Bug 234307 - Java Model Exception from org.eclipse.jdt.internal.core.Buffer.save(Buffer.javga:361)
Summary: Java Model Exception from org.eclipse.jdt.internal.core.Buffer.save(Buffer.ja...
Status: VERIFIED FIXED
Alias: None
Product: JDT
Classification: Eclipse Project
Component: Core (show other bugs)
Version: 3.4   Edit
Hardware: PC Linux
: P3 blocker (vote)
Target Milestone: 3.4 RC3   Edit
Assignee: Jerome Lanneluc CLA
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2008-05-28 02:49 EDT by Jack Fong CLA
Modified: 2008-06-04 11:00 EDT (History)
5 users (show)

See Also:
philippe_mulet: review+
frederic_fusier: review+
david_audel: review+

Attachments
Proposed fix and regression test (3.76 KB, patch)
2008-05-29 05:43 EDT, Jerome Lanneluc CLA 		no flags Details | Diff
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 Jack Fong CLA 2008-05-28 02:49:26 EDT 
Build ID: eclipse-SDK-3.4RC1

Steps To Reproduce:
1.  
2.
3.


More information:
I'm only getting this exception on the linux platform.  My program works fine in Windows.

Please take a look at this exception:
Java Model Exception: Core Exception [code 368] Resource '/TestFile.java' does not exist.
        at org.eclipse.jdt.internal.core.Buffer.save(Buffer.java:384)
        at org.eclipse.jdt.internal.core.Openable.save(Openable.java:480)
        at org.eclipse.jdt.internal.core.CompilationUnit.save(CompilationUnit.java:1270)
        at foo.DumpApplication.externalize(DumpApplication.java:123)
Caused by: org.eclipse.core.internal.resources.ResourceException: Resource '/TestFile.java' does not exist.
        at org.eclipse.core.internal.resources.Resource.checkExists(Resource.java:317)
        at org.eclipse.core.internal.resources.Resource.checkAccessible(Resource .java:194)
        at org.eclipse.core.internal.resources.File.getContentDescription(File.java:277)
        at org.eclipse.jdt.internal.core.Buffer.save(Buffer.java:361)
        at org.eclipse.jdt.internal.core.Openable.save(Openable.java:480)
        at org.eclipse.jdt.internal.core.CompilationUnit.save(CompilationUnit.java:1270)
        at foo.DumpApplication.externalize(DumpApplication.java:123)

When I debug into JDT Buffer.java, I found that at line 361, we should only call:
  IContentDescription description = this.file.getContentDescription();
if this.file.exists() == true similar to line 373.
Comment 1 Thomas Hallgren CLA 2008-05-28 03:10:24 EDT 
Reassigning to jdt-core-inbox since I don't think this is in any way related to the spaces project.
Comment 2 Jerome Lanneluc CLA 2008-05-29 05:34:39 EDT 
Looking at the trace, it appears that it is impossible to get in this situation since Openable.save(...) first acquires a buffer which ensures that the file exists. 

In fact, this can happen if a race condition occurs: the file is deleted after the buffer is acquired, and thus the Buffer.save(...) operation fails if the encoding is "UTF-8". I was able to confirm this behavior on Jack's machine.

I entered bug 234551 to capture the race condition issue.

However one can see this issue by using IBuffer.save(...) directly. The following snippet shows the problem:

IBuffer buffer = (new WorkingCopyOwner() {}).createBuffer(
  getCompilationUnit("/P/X234307.java"));
buffer.setContents(
  "public class X234307 {\n" +
  "}"
);
IProject project = getProject("P");
String defaultCharset = project.getDefaultCharset();
try {
  project.setDefaultCharset("UTF-8", null);
  String newContents = 
    "public interface X234307 {\n" +
    "}";
  buffer.setContents(newContents);
  buffer.save(null, false/*don't force*/);
  char[] contents = Util.getFileCharContent(
    getFile("/P/X234307.java").getLocation().toFile(), null);
  assertSourceEquals("Unexpected source", newContents, new String(contents));
} finally {
  deleteBuffer(buffer); // this deletes the file as well
  project.setDefaultCharset(defaultCharset, null);
}

Since there is no workaround, and the reporter marked the bug as blocker we should fix this bug for 3.4RC3
Comment 3 Jerome Lanneluc CLA 2008-05-29 05:43:10 EDT 
Created attachment 102594 [details]
Proposed fix and regression test
Comment 4 Jerome Lanneluc CLA 2008-05-29 05:44:41 EDT 
Philippe, Frederic, David, can you please approve/review?
Comment 5 Frederic Fusier CLA 2008-05-29 06:08:45 EDT 
Patch looks good: +1
Comment 6 David Audel CLA 2008-05-29 07:02:16 EDT 
Patch looks good: +1
Comment 7 Philipe Mulet CLA 2008-05-29 09:25:19 EDT 
+1 for 3.4rc3
Comment 8 Jerome Lanneluc CLA 2008-05-29 09:35:06 EDT 
Fix and test released for 3.4RC3
Comment 9 Jack Fong CLA 2008-05-29 15:01:43 EDT 
Tested patch and it is good.  Thank you Jerome.
Comment 10 David Audel CLA 2008-05-30 10:59:15 EDT 
Verified for 3.4RC4.
Comment 11 Olivier Thomann CLA 2008-06-04 11:00:11 EDT 
The copyright was not updated!