Log debug info for #1897

[chanseokoh]

For #1897. SWTBotTreeItem.toString() will do the job.

[codecov-io]

Codecov Report

Merging #1975 into master will increase coverage by 0.12%.
The diff coverage is n/a.

@@             Coverage Diff              @@
##             master    #1975      +/-   ##
============================================
+ Coverage     65.27%   65.39%   +0.12%     
- Complexity     1804     1817      +13     
============================================
  Files           320      320              
  Lines         12382    12428      +46     
  Branches       1110     1118       +8     
============================================
+ Hits           8082     8127      +45     
- Misses         3891     3892       +1     
  Partials        409      409

Impacted Files	Coverage Δ	Complexity Δ
...roject/maven/MavenAppEngineStandardWizardPage.java	80.7% <0%> (+0.61%)	37% <0%> (+4%)	⬆️
...loud/tools/eclipse/swtbot/SwtBotTreeUtilities.java	18.75% <0%> (+1.04%)	5% <0%> (ø)	⬇️
.../appengine/localserver/launching/LaunchHelper.java	46.91% <0%> (+1.23%)	16% <0%> (+1%)	⬆️
...e/appengine/newproject/AppEngineProjectWizard.java	64.4% <0%> (+4.79%)	7% <0%> (+3%)	⬆️
.../newproject/maven/MavenArchetypeProjectWizard.java	75.3% <0%> (+5.66%)	8% <0%> (+4%)	⬆️
...oject/standard/AppEngineStandardProjectWizard.java	74.19% <0%> (+7.52%)	6% <0%> (+1%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 3734b76...1436b27. Read the comment docs.

[chanseokoh]

This is funny. The Kokoro failure of this PR (the initial commit 1436b27 above) triggered #1897 (very opportunistic), but there was no debug info logged by this PR. This doesn't make sense, so I'm a bit perplexed.

[chanseokoh]

Ah, okay, the log was quite a bit more above the assertion failure message:

INFO: submitting command: C:\Program Files\Java\jdk1.8.0_131\bin\java.exe -Xdebug -Xrunjdwp:transport=dt_socket,server=n,suspend=y,quiet=y,address=52119 -javaagent:C:\Program Files (x86)\Google\Cloud SDK\google-cloud-sdk\platform\google_appengine\google\appengine\tools\java\lib\agent\appengine-agent.jar -Dappengine.sdk.root=C:\Program Files (x86)\Google\Cloud SDK\google-cloud-sdk\platform\google_appengine\google\appengine\tools\java -cp C:\Program Files (x86)\Google\Cloud SDK\google-cloud-sdk\platform\google_appengine\google\appengine\tools\java\lib\appengine-tools-api.jar com.google.appengine.tools.development.DevAppServerMain --address=localhost --port=8080 --allow_remote_shutdown --disable_update_check T:\src\github\google-cloud-eclipse\plugins\com.google.cloud.tools.eclipse.integration.appengine\target\work\data\.metadata\.plugins\org.eclipse.wst.server.core\tmp0\testapp
    [GitHub #1897 debug info] TreeItem with text {}
12:45:19.861 [main] WARN  o.e.s.s.finder.widgets.SWTBotButton - Widget is not enabled: (of type 'Button' and with mnemonic 'OK' and with style 'SWT.PUSH')

There certainly exists one tree item, but its text is empty. Maybe a race condition: we wait until there becomes one tree item, but we check its text too early before the UI thread sets a proper text.

[briandealwis]

There may have been an execution error with launching devappserver. As part of trying to diagnose the very strange test problems I encountered in #1529, I had added some support to the tests to dump the console output from launching the devappserver.

(I never did get to the bottom of #1529)

[chanseokoh]

@briandealwis That's very interesting.

Before that, what seems weird though is that the test calls SwtBotTreeUtilities.waitUntilTreeHasText(bot, allItems[0]), which is supposed to wait until allItems[0] has a non-empty text. However, we end up seeing an empty text, so I think the method is either not functioning properly or timing out.

[briandealwis]

or timing out

As might happen if the launched JVM never actually ran for some reason…

[briandealwis]

I wonder if ProjectUtils#waitForProjects() shouldn't instead forcibly trigger a rebuild via IWorkspace#build() rather than wait for the auto build job.

[chanseokoh]

Hopefully #1897 is fixed by #2047. Closing.