* Christopher Faylor (Fri, 23 Nov 2007 11:25:50 -0500) > On Thu, Nov 22, 2007 at 06:05:49PM -0000, Thorsten Kampe wrote: > >* Corinna Vinschen (Thu, 22 Nov 2007 16:42:10 +0100) > >> ==== CYGWIN THANKSGIVING EDITION ==== > >> > >> I've made the new version of the Cygwin DLL, 1.5.25-1, and associated > >> utilities available for testing. Version 1.5.24-2 remains the current > >> version for now. > >> > >> This is a bug fix release. Important changes since 1.5.24-2: > > > >Unfortunately the changes that made rsync run about 2 - 6 times faster > >[1] (on a backup of about 100,000 files residing on a FAT32 volume) > >didn't make it into the prelease. > > > >The first snapshot that exhibited the stunning increase was the > >snapshot from 2007-08-02 and the last one was the snapshot from 2007- > >08-13. > > > >[1] http://thread.gmane.org/gmane.os.cygwin/91350/focus=91509 > > So, does this mean that the snapshot was faster and then it wasn't?
Yes, all the snapshots between 2007-08-02 and 2007-08-13 were much faster. You remarked "does rsync use pipes by any chance? If so, that may be at least part of the reason since that code has been revamped." Unfortunately I don't really understand Corinna's suggestion "Using gcc and cvs you should be able to find out which change after the 2007-08-13 snapshot slowed rsync down again." Does it mean that I should compile any committed change after August 13th? That won't work because I never used cvs or gcc on my own. > Or that the snapshot was exhibiting a problem with unclosed file handles. There was an issue but it was fixed right away by Corinna[1] Thorsten http://article.gmane.org/gmane.os.cygwin/91507 -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
