Bugzilla – Full Text Bug Listing
| Summary: | 64kb method should be a configurable problem (1FJHGVF) | ||
|---|---|---|---|
| Product: | [Eclipse Project] JDT | Reporter: | Philipe Mulet <philippe_mulet> |
| Component: | Core | Assignee: | Philipe Mulet <philippe_mulet> |
| Status: | RESOLVED WONTFIX | QA Contact: | |
| Severity: | normal | ||
| Priority: | P3 | ||
| Version: | 2.0 | ||
| Target Milestone: | 2.0 M1 | ||
| Hardware: | All | ||
| OS: | Windows 98 | ||
| Whiteboard: | |||
Bytecode generation got fixed. Nobody mentionned needing a warning for it. Closing PRODUCT VERSION: lf315 |
Sun javac 1.2.2 will only issue a warning by default: A.java:3: This code requires generating a method with more than 64K bytes. Virtual machines may refuse the resulting class file. int[] intarray = { ^ 1 warning NOTES: Philippe (10/4/99 3:16:38 PM) In presence of wide branches in such code, javac is not using gotow bytecodes (like us), thus leading to illegal code generation. Need to see any use for only making it a warning in order to comply with Sun's behavior. Philippe (10/4/99 4:07:19 PM) New javac 1.3 will properly generate a goto_w bytecodes, and not even complain for a warning. Then we should also fix the bytecode and issue a warning. PM (7/25/2001 1:59:55 PM) We should only fix the bytecode, and make a potential warning (in case embedded customer are using it).