Bug 61831 - Full build happens on every start of Eclipse
Summary: Full build happens on every start of Eclipse
Status: VERIFIED FIXED
Alias: None
Product: JDT
Classification: Eclipse Project
Component: Core (show other bugs)
Version: 3.0   Edit
Hardware: PC Windows XP
: P3 normal (vote)
Target Milestone: 3.0 M9   Edit
Assignee: Jerome Lanneluc CLA
QA Contact:
URL:
Whiteboard:
Keywords:
: 62176 (view as bug list)
Depends on:
Blocks:
 
Reported: 2004-05-11 19:07 EDT by Dirk Baeumer CLA
Modified: 2004-05-19 06:03 EDT (History)
1 user (show)

See Also:

Attachments
The trace (317.28 KB, text/plain)
2004-05-11 19:08 EDT, Dirk Baeumer CLA 		no flags Details
A trace with classpath resolution debugging on (432.65 KB, text/plain)
2004-05-14 11:04 EDT, Dirk Baeumer CLA 		no flags Details
View All 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 Dirk Baeumer CLA 2004-05-11 19:07:49 EDT 
I20040506 + Plug-in export (contains newwer JDT/Core).

Whenever I start Eclipse I see the following behaviour

- some sort of incremental build kicks in which takes a couple of seconds
- after that a full build kicks in.

Attached the beginning of a trace (contains the FULL BUILD request). The whole 
trace is really large.
Comment 1 Dirk Baeumer CLA 2004-05-11 19:08:24 EDT 
Created attachment 10523 [details]
The trace
Comment 2 Philipe Mulet CLA 2004-05-12 06:23:19 EDT 
During incremental build, the Java builder is correctly reacting to what looks 
like classpath changes, and triggers a full build instead. 

The classpath used to build in previous session is different from the one 
discovered after restart. 

Suspecting PDE containers to answer differently... 

How did you construct your workspace ?
Comment 3 Dirk Baeumer CLA 2004-05-12 06:51:37 EDT 
I have this workspace for almost 6 months now. It consist of binary plug-ins 
with source from CVS. The source plug-ins use PDE libarary containers. I 
import the binary plug-ins via PDE. The plug-ins get copied into my workspace.
Comment 4 Philipe Mulet CLA 2004-05-12 09:15:23 EDT 
I have a similar setup, and am not observing this behavior.
Comment 5 Philipe Mulet CLA 2004-05-14 10:43:38 EDT 
Also see bug 62120 for intriguing trace. It seems that JRE container is not 
remembering its old value when rebound.

Wondering if this wouldn't be a consequence of our added support for 
flattening classpath initialization. Instead of reusing empty classpath, it 
should rather consider previous session classpaths (using persisted container 
values).
Comment 6 Dirk Baeumer CLA 2004-05-14 11:04:55 EDT 
Created attachment 10647 [details]
A trace with classpath resolution debugging on
Comment 7 Jerome Lanneluc CLA 2004-05-17 07:39:19 EDT 
Fixed all internal JDT Core calls to use getResolveClasspath(..., false/*don't 
returnResolutionInProgress*/) so that an empty classpath is never returned by 
internal calls.

Note that the delta will always show external jars added until bug 39856 is 
fixed.

Dirk I was never able to reproduce this problem. Can you please verify and 
reopen if you see the problem again?
Comment 8 Dirk Baeumer CLA 2004-05-17 12:08:04 EDT 
Which build will have the change so that I test the correct environment.
Comment 9 Jerome Lanneluc CLA 2004-05-17 12:11:18 EDT 
Tomorrow's 00:10 AM integration build will have the fix in, or if you want you 
can take o.e.jdt.core v_428b.
Comment 10 John Arthorne CLA 2004-05-18 16:19:08 EDT 
*** Bug 62176 has been marked as a duplicate of this bug. ***
Comment 11 Frederic Fusier CLA 2004-05-19 06:03:18 EDT 
Verified for 3.0 M9 with build I200405190010.