Summary: | [compiler] Consider disabling the syncOverride warning/error by default | ||||||
---|---|---|---|---|---|---|---|
Product: | [Eclipse Project] JDT | Reporter: | Thomas Watson <tjwatson> | ||||
Component: | Core | Assignee: | Philipe Mulet <philippe_mulet> | ||||
Status: | VERIFIED FIXED | QA Contact: | |||||
Severity: | normal | ||||||
Priority: | P3 | CC: | eclipse, hargrave, john.arthorne, markus.kell.r | ||||
Version: | 3.4 | ||||||
Target Milestone: | 3.5 M3 | ||||||
Hardware: | PC | ||||||
OS: | Windows XP | ||||||
Whiteboard: | |||||||
Attachments: |
|
Description
Thomas Watson
2008-10-02 15:18:21 EDT
There was one warning in core.resources, and it was invalid. It was an empty method so synchronization was not needed. There was one warning in JDT APT, on a method that did a little extra unsynchronized work and then called the superclass. Adding synchronization was unnecessary but didn't hurt anything. I agree with the assertion that the synchronized keyword is not a public contract. Its absence does not mean that a method is not threadsafe; its presence does not mean that the method is threadsafe or deadlock safe. JSR-305 proposes annotations which do describe synchronization contracts. Created attachment 114159 [details]
Proposed patch
Released for 3.5M3. Fixed Verified for 3.5M3 using I20081026-2000 build. |