This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
Summary: | incorrect error badges creates infinite loop | ||
---|---|---|---|
Product: | java | Reporter: | Rich Unger <richunger> |
Component: | Source | Assignee: | Rastislav Komara <moonko> |
Status: | RESOLVED WONTFIX | ||
Severity: | blocker | ||
Priority: | P3 | ||
Version: | 6.x | ||
Hardware: | All | ||
OS: | All | ||
Issue Type: | DEFECT | Exception Reporter: | |
Attachments: | simple project |
Description
Rich Unger
2008-05-29 00:22:39 UTC
Steps to reproduce would be perfect. Otherwise, it may take a while until we find the cause. First, the errors that are causing the incorrect error badge may give us a lead. These can be found like this: -in ${nbuser.dir}/var/cache/index/<highest-number>/s<number>/errors are stored all errors from batch "compile". The number maps to the given source root through ${nbuser.dir}/var/cache/index/<highest-number>/segments . -while having the file with incorrect badge open, select Tools/Tasklist, and select Current File in the tasklist window. Should show the batch compile errors for the given file. Thanks. The errors are: ERROR:105:cannot find symbol symbol \d method open(org.apache.lucene.store.Directory,java.util.List<org.apache.lucene.index.Term>) location\d class org.apache.lucene.index.IndexReader ERROR:105:incompatible types found \d org.apache.lucene.index.IndexReader.open required\d org.apache.lucene.index.IndexReader It looks like the problem stems from the handling of files that we patch. So, IndexReader is a class in Lucene which we patch (kind of like NB modules do in http://wiki.netbeans.org/DevFaqModulePatching), and it's referred to from TermDeleter. Perhaps the source parser is getting the 2 versions confused? BTW, when I open TermDeleter, there's an error badge in the editor's tab, but the square at the top-right is green. Created attachment 62773 [details]
simple project
Attached project, when loaded with contrib/autoproject, produces an incorrect error badge for, I believe, the same reason my large codebase does. This particular example doesn't go into the same infinite loop, though. Don't know if that helps. Just FYI, I'm still seeing this, despite the attached sample project no longer creating an error badge. My actual sources still create an error badge that results in an infinite loop, though thus far I've been unable to isolate a good test case. Precise steps to reproduce would be welcome. Overtake. Can't successfully reproduce. Awaiting reporter interaction (Exact steps to reproduce, userdir (at least messages.log)) . Planing for future. Will be closed within next 6 months if there will be no interaction from reporter. Resolving all issues with milestone "future" as LATER. If you feel strongly that it should be implemented please reopen and set the target milestone to "next". NetBeans.org Migration: changing resolution from LATER to WONTFIX |