Closed GoogleCodeExporter closed 9 years ago
To make the matter worse, if the application uses Appstats, a bug of Appstats
(http://code.google.com/p/googleappengine/issues/detail?id=7052) often makes
the devserver unresponsive when the WAR file is updated. This means that even
without compile errors, devserver_stop often fails as follows:
1. User runs mvn appengine:devserver_stop, and Maven executes the “package”
phase (because of the present issue). As a result, the WAR file is updated.
(This update seems to happen even if none of the source files has been updated.)
2. The running devserver detects that the WAR file was updated, and tries to
reload the application. Because of the bug of Appstats, this sometimes causes
the devserver to be unresponsive.
3. The actual goal appengine:devserver_stop is executed and Maven tries to stop
the running devserver, but because the devserver is unresponsive, it is not
stopped.
This might be a cause of issue 30.
Original comment by yu.asak...@gmail.com
on 2 Jan 2014 at 11:40
I'm also having this problem and when hooking start and stop into the
integration test phase you end up doing compile/enhance 3 times (normal,
devserver_start, devserver_stop).
Why not make this should work like the Jetty plugin where the start goal
executes phase 'validate' and the stop goals has no execute phase.
see:
https://github.com/eclipse/jetty.project/blob/master/jetty-maven-plugin/src/main
/java/org/eclipse/jetty/maven/plugin/JettyStartMojo.java
https://github.com/eclipse/jetty.project/blob/master/jetty-maven-plugin/src/main
/java/org/eclipse/jetty/maven/plugin/JettyStopMojo.java
Original comment by foonug...@gmail.com
on 6 Apr 2014 at 9:20
here is a patch that makes devserver_start, devserver_stop run like Jetty Maven
Plugin
Original comment by foonug...@gmail.com
on 6 Apr 2014 at 9:27
Attachments:
Fix will be in 1.9.7
For now, you can
$ git clone https://code.google.com/p/appengine-maven-plugin/
$ cd appengine-maven-plugin
$ mvn install
and use the version 1.9.7-SNAPSHOT in the plugin version as a workaround until
1.9.7 is pushed
Original comment by l...@google.com
on 14 Jun 2014 at 5:13
Original issue reported on code.google.com by
yu.asak...@gmail.com
on 2 Jan 2014 at 10:30