You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@felix.apache.org by "Richard S. Hall (JIRA)" <ji...@apache.org> on 2010/12/26 21:43:45 UTC

[jira] Closed: (FELIX-2721) [Framework] Implement custom manifest parser and avoid JarFile

     [ https://issues.apache.org/jira/browse/FELIX-2721?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Richard S. Hall closed FELIX-2721.
----------------------------------

    Resolution: Fixed

Karl committed finishing touches on this.

> [Framework] Implement custom manifest parser and avoid JarFile
> --------------------------------------------------------------
>
>                 Key: FELIX-2721
>                 URL: https://issues.apache.org/jira/browse/FELIX-2721
>             Project: Felix
>          Issue Type: Improvement
>          Components: Framework
>    Affects Versions: framework-3.0.6
>            Reporter: Richard S. Hall
>            Assignee: Richard S. Hall
>            Priority: Minor
>             Fix For: framework-3.0.7
>
>
> In previous performance tuning work, I noticed that JarFile consumes a lot of memory, particularly due to the use of JarFile.getManifest(). This caches the manifest, perhaps the entire manifest, which is unnecessary for OSGi since we only use the main attributes. At the time I wasn't trying to address memory consumption, but speed and this didn't appear to improve it. However, now I have received reports of memory consumption issues in the field from people trying to use the Felix framework on embedded devices.
> After having sent a patch framework with custom manifest parsing, I have reports of considerable decreases in memory consumption. I think it makes sense to make this change. To avoid indirectly reading the manifest, we will need to avoid using JarFile altogether, since it will implicitly read the manifest and cache it during some operations.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.