This is the mail archive of the
cygwin-talk
mailing list for the cygwin project.
Re: Fw: [1.7] "broken" pipe implementation?
- From: Christopher Faylor <cgf-use-the-mailinglist-please at cygwin dot com>
- To: The Vulgar and Unprofessional Cygwin-Talk List <cygwin-talk at cygwin dot com>
- Date: Fri, 27 Feb 2009 17:07:20 -0500
- Subject: Re: Fw: [1.7] "broken" pipe implementation?
- References: <tcqqp4pir13cnq0oa8psu1uq4lso1q2550@4ax.com> <20090219142734.GA22666@ednor.casa.cgf.cx> <20090227005010.GA11675@ednor.casa.cgf.cx> <re3gq45oj4clmern31thehbrupl4ed6fvr@4ax.com> <49A81791.8010002@gmail.com>
- Reply-to: cygwin-talk at cygwin dot com
- Reply-to: The Vulgar and Unprofessional Cygwin-Talk List <cygwin-talk at cygwin dot com>
On Fri, Feb 27, 2009 at 04:40:49PM +0000, Dave Korn wrote:
>Andrew Schulman wrote:
>>>It qualifies as a "What in the world was I thinking?" problem.
>>
>>No idea what that is.
>
>The kind of bug you write when you've been up all night writing code,
>and you finally crash out, and when you wake up refreshed a few hours
>later you look at it and go "My god, this is all gibberish! I must have
>gone temporarily mad!".
>
>Happened to me just the other night. :-P
Ah, so it isn't just me.
In this case, I cleverly rewrote the named pipe handling to use
overlapped I/O and real named pipes but only did 95% of the
implementation. I missed the select() part although I have a clear
remembrance of implementing it. Must have been a late-night
hallucination.
That's why I spent the first hour of debugging setting breakpoints
and saying "Huh? Why didn't it get there???"
cgf