On Mon, Jul 20, 2009 at 10:27:18AM -0400, Christopher Faylor wrote: >On Mon, Jul 20, 2009 at 01:57:59PM +0200, Corinna Vinschen wrote: >>On Jul 19 19:35, Christopher Faylor wrote: >>> On Tue, Jul 07, 2009 at 02:02:12PM -0400, Christopher Faylor wrote: >>> >On Mon, Jul 06, 2009 at 10:18:39PM -0400, Christopher Faylor wrote: >>> >>On Mon, Jul 06, 2009 at 07:03:11AM -0600, Eric Blake wrote: >>> >>>But on cygwin 1.7, when you attempt to create multiple writers to a >>> >>>single >>> >>>fifo, the second writer creates an fd just fine but then fails on any >>> >>>attempt to write to that fd: >>> >> >>> >>I see the problem but I don't know how to fix it yet. It may be a >>> >>couple of days before I can come up with a solution. >>> > >>> >Just an update: I now have a headache. Thanks again Microsoft for >>> >making it all so complicated. This is YA example of a feature (Windows >>> >Named Pipes) which behave 98% of the way you'd expect it to work. The >>> >2% is the killer. >>> >>> Just an update: @#$%^&*(! >> >>This sounds promising. > >I did get Eric's test case working last night with only an addition of >one or two hundred lines of extra code. I ended up fixing a >long-standing bug in fhandler_base::read, and found a strange problem >where fd's are inexplicably changed to textmode. I haven't tracked that >one down yet. > >Then, as I was drifting off to sleep, I realized that my fix only goes >75% of the way towards getting full fifo behavior. So, I probably still >won't be checking anything in anytime soon.
I've checked this in but I'm not done yet. I was hoping to implement this fully but I probably won't have a chance for at least a week. cgf -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple