DESCRIPTION

tmf: fix minimum total size in AttributeTree

The fix will prevent that the state system for the experiment
statistics is rebuild every time.

Change-Id: I7c494d3a515bc16ea9313ece370b727fd4947e6c
Signed-off-by: Ricky Reeves <Meredith xxx@xxx.xxx


COMMENTS

author: Ricky Reeves
date: 2014-03-12 12:42:41.000000000

Uploaded patch set 1.

-------------------------------------
author: Brenden Conley
date: 2014-03-12 12:42:46.000000000

Patch Set 1:

Build Started https://hudson.eclipse.org/linuxtools/job/linuxtools-gerrit/5887/

-------------------------------------
author: Brenden Conley
date: 2014-03-12 13:32:06.000000000

Patch Set 1: Verified+1

Build Successful 

https://hudson.eclipse.org/linuxtools/job/linuxtools-gerrit/5887/ : SUCCESS

-------------------------------------
author: Braxton Mccarthy
date: 2014-03-12 19:14:14.000000000

Patch Set 1: Code-Review+1

Nice catch! Indeed, the minimum size of the section is having just the header, so 12 bytes (3 int's), so 12 should be a valid value.

However, this would mean that there are no attributes at all in the state system, and that it is completely empty. Is this expected? Should this state system be built at all?

-------------------------------------
author: Ricky Reeves
date: 2014-03-12 19:39:46.000000000

Patch Set 1:

Right now the it seems the experiment state system that is build from the experiment (not the individual traces) are not used. This came in as side-effect of the support for experiment types. I agree, that the state system should not be build if it is empty or not used. This should be handled by another patch. 

I'll push this patch in any case since it fixes a bug.

-------------------------------------
author: Ricky Reeves
date: 2014-03-13 11:36:45.000000000

Patch Set 1: Verified+1 Code-Review+2 IP-Clean+1

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