On Nov 30, 2010, at 10:01 AM, Andrew Hanushevsky wrote: > OK, the penultimate patch has been applied from Brian (thanks Brian -- but you really need an account to do this :-) Hi Pete, Once this goes to upstream ROOT, you can remove the corresponding CMS patch. I don't think this will happen before your current build. > The only thing on my list is to disable generating shared libraries in various packages via autotools. It's not critical, if I get it done then could someone test it to see if it actually still works? Anyway, if Gerri gets ahead that's ok. > I should be able to test things later this week, but not before Thursday. Brian > Andy > > On Tue, 30 Nov 2010, Lukasz Janyst wrote: > >> On Tue, Nov 30, 2010 at 3:37 PM, Brian Bockelman <[log in to unmask]> wrote: >>> You should be able to reproduce this by hand without ROOT, right? >>> >>> Create a client, call ReadV with buf=NULL and one entry in the vector, and verify the return code is zero instead of the length of the request. >> >> Sure, that's trivial. >> >>> Within ROOT, create a TTree with one branch and one entry and read it with TTreeCache. >> >> This I cannot see however (with your >> 087FAA16-F9E6-DF11-9D26-0030487CAF0E.root, and its EventMetaData >> tree), let's not waste more time on this though. I hope this fixes >> what you need fixed (and doesn't have any unexpected side effects). >> >> Lukasz >>