Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.
Bug 314433 - [efs][testing] testDeleteSpecialCases runs into an NPE on Linux
Summary: [efs][testing] testDeleteSpecialCases runs into an NPE on Linux
Status: RESOLVED FIXED
Alias: None
Product: Target Management
Classification: Tools
Component: RSE (show other bugs)
Version: 3.2   Edit
Hardware: PC Linux
: P3 normal (vote)
Target Milestone: 3.2 RC3   Edit
Assignee: Martin Oberhuber CLA
QA Contact: Martin Oberhuber CLA
URL:
Whiteboard:
Keywords:
Depends on: 314496
Blocks: 247908 314439
  Show dependency tree
 
Reported: 2010-05-26 07:15 EDT by Martin Oberhuber CLA
Modified: 2010-05-26 16:04 EDT (History)
1 user (show)

See Also:
dmcknigh: review+

Attachments
patch v1 (2.23 KB, patch)
2010-05-26 07:29 EDT, Martin Oberhuber CLA 		no flags Details | Diff
patch v2 (2.50 KB, patch)
2010-05-26 14:42 EDT, Martin Oberhuber CLA 		no flags Details | Diff
Show Obsolete (1) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Martin Oberhuber CLA 2010-05-26 07:15:37 EDT 
Build ID: TM 3.2rc2 on Linux-GTK-RHEL-4

RSEFileStoreTest#testDeleteSpecialCases() runs into an NPE in line 178, because RSEFileStore#openOutputStream() violates the API contract by returning null.
Comment 1 Martin Oberhuber CLA 2010-05-26 07:29:24 EDT 
Created attachment 169969 [details]
patch v1

Patch fixing the issue.

The problem was, that for a broken symbolic link, "exists()" is false, but neither "isFile()" nor "isDirectory()" is true. As a result, the EFS code ran into a case where it returned null (marked by a TODO comment).

Correct behavior is throwing an error if the output stream is opened on a folder, but try creating the stream in all other cases. Also, never return null but throw a "cannot write" core exception in unexpected cases.
Comment 2 Martin Oberhuber CLA 2010-05-26 07:30:02 EDT 
Dave can you review this please? I want to get rid of known unittest errors for the TM 3.2 release.
Comment 3 David McKnight CLA 2010-05-26 10:44:46 EDT 
(In reply to comment #2)
> Dave can you review this please? I want to get rid of known unittest errors for
> the TM 3.2 release.

I'm hitting a conflict when I try to apply the patch with the latest from HEAD.
Comment 4 Martin Oberhuber CLA 2010-05-26 14:42:11 EDT 
Created attachment 170069 [details]
patch v2

Updated patch - this time it should apply cleanly.
Please review, thanks!
Comment 5 Martin Oberhuber CLA 2010-05-26 16:04:58 EDT 
Released > I20100526 - thanks for the review.