Bug 141677 - [rename] Suppress refactoring-may-fail alert when the errors are just warnings treated as errors
Summary: [rename] Suppress refactoring-may-fail alert when the errors are just warning...
Status: NEW
Alias: None
Product: JDT
Classification: Eclipse Project
Component: UI (show other bugs)
Version: 3.2   Edit
Hardware: Macintosh Mac OS X - Carbon (unsup.)
: P3 enhancement (vote)
Target Milestone: ---   Edit
Assignee: Markus Keller CLA
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on: 147029
Blocks:
  Show dependency tree
 
Reported: 2006-05-13 15:28 EDT by Gordon Henriksen CLA
Modified: 2007-07-29 09:19 EDT (History)
0 users

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Gordon Henriksen CLA 2006-05-13 15:28:43 EDT
If the compilation errors in question are merely warnings which Eclipse is configured to treat as errors through Preferences > Java > Compiler > Errors/Warnings, it seems logical that Eclipse could safely suppress the alert.

Eclipse SDK
Version: 3.2.0
Build id: I20060505-1306
Comment 1 Olivier Thomann CLA 2006-05-14 21:11:15 EDT
If the user wants them to be reported as error, then this means these errors should be fixed. If they should be "ignored", then they should be set as warnings/ignore.
Comment 2 Gordon Henriksen CLA 2006-05-14 22:49:10 EDT
Seems there are already numerous special cases for this. For instance, Change Method Signature to remove a parameter does not produce an alert if the parameter is never used, even if that's set to be an error. Extract Local Variable will work without complaint even in the face of numerous errors. Here's an example that has maybe 5 errors (promoted warnings):

	package eg;
	import java.util.*;
	public class Example {
		public void greet(String n) {  // warning: parameter never read
			String name = "Bill"; // warning: local never read
			List x = new ArrayList(); // warning(3): {{List,ArrayList} is a raw type,local var never read}
			System.out.println("hi");
		}
	}

Inconsistent with those above cases, a Rename refactoring on the class complains that it might fail, although it clearly won't, and the warning of course goes away if the relevant messages are demoted to warnings.

The alert is merely annoying, but the warning configuration and this alert seem as if they should be fully orthogonal, with the refactoring alert reserved for non-discretionary failures.
Comment 3 Jerome Lanneluc CLA 2006-05-15 09:12:38 EDT
Moving to JDT UI
Comment 4 Philipe Mulet CLA 2006-05-15 09:14:52 EDT
Haven't checked (maybe already so), but a criteria could also be to avoid producing more diagnostics than previously.

Anyhow, this belongs to UI.
Comment 5 Martin Aeschlimann CLA 2006-06-10 06:49:07 EDT
Should then be fixed in rename refactoring to work better with the optional warnings.
Comment 6 Eclipse Webmaster CLA 2007-07-29 09:19:58 EDT
Changing OS from Mac OS to Mac OS X as per bug 185991