Bugzilla – Full Text Bug Listing
| Summary: | [assist] completion proposes callout override with nothing to override | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Tools] Objectteams | Reporter: | Stephan Herrmann <stephan.herrmann> | ||||
| Component: | OTDT | Assignee: | Stephan Herrmann <stephan.herrmann> | ||||
| Status: | VERIFIED FIXED | QA Contact: | |||||
| Severity: | normal | ||||||
| Priority: | P3 | ||||||
| Version: | 0.7 | ||||||
| Target Milestone: | 0.7 M4 | ||||||
| Hardware: | PC | ||||||
| OS: | Linux | ||||||
| Whiteboard: | |||||||
| Attachments: |
|
||||||
Created attachment 170775 [details]
test & fix
The new test testCompletionCalloutDeclaration1a witnesses why this happened:
The sorted array of base method contains "fubar" (the test object) and "toString".
When processing this array back to front inspecting "toString" causes kind
to be set to CALLOUT_OVERRIDE, which will then be applied to all subsequent
proposals, too.
Fixed by using a tmp local "curKind".
Patch has been released as r409. Verified for M4 using build 201006111044 (In reply to comment #3) > Verified for M4 using build 201006111044 OT build was actually 201006111053. |
When creating a method binding using completion, sometimes a callout override binding is proposed although no role method exists that would be overridden by the callout. The simplest example to reproduce this is public class MyBase { void foo() {} public String toString() { return ""; } } public team class MyTeam { protected class R playedBy MyBase { // complete here and select binding to "foo" } } Moreover, the compiler doesn't actually complain about the bogus "=>" binding token.