On Wed, Mar 09, 2016 at 09:40:38PM +0100, Georg Baum wrote: > Scott Kostyshak wrote: > > > Dear all, > > > > I think we are getting close to being able to release 2.2.0rc1. We have > > not discussed whether to release a beta3 or move directly to rc1. If > > others disagree with moving directly to rc1, please do speak up and > > let's discuss. > > rc1 is fine. > > > Below are the items to consider fixing for rc1. > > > > 1. Although #9968 is not a fresh regression (there was a problem in > > 2.1.x), it seems to be worse for 2.2.0dev (although I'm guessing this is > > due to different Qt versions). Georg has posted a patch which hopefully > > improves the situation. We just need testing of the patch and we need > > to decide whether it is good enough for 2.2.0. > > Unfortuntely the patch will not really help, the effect is barely visible. > > I would postpone this. We neither have somebody who can debug the crash on > windows, nor somebody who knows QAbstractItemModel better than me stepped up > to implement symbol computation on demand (triggered by scrolling through > the list). Of course it would be nice to fix at least the crash, but I don't > see how that could be done. My previous attempts to explain how to get a > backtrace from the MSVC debugger failed.
Guillaume committed a fix that was tested by Enrico and Uwe. If you have time to check it, take a look at b3bed292. > > 8. What am I missing? > > There is also the question which compiler to use on windows. In general it > is unwise to do all the testing with one compiler and then switch the > compiler for the release. IMHO the only justification for switching > compilers would be known serious problems with the old one which are fixed > by the new one, so that the risk of yet unknown problems would be > acceptable. +1 Scott
signature.asc
Description: PGP signature
