Noel J. Bergman wrote:
Stefano Bagnara wrote:
You also left in the removal of the ShareFileInputStream, right?
Yes, but ...
I think that one could be one of the problems for the file not being
deletable.
I disagree because although I do understand your reasoning, I have not seen
one of the temporary files leaked except in conjunction with an I/O
exception occurring during the populating of the file, e.g., the socket
closing prematurely.
--- Noel
Ok, I just wanted to be sure that we also test that revert before
introducing the FileCleaner. I've not anything against FileCleaner
itself, but
1) we are in RC and I prefer to add a catch than adding a new behaviour
(and a new thread)
2) I prefer to really fix the bug instead of relying on a thread to
workaround it ;-)
That said if even with SharedFileInputStream it works fine imho we can
keep it there.
What's the plan now?
Do you plan to commit your current version or have you more tests to do
before?
Imho all new exception handling must be applied to both branches. The
sharedfileinputstream revert (if done on 2.3) should not be applied to
trunk (by now) and FileCleaner (if at the end we'll need it) should not
be applied to trunk (by now).
Stefano
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
