On Sun, Apr 22, 2018 at 09:09:40PM +0200, Maxime Villard wrote: > I recently told membership-exec that I would be less outspoken, and more > convivial, so here's a try: > > Le 22/04/2018 à 20:51, Joerg Sonnenberger a écrit : > > On Sun, Apr 22, 2018 at 12:36:36PM +0200, Maxime Villard wrote: > > > Where are they? I haven't been made aware of any issue related to > > > SVS+clang. > > > > Yes, I did make you aware that SVS killed VirtualBox. > > You are being dishonest. You did tell me that SVS didn't work with your > VirtualBox. At no point in time did you tell me that it was related to clang > or anything close to being a compiler issue, and not an implementation > issue.
I didn't claim that now either. All I said is that SVS was known to be broken in my environment. Understanding the issue took a while as reproduction was annoying given that people continued to break the LLVM build every other day, so it was hard to use official images for testing. > In fact, if you want my point of view, you reported your "problem" in a way > that made me just unable to understand what it was about. I had to ask you > repeatedly, question after question, what is your virtualbox, what is your > cpu, is it hw-assisted, and so on. Shockingly, I would have included more data if I know whether any of the parameters are relevant. I originally ruled out LLVM since I thought it worked on a different (physical) machine. No longer sure I did, given that the machine is not supposed to use SVS for the obvious performance implications. Joerg