| Issue 24: | Release problem in operation autoclean | |
| 3 people starred this issue and may be notified of changes. | Back to list |
From an email report: " The crashes seem to occur when I am uploading multiple files at once (e.g. I have dragged ~5 files over a bucket window). However, it's not related to the size of the files as they can be quite small and I still get the crash. The problem MAY however be related to (or exacerbated by) having a somewhat slow connection. It seems I get the crashes at home (w/ slow connection) more often than at work. "
|
|||
,
Jan 18, 2007
I also see this crash, and as you said in Issue 17 , they're probably related. I'm hoping your patch fixes both of them, as I'm rather tired of getting halfway through file transfers (being charged Amazon's nominal rate), and then crashing before finishing. Last night, I left a transfer running before heading off to bed, giving S3 Browser "free reign" of the system so it wasn't competing with my browser, etc., and it still crashed. |
|||
,
Jan 20, 2007
Attached you will find a diff that fixes this bug. The summary should be changed to something like 'Operation console autoclean causes crash'. In the S3ObjectListController class when operationDidFinsh: was called it first called its super. Doing this removed the operation from the current operations. However, when the unlog operations in enabledit removed the operation from all holding containers decrementing the retain count by 1 each time untimately dealloc'ing the object. The solution is to retain the object if it is being removed from the log array and immediately autorelease it so it won't be dealloc'ed until the end of the run loop.
|
|||
,
Jan 21, 2007
That would also explain why I had trouble reproducing this, I'm usually running without autoclean. r82
Status: Integrated
|
|||
,
Jan 21, 2007
(No comment was entered for this change.)
Summary: Release problem in operation autoclean
|
|||
,
Jan 23, 2007
Please verify in 1.0.6 or later.
Status: Fixed
|
|||
| ► Sign in to add a comment |
