Richard Heck wrote: > On 07/07/2014 06:46 AM, Jürgen Spitzmüller wrote: >> Am Sonntag 06 Juli 2014, 21:03:14 schrieb Kornel Benko: >>>> This one was the only one where I realy saw a crash. From the remaining >>>> ones 2a677592 and 5093893b5 also fixed other bugs, and 5a8b8ba8 is >>>> important, so I propose these as well. For the remaining ones I don't >>>> know: I am pretty sure that these fixes are correct, but they are morre >>>> exotic. What do others think? >>> I think all of them should go to branch as well. >> Me, too. Thanks for this work, BTW. > > Go ahead, then, Georg. (And thanks also from me.)
OK, I'll merge all fixes then. You are all welcome. It was quite easy btw: We have two needed tools (a mutex and local thread storage), so my algorithm was: 1) search for FIXME THREAD, pick one that looks interesting 2) decide what to do: Is it possible to get rid of the static data completely? If not, is it better to use a mutex or local thread storage? 3) implement and test the solution 4) goto 1) until I could not be bothered anymore Unfortunately some of the remaining FIXME THREAD occurances are more difficult. I might have a look at them as well, but I don't know yet. Georg
