https://bz.apache.org/bugzilla/show_bug.cgi?id=64202
Mark Thomas <ma...@apache.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution|--- |FIXED
--- Comment #15 from Mark Thomas <ma...@apache.org> ---
Found the issue.
The read needs to be in a loop.
A TLS Network read can result in zero application bytes. Therefore you have to
keep reading until you get a timeout, EOF or data. The original patch read a
second time but did not handle the case of that read also resulting in zero
byte. The current code also keeps track of the remaining read timeout int he
case of multiple reads which adds a little complexity.
I haven't looked a write. It is possible a similar issue exists there.
Marking this as FIXED since the commit that triggered this has been reverted.
--
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org
For additional commands, e-mail: dev-h...@tomcat.apache.org
