, you can see > the getAS() method where the code for the test is generated. If > playerglobal.swc is available, it creates a main class for Flash that > extends Sprite. If playerglobal.swc is not available, the main class does > not extend Sprite. > > With this in mind, it looks like the tests can't find playerglobal.swc, > for some reason. I can reproduce this issue locally on my computer by > running the tests without the PLAYERGLOBAL_HOME variable set (or by setting > it to a bad location). I can also reproduce if PLAYERGLOBAL_VERSION is set > to the wrong version. > > I'll try to help you investigate some more, but I thought I'd share what I > found so far. > > -- > Josh Tynjala > Bowler Hat LLC <https://bowlerhat.dev> > > > On Mon, Jun 28, 2021 at 5:12 AM Yishay Weiss <yishayj...@hotmail.com> > wrote: > >> The standalone flash player pops up messages such as >> >> >> TypeError: Error #2023: Class ASDateTests4516718174320608253$ must >> inherit from Sprite to link to the root. >> >> If I click ‘continue’ then the tests seem to pass, but I can’t do that in >> a build script. >> >> You can see log [1] of tests failure here >> >> [1] Royale_Release_Step_013 #12 Console [Jenkins] (azure.com)< >> http://apacheroyaleci2.westus2.cloudapp.azure.com:8080/job/Royale_Release_Step_013/12/console >> > >> >> Any thoughts? >> >
