Bug 273308

Summary:

[perfs] Save very slow because DeltaProcessor#resourceChanged(..) resolves classpath

Product:

[Eclipse Project] JDT

Reporter:

Markus Keller <markus.kell.r>

Component:

Core

Assignee:

Jerome Lanneluc <jerome_lanneluc>

Status:

VERIFIED FIXED

QA Contact:

Severity:

major

Priority:

CC:

daniel_megert, frederic_fusier, greensopinion, leo.dos.santos, robert.elves

Version:

3.5

Keywords:

performance

Target Milestone:

3.5 M7

Hardware:

OS:

Windows XP

Whiteboard:

Bug Depends on:

Bug Blocks:

270784, 270824

Attachments:

Description	Flags
trace	none
Proposed fix	none

Description Markus Keller

2009-04-22 13:32:25 EDT

Created attachment 132815 [details]
trace

I20090421-0930, I guess this is a consequence of the fix for bug 271102

With this build, normal Save actions in a Java Editor often take a long time. I took a trace (sampling) and saw that most of the 4 seconds were spent in DeltaProcessor#resourceChanged(IResourceChangeEvent), split up in two invocations, one from the CleanUp (Save Action), and the other from the actual save operation.

Comment 1 Frederic Fusier

2009-04-23 03:42:30 EDT

This regression is also seen in our JDT/Core model performance test: testCloseProject() (see bug 270784).

Comment 2 Jerome Lanneluc

2009-04-23 06:46:58 EDT

Created attachment 132915 [details]
Proposed fix

Comment 3 Frederic Fusier

2009-04-23 07:07:12 EDT

(In reply to comment #2)
> Created an attachment (id=132915) [details]
> Proposed fix
> 
Here are the numbers I get for the Elapsed Process:
v_949 + bug 270784 fix: 1.08s  (95% in [997ms, 1.16s])
v_950:                  7.14s  (95% in [6.69s, 7.59s]) 
v_950 + proposed fix:   0.992s (95% in [885ms, 1.1s])

Comment 4 Frederic Fusier

2009-04-23 09:50:52 EDT

(In reply to comment #3)
> (In reply to comment #2)
> > Created an attachment (id=132915) [details] [details]
> > Proposed fix
> > 
> Here are the numbers I get for the Elapsed Process:
> v_949 + bug 270784 fix: 1.08s  (95% in [997ms, 1.16s])
> v_950:                  7.14s  (95% in [6.69s, 7.59s]) 
> v_950 + proposed fix:   0.992s (95% in [885ms, 1.1s])
> 
As per Jerome's remark, I forgot to say that these numbers are times for the JDT/Core perf test testCloseProjects()...

Comment 5 Jerome Lanneluc

2009-04-23 11:31:37 EDT

Fix released for 3.5M7

Comment 6 Kent Johnson

2009-04-28 12:36:14 EDT

Verified for 3.5M7 using Frederic's performance numbers

Comment 7 David Green

2009-04-29 16:42:36 EDT

*** Bug 274232 has been marked as a duplicate of this bug. ***