On Jun  1 16:01, Ken Brown wrote:
> On 6/1/2016 3:54 PM, Corinna Vinschen wrote:
> > On Jun  1 15:05, Ken Brown wrote:
> > > On 6/1/2016 2:28 PM, Corinna Vinschen wrote:
> > > > On Jun  1 11:22, Ken Brown wrote:
> > > > > And this also fixes the original problem in emacs that started the thread:
> > > > > 
> > > > >   https://www.cygwin.com/ml/cygwin/2016-05/msg00318.html
> > > > 
> > > > I reverted commit 252a07b0ad3353abcd0fcd9b1b65ff977acd679e in git master
> > > > for now, unless we find a reason to rework this again.
> > > 
> > > Thanks.  By the way, I'd like to take the opportunity to thank Peter Castro
> > > for the Cygwin Time Machine
> > > (http://www.fruitbat.org/Cygwin/#cygwintimemachine).  The bisection I did,
> > > spanning a period of about 4 years, would have been virtually impossible
> > > without it.
> > 
> > Huh.  I was thinking of git bisect.  It's incredibly helpful.  But, yes,
> > all power to the Cygwin time machine :)
> 
> I first used the time machine to narrow down the problem to between 1.7.35
> and 2.0.0.  Then I used git bisect to finish the job.  But even for that I
> needed the time machine; for example, I needed an old version of the w32api
> headers in order to build an old cygwin1.dll.

Wow, thanks for going to such great lengths to track down the
regression.  Andrew, do we have a GOLDSTAR for Ken here?  He really
deserves it.


Thanks,
Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Maintainer                 cygwin AT cygwin DOT com
Red Hat