On Wed, Feb 12, 2020 at 6:42 PM Mark Thomas <ma...@apache.org> wrote:
> On 12/02/2020 17:36, ma...@apache.org wrote: > > This is an automated email from the ASF dual-hosted git repository. > > > > markt pushed a commit to branch master > > in repository https://gitbox.apache.org/repos/asf/tomcat.git > > > > > > The following commit(s) were added to refs/heads/master by this push: > > new 5a36c39 Fix timeout on zero length read/write with useAsyncIO > for NIO and APR > > 5a36c39 is described below > > > > commit 5a36c3941e99d7e1554bc68cd2acef53c76d5fcb > > Author: Mark Thomas <ma...@apache.org> > > AuthorDate: Wed Feb 12 17:27:25 2020 +0000 > > > > Fix timeout on zero length read/write with useAsyncIO for NIO and APR > > For those wondering about this, I found this while testing the Jakarta > WebSocket 2.0 TCK with Tomcat 10. A couple of the tests switch from > buffered to non-buffered messages with any empty buffer and the > subsequent flush triggers a zero length write. Whether that write should > even be triggered is a whole other question... > I think I remember I had problems already before the async API with those weird tests doing zero length writes in that TCK :( I may make changes to the fix, since: - I notice arrayHasData doesn't use offset and length, it's better if it does - With NIO arrayHasData is also used for a read, I suppose it's not a problem but I'll change it anyway Rémy
