On 8/14/12 2:51 PM, Rob Weir wrote: > On Tue, Aug 14, 2012 at 4:42 AM, FR web forum <ooofo...@free.fr> wrote: >> Hello list, >> >> French user has open this report: >> https://issues.apache.org/ooo/show_bug.cgi?id=120481 >> >> Unable to drag & drop column header from data sources beamer (F4) to Writer >> document. >> This happen only with MacOS 10.8 >> > > What I look for in a release blocker at this point: > > 1) Is the impact severe, i.e., data loss, blocks use of a major features, etc. > > 2) Will it impact many users? > > 3) Is there no adequate workaround that we can put in the release notes? > > > Again, there is no risk-free fix. The fact that this bug was > introduced via an earlier fix proves that. > > > It is safe to assume that with more days of testing we will find more > bugs. We could probably test for another 6 months and still find new > bugs. That is the nature of complex software. But the question is > not whether or not we fix this bug. Of course we will. The question > is between: > > A) Fix this bug later in 3.4.3 or in 3.5 > > or > > B) Fix this bug in 3.4.1, and delay the delivery of this release, > which contains many other critical fixes for more severe issues. Also > introduce further risks for 3.4.1 by changing more code. > > IMHO the decision between A and B should be based on impact, severity and > risk. > > Does anyone have an opinion on the impact and severity? >
I think as release manager I should have an opinion ;-) Ok to be more serious, we have analyzed the problem. Armin remembered the issue 117990 that we have integrated early in the 3.4.1 branch and that solved another serious problem with the clipboard in the calc application. This fix was tested and verified but as side effect we have now this new problem. It is definitely a regression to 3.4. Andre and I have analyzed the code and we have a potential fix that solves of course both problems. But we are not 100% sure if this fix is good enough or if it introduce new problems. As Rob mentioned every fix contains a risk to break something else. At the moment I am unsure but I tend to leave it out of 3.4.1. The question is how many users are affected by this problem or if the workaround via Insert -> Fields -> ... is ok for now. Reverting the fix for 117990 is no option because it is serious as well and probably more users are affected by this issue as by the new problem. For now I would propose that we continue to analyze the fix a little bit more and try to understand the code better. And when we have to rebuild for another critical show stopper, we will potentially integrate the fix for this problem as well if we feel more comfortable with the fix. Juergen