Jetty

Problem with array index exception with ResourceCache.java

Details

  • Type: Bug Bug
  • Status: Resolved Resolved
  • Priority: Major Major
  • Resolution: Fixed
  • Affects Version/s: 6.0.1
  • Fix Version/s: 6.1.0pre0
  • Component/s: Servlet
  • Labels:
    None
  • Environment:
    Mac OS X - jdk-1.5_06 / Linux - jdk-1.5_08
  • Number of attachments :
    1

Description

When using Jetty in an embedded manner with no extraction of war files, Jetty has a array index exception when trying to use the fill method within ResourceCache. I have not been able to reproduce this problem when simply using the out of the box jetty. The following code shows the problem that I am seeing.

protected void fill(Content content)
throws IOException

{ InputStream in = content.getResource().getInputStream(); int len=(int)content.getResource().length(); // XXX Returns -1 which causes an exception on the next line Buffer buffer = new ByteArrayBuffer(len); buffer.readFrom(in,len); in.close(); content.setBuffer(buffer); }

I've attached a patch that addresses the problem, but I don't believe that it handles the root of the problem. For some reason, the length in ResourceCache.java:146 returns the correct value, but the fill method sometimes returns -1.

Activity

Ascending order - Click to sort in descending order
Hide
Permalink
Greg Wilkins added a comment -

Well the ResourceCache is one of my least favourite parts of Jetty. It is rather ugly.

I can't see why that length should be -1? It could be that the resource is being released
from the cache just as you try to use it. So I am not sure if ignoring the -1 is the
right thing to do.

I may try to remake the resource if it returns -1 and see if that makes a difference.
if not, I will still throw an exception

Show
Greg Wilkins added a comment - Well the ResourceCache is one of my least favourite parts of Jetty. It is rather ugly. I can't see why that length should be -1? It could be that the resource is being released from the cache just as you try to use it. So I am not sure if ignoring the -1 is the right thing to do. I may try to remake the resource if it returns -1 and see if that makes a difference. if not, I will still throw an exception
Hide
Permalink
Greg Wilkins added a comment -

I found a funny call of fill twice in a row, that could cause some grief to the jarfileResource type.
I have fixed this and checked it into trunk.

If you can try trunk, that would be good.

If it does reoccur, any info you can give me on the size of the cache and if the resource may be being flushed or
accessed for the first time would be good.

you can also try turning the cache off.

Show
Greg Wilkins added a comment - I found a funny call of fill twice in a row, that could cause some grief to the jarfileResource type. I have fixed this and checked it into trunk. If you can try trunk, that would be good. If it does reoccur, any info you can give me on the size of the cache and if the resource may be being flushed or accessed for the first time would be good. you can also try turning the cache off.
Hide
Permalink
Shawn Ellis added a comment -

I tried out the latest version from the trunk and I'm no longer seeing the problem. Thanks for the help.

Show
Shawn Ellis added a comment - I tried out the latest version from the trunk and I'm no longer seeing the problem. Thanks for the help.

People

  • Assignee:
    Greg Wilkins
    Reporter:
    Shawn Ellis
Vote (0)
Watch (0)

Dates

  • Created:
    Updated:
    Resolved: