Unable to run gae:debug

[Moritz]

Hi,

I'm unable to run the gae:debug goal. Anyone having more success?
There is already an issue: http://code.google.com/p/maven-gae-plugin/issues/detail?id=88

My setup is:

Plugin version: 0.7.1
GAE version: 1.3.7 (installed using gae:unpack)
IDE / debugger: Eclipse
Maven version: 2.2.0
Output:
...

INFO: The server is running at http://192.168.178.24:8080/
[INFO]
------------------------------------------------------------------------
[INFO] BUILD SUCCESSFUL
[INFO]
------------------------------------------------------------------------
[INFO] Total time: 49 seconds
[INFO] Finished at: Mon Sep 06 21:01:50 CEST 2010
[INFO] Final Memory: 23M/81M
[INFO]
------------------------------------------------------------------------
Destroying 1 processes
Destroying process..
Destroyed 1 processes


(yes, I'm binding the Eclipse debugger to 192.168.178.24)

It worked with previous versions.

Cheers,
Moritz

[Rodolfo Hansen]

Hi,

I just ran gae:debug on my machine. the process is forking, but the
debug session is accesible.

[David C. Hicks]

Rodolfo,

I was a bit confused by your statement here.  It was hard to know if you meant that it's working or not.  So, I figured I'd try it myself.  The plugin starts up the Debug session just fine (plugin version 0.7.1), but as soon as I try to connect to it using Eclipse Helios, it terminates.  What I notice is that the debug session seems to start and listen on port 8000, but only when I try to connect with Eclipse does it attempt to listen on port 8080 for regular HTTP activity.  But, then it goes on to terminate as if I had killed the whole session.

In case it's helpful, here is the output from my terminal:

[INFO] [gae:debug {execution: default-cli}]
Listening for transport dt_socket at address: 8000
Sep 7, 2010 12:37:04 AM com.google.apphosting.utils.jetty.JettyLogger info
INFO: Logging to JettyLogger(null) via com.google.apphosting.utils.jetty.JettyLogger
Sep 7, 2010 12:37:04 AM com.google.apphosting.utils.config.AppEngineWebXmlReader readAppEngineWebXml
INFO: Successfully processed /home/dhicks/workspace/amedia-struts/target/amedia-struts-1.0-SNAPSHOT/WEB-INF/appengine-web.xml
Sep 7, 2010 12:37:04 AM com.google.apphosting.utils.config.AbstractConfigXmlReader readConfigXml
INFO: Successfully processed /home/dhicks/workspace/amedia-struts/target/amedia-struts-1.0-SNAPSHOT/WEB-INF/web.xml
log4j:WARN No appenders could be found for logger (com.opensymphony.xwork2.config.providers.XmlConfigurationProvider).
log4j:WARN Please initialize the log4j system properly.
Sep 6, 2010 8:37:06 PM com.google.appengine.tools.development.DevAppServerImpl start
INFO: The server is running at http://localhost:8080/

[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESSFUL
[INFO] ------------------------------------------------------------------------

[INFO] Total time: 45 seconds
[INFO] Finished at: Mon Sep 06 20:37:06 EDT 2010
[INFO] Final Memory: 23M/302M

[INFO] ------------------------------------------------------------------------
Destroying 1 processes
Destroying process..
Destroyed 1 processes

[David C. Hicks]

I must be doing something wrong.  I've tried several versions of the plugin, and they're all behaving the same way - terminating when I try to connect.  It does this for me with versions 0.7.1, 0.7.0, and 0.6.0.

Dave

[Rodolfo Hansen]

It should not be terminated.

It has forked; check your eclipse to see if it is actually disconected.

This is the difference between gae:start and gae:run where the former
actually forks a separate process to run the goal.

[David C. Hicks]

OK. I didn't quite understand that before.

However, it does appear to be terminating.
Eclipse does not end up connected, and the app container on 8080 is not
there, either.

As soon as I attempt to connect the debugger, I get the output that
follows "Listening for transport dt_socket at address: 8000" and
everything appears to terminate.

Is it possible I just have something not configured properly? My
project is based on the JSP archetype from 0.6.0. I never actually
tried to run gae:debug previously, because I had been using GPE.
However, I was unhappy with GPE and removed it from the project.

gae:run works just fine for me.

Dave

[Moritz]

I can confirm everything David is writing here. Except that it was
working with previous versions of the plugin or of GAE itself.
Rodolfo, so you are able to use the debugger? Can you maybe tell us
the debugger settings you are using?

Cheers,
Moritz

> >>          INFO: The server is running athttp://localhost:8080/

[David C. Hicks]

Has anyone got a solution for this problem, yet? It's going to be a
real issue for a project I'm starting.

I guess, for now, I'll have to configure the Google Eclipse Plugin and
do my debugging using that web container.

[Taylor Leese]

David,

Rodolfo posted a workaround in Issue 88 using, "mvn gae:debug -
Dgae.wait=true". However, this really shouldn't be anything that holds
you up because you can debug using gae:run with the relevant jvmFlag
settings. I posted these in Issue 88 as well.

Thank you,
Taylor

[Rodolfo Hansen]

I'm sorry conversion continued on the issue.

Basically:

mvn gae:deploy -Dgae.wait=true works as expected

I am waiting for someone who has windows to send me the output of

version:   0.7.2-SNAPSHOT
with:

mvn -X gae:deploy

or

mvn -X gae:start

since they both suffer from the same issue in windows.

--
Rodolfo Hansen
CTO, KindleIT Software Development
Email: rha...@kitsd.com
Mobile: +1 (809) 860-6669

[David C. Hicks]

No problem.  I guess I should get in on the issue database mailing list to keep track.
Thanks for the workaround!  I'll give it a shot.

Dave

[David C. Hicks]

I had an issue come up with my M2 repository and had to start fresh.
An existing project using plugin 0.7.1 is now having build failures. I
ran the gae:unpack goal to pull the AppEngine artifacts into my
repository, but something is clearly amiss.

The first change I had to make was to look for
datanucleus-appengine-1.0.7 jar. The original project created by the
archetype wanted datanucleus-appengine-1.0.7-final.jar. No big deal
there. However, now, the project can't find the datanucleus-jpa jar
it's looking for. In my repository, I see
"datanucleus-jpa-1.1.5.jar.lastUpdated".

Can someone tell me how I should alter either my pom.xml or how I'm
getting artifacts into the repository in order for this to work?

Thanks,
Dave

[Rodolfo Hansen]

Hmm, when did you create your project? because the 0.7.1 archetypes have been updated..

ORM related packages have been moved by google into an ORM subfolder

checkout which project you are looking for:

http://repo2.maven.org/maven2/com/google/appengine/

and

http://repo2.maven.org/maven2/com/google/appengine/orm

[David C. Hicks]

The project was originally created with the 0.6.0 plugin.
I'll create a 0.7.1 project and use it as a guide to update my ORM artifact versions.

I'm seeing a similar issue with the geronimo-jpa_3.0_spec.  The jar has the "lastUpdated" suffix on the file name.

Thanks for the quick feedback!
Dave

[Rodolfo]

Your welcome,

I don't if you've used sites like this one, but just in case:
http://mvnrepository.com

> > Email: rhan...@kitsd.com <mailto:rhan...@kitsd.com>
> > Mobile: +1 (809) 860-6669

[David C. Hicks]

Yes, I've used search sites for Maven repositories before. I was just
surprised when the artifacts that had been in my local repository
suddenly couldn't be found.

It also looks like the new JSP archetype for 0.7.1 doesn't include any
of the JDO dependencies. So, I'll have to figure those out on my own.
At least, now I know they should all be in the "orm" folder. It would
be nice if Google would quit moving things around with each new release,
wouldn't it? :-)

Dave

[Rodolfo Hansen]

I'll have to side with Linus on changes... :p

Also, before Google published to maven, I had to choose the locations... :/

For most people (who don't delete their repo) its going to be a smooth transition.

I'll write something in the doc for the artifact locations.

[David C. Hicks]

I see.  Figures that I'd go and throw my own wrench into the works.  :-)
Thanks for the help figuring it out.