| Summary: | Code completion should prefer methods with the correct return type | ||||||
|---|---|---|---|---|---|---|---|
| Product: | z_Archived | Reporter: | Lars Vogel <Lars.Vogel> | ||||
| Component: | Recommenders | Assignee: | Marcel Bruch <marcel.bruch> | ||||
| Status: | CLOSED DUPLICATE | QA Contact: | |||||
| Severity: | enhancement | ||||||
| Priority: | P3 | CC: | Lars.Vogel, lerch, sebastian | ||||
| Version: | unspecified | Keywords: | helpwanted | ||||
| Target Milestone: | --- | ||||||
| Hardware: | PC | ||||||
| OS: | Windows 7 | ||||||
| Whiteboard: | |||||||
| Attachments: |
|
||||||
|
Description
Lars Vogel
Created attachment 206707 [details]
Current behavior
I think the current JDT does actually sort already by exact matches.... So maybe an extension would be to use hierachy or chains (the return type of a method does not fit but a sub-method) We currently rely in the the scoring system of JDT, i.e., for ranking proposals we use "JDT's proposal relevance + our probability". While perfect matches are already covered by JDT (and thus implicitly by in Code Recommenders too), it seems reasonable to perform an "is assignable" check to further improve the probability/score of a proposal. We are currently redesigning some internal APIs for extensions of Queens University and McGill University. Afterwards, we'll evaluate how costly supertype checking is and decide how/whether to implement this. We also welcome contributions for this feature. (In reply to comment #3) > Afterwards, we'll evaluate how costly > supertype checking is and decide how/whether to implement this. *very* costly. This feature cannot be implemented using JDT means. marking this one as duplicate of Bug 384443 as they basically request the same feature. let's track this in one bug report. *** This bug has been marked as a duplicate of bug 384443 *** |