The following comment has been added to this issue:

     Author: Brett Porter
    Created: Wed, 14 Apr 2004 6:32 PM
       Body:
I have no problems on Solaris, although with somewhat fewer projects.

I'd look more at what is happening on the box. Perhaps your Solaris box has less 
physical RAM and is swapping a lot? What JDK versions are you running across each?

AFAIK Goal.addPrecursor is not run inside the dependency resolution code.
---------------------------------------------------------------------
View this comment:
  
http://jira.codehaus.org/secure/ViewIssue.jspa?key=MAVEN-1235&page=comments#action_18626

---------------------------------------------------------------------
View the issue:
  http://jira.codehaus.org/secure/ViewIssue.jspa?key=MAVEN-1235

Here is an overview of the issue:
---------------------------------------------------------------------
        Key: MAVEN-1235
    Summary: Performance problem with Maven reactor on Solaris
       Type: Bug

     Status: Unassigned
   Priority: Major

 Original Estimate: Unknown
 Time Spent: Unknown
  Remaining: Unknown

    Project: maven
 Components: 
             core
   Versions:
             1.0-beta-10
             1.0-rc3
             1.0-rc1
             1.0-rc2

   Assignee: 
   Reporter: Vincent Massol

    Created: Wed, 14 Apr 2004 5:01 AM
    Updated: Wed, 14 Apr 2004 6:32 PM

Description:
Context: 

We have about 58 Maven projects. Each project has about 85 dependencies. About 50 of 
these are dependencies to external jars. The rest are inter-project dependencies.

Problem:

When running the Maven reactor on windows machine. The reactor correctly and 
relatively quickly (30-40 seconds) finds out the ordered list of projects to run. 
However, on Solaris, it takes about 215 minutes (almost 4 hours!) to find out the 
ordered list of projects.

Analysis:

It seems the problem is somewhere in the werkz code. When using a profiling tool, 100% 
of the time is spent in the Goal.addPrecusor method (which takes as an average about 1 
minute per project per dependency to run). I don't see how the problem could be in 
werkz. But still the profiling tool points there.

Could it be the checkForCycle which is somehow having recursivity problems on Solaris?

Any idea?

Thanks


---------------------------------------------------------------------
JIRA INFORMATION:
This message is automatically generated by JIRA.

If you think it was sent incorrectly contact one of the administrators:
   http://jira.codehaus.org/secure/Administrators.jspa

If you want more information on JIRA, or have a bug to report see:
   http://www.atlassian.com/software/jira


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to