[corinna-cygwin@cygwin.com: Re: ENOTSOCK errors with cygwin dll 1.3.21 and 1.3.22]

Pierre A. Humblet Pierre.Humblet@ieee.org
Sun May 25 23:21:00 GMT 2003


At 05:20 PM 5/25/2003 -0400, Christopher Faylor wrote:
>On Sun, May 25, 2003 at 05:03:02PM -0400, Pierre A. Humblet wrote:
>>At 11:30 AM 5/24/2003 +0200, Corinna Vinschen wrote:
>>>Perhaps we should revert to Winsock2 duplicating and reactivate
>>>Jason's idea to just ignore the WSAENOTSOCK from close()?
>>>
>>>Corinna
>>
>>It's more complicated than that. The old behavior is that the dup
>>was failing silently and returning the dupeed (old) handle instead 
>>of a duped (new) one. That's not acceptable IMO, dup should fail
>>if it can't do what it's supposed to.
>>Later the program was closing both the old and new handles,
>>with the second close failing (as it should).
>>I would like to better understand what's going on, at least until
>>we are close to a release. 
>
>I intend on making a 1.5.0 release next week, as soon as I return from
>my current trip.
>
>It would be nice to figure out what's going on here, though.

I don't think I figure it in that time frame. The problem doesn't
happen on my systems, I need to find a sick machine first.
Meanwhile, looking up the current code, we could make the test
" if (!wincap.has_security ())" in the socket::dup succeed all the 
time. That would help Abraham Backus but hurt exim with Jason's 
configuration :( 

Don't you think we should improve the sparse file behavior? I could
submit a fix in a few days. 

Pierre



More information about the Cygwin-developers mailing list