DESCRIPTION

TMF: Port the LTTng Kernel state system to an analysis module

And update the views accordingly.

Also, for the sake of uniformisation, with the analysis module state system,
old state systems should now use this mechanism.  The old get- and register-
state system methods are deprecated.

Change-Id: I90975b5772db24a784d99d2d0232f6f76a8e5e03
Signed-off-by: Aryana Jones xxx@xxx.xxx


COMMENTS

author: Annalise Jimenez
date: 2013-11-14 19:53:41.000000000

Uploaded patch set 2.

-------------------------------------
author: Brenden Conley
date: 2013-11-14 19:53:52.000000000

Patch Set 2:

Build Started https://hudson.eclipse.org/linuxtools/job/linuxtools-gerrit/3928/

-------------------------------------
author: Brenden Conley
date: 2013-11-14 19:53:56.000000000

Patch Set 2: Verified-1

Build Failed 

https://hudson.eclipse.org/linuxtools/job/linuxtools-gerrit/3928/ : FAILURE

-------------------------------------
author: Brenden Conley
date: 2013-11-14 21:38:24.000000000

Patch Set 2: -Verified

Build Started https://hudson.eclipse.org/linuxtools/job/linuxtools-gerrit/3938/

-------------------------------------
author: Braxton Mccarthy
date: 2013-11-14 23:43:02.000000000

Patch Set 2: Code-Review-1

(5 comments)

Some quick comments.

Line:13, /COMMIT_MSG -> Isn't this (the second paragraph) more related to the previous patch?

Line:33, lttng/org.eclipse.linuxtools.lttng2.kernel.ui/src/org/eclipse/linuxtools/lttng2/kernel/ui/analysis/LttngKernelAnalysisModule.java -> Is this still needed? (seems unused)

Line:36, lttng/org.eclipse.linuxtools.lttng2.kernel.ui/src/org/eclipse/linuxtools/lttng2/kernel/ui/analysis/LttngKernelAnalysisModule.java -> Should we make analysis module IDs finish with ".analysis" or something? Just to be sure it doesn't clash with views or tracetypes. I'm asking since I'll do the same with the statistics patches.

Line:49, lttng/org.eclipse.linuxtools.lttng2.kernel.ui/src/org/eclipse/linuxtools/lttng2/kernel/ui/analysis/LttngKernelAnalysisModule.java -> Since super.fTrace is not final and has a setter, to be extra-safe it should be extracted in a local variable first (which you can then instanceof and cast to the right type). That's probably a Sonar warning too.

Line:16, lttng/org.eclipse.linuxtools.lttng2.kernel.core/META-INF/MANIFEST.MF -> We could also export the package publicly at that point. I don't mind either way.

-------------------------------------
author: Annalise Jimenez
date: 2013-11-15 14:11:35.000000000

Patch Set 2:

(3 comments)

Line:33, lttng/org.eclipse.linuxtools.lttng2.kernel.ui/src/org/eclipse/linuxtools/lttng2/kernel/ui/analysis/LttngKernelAnalysisModule.java -> Oups... got lost between a few versions of this patch ;-)

Line:36, lttng/org.eclipse.linuxtools.lttng2.kernel.ui/src/org/eclipse/linuxtools/lttng2/kernel/ui/analysis/LttngKernelAnalysisModule.java -> Ok, let's do that.  We're inventing the convention anyway

Line:16, lttng/org.eclipse.linuxtools.lttng2.kernel.core/META-INF/MANIFEST.MF -> Anyway, it may become an XML file at some point soon... Let's leave it there

-------------------------------------
author: Brenden Conley
date: 2013-11-15 17:03:21.000000000

Patch Set 2: Verified-1

Build Failed 

https://hudson.eclipse.org/linuxtools/job/linuxtools-gerrit/3938/ : ABORTED

-------------------------------------
