Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.
Bug 322989 - MoDisco Browser : resources loading depth specific to one metamodel
Summary: MoDisco Browser : resources loading depth specific to one metamodel
Status: CLOSED FIXED
Alias: None
Product: MoDisco
Classification: Modeling
Component: Infrastructure (show other bugs)
Version: 0.9.0   Edit
Hardware: PC Windows XP
: P3 normal (vote)
Target Milestone: 0.9.0 M2   Edit
Assignee: Nicolas Bros CLA
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-08-18 04:21 EDT by Fabien Giquel CLA
Modified: 2010-10-18 08:37 EDT (History)
3 users (show)

See Also:
nicolas.bros: indigo+

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Fabien Giquel CLA 2010-08-18 04:21:51 EDT 
For some models loading, it is a critical need to dispatch model elements on
many XMI resources from a memory usage point of view. In such cases, MoDisco
browser resources loading depth need to be set to a convenient value.

It would be convenient to have the following evolution
- a new extension point which allows to specify a default "MoDisco resources loading depth" for one given metamodel (using EPackage uri).
- a checkbox in "Window->preferences->MoDisco Model Browser"  for specifying that the global setting "loading depth" should overrides specific metamodel "loading depth" (unchecked by default)
- The MoDisco browser takes in account theses informations to decide which level of resources to load at opening.
Comment 1 Nicolas Bros CLA 2010-09-10 10:45:03 EDT 
modisco-nightly took 5 hr 35 min on hudson-slave1:
BUCKMINSTER SETPREF took : 6s
BUCKMINSTER IMPORT took : 19302s (5 hours and 22 minutes!)
BUCKMINSTER RESOLVE took : 76s
BUCKMINSTER BUILD took : 74s
BUCKMINSTER JUNIT took : 591s
BUCKMINSTER PERFORM 'site.p2' took : 85s

Whereas the same job took 48 min on master on build.eclipse.org:
BUCKMINSTER SETPREF took : 25s
BUCKMINSTER IMPORT took : 1213s
BUCKMINSTER RESOLVE took : 90s
BUCKMINSTER BUILD took : 164s
BUCKMINSTER JUNIT took : 1097s
BUCKMINSTER PERFORM 'site.p2' took : 234s

So, this is definitely the import phase that is slow, as it runs 16 times more slowly on hudson.eclipse.org than on build.eclipse.org.
The other major steps in the build are actually *faster* on hudson.eclipse.org.
Comment 2 Nicolas Bros CLA 2010-09-10 10:47:47 EDT 
(In reply to comment #1)
Oops, wrong bug!
Comment 3 Nicolas Bros CLA 2010-09-13 08:38:11 EDT 
Done in revision 3021 (extension point "loadingDepth").
Comment 4 Fabien Giquel CLA 2010-09-16 14:14:27 EDT 
It works fine. Thanks.
Comment 5 Hugo Bruneliere CLA 2010-10-05 04:49:16 EDT 
Bug fixed.