On Fri, Jan 20, 2012 at 08:56:28AM -0500, Jeff Trawick wrote:
> On Fri, Jan 20, 2012 at 7:41 AM, <[email protected]> wrote:
> > Author: jorton
> > Date: Fri Jan 20 12:41:18 2012
> > New Revision: 1233882
> >
> > URL: http://svn.apache.org/viewvc?rev=1233882&view=rev
> > Log:
> > * server/core_filters.c (ap_core_input_filter): Only treat EAGAIN as
> > success if a non-blocking read was requested; for a blocking read,
> > it is an error condition.
>
> In case anyone remembers: Is this the same area discussed by
> Jim+Joe+Jeff at AC2011 based on Jim's Jenkins testcase?
I was reading that code for PR 52476.
I just re-read that thread on Jenkins. I am not convinced by the
comments in get_remaining_chunk_line(). This:
if (lineend[len - 1] != APR_ASCII_LF) {
return APR_EAGAIN;
}
looks suspect. If the server sent a few bytes of chunk-line (*no* LF)
then closed the socket, we'd hit that condition and should fail with an
error, not EAGAIN. Possibly. I'll try some test cases.
