> On November 19, 2018 at 5:17 PM Andrea Pescetti <pesce...@apache.org> wrote:
> Does the current explanation actually explain why the first command > succeeds on Fedora? (of course, this also reflects in GUI, where the > check for updates only succeeds on Fedora). Not at all. If the problem was simply on the TLS version then it shouldn't work on Fedora also. If it only worked on CentOS, it could have been some compilation glitch. But this is really puzzling! > > Can this change be reversed or an exception opened for AOO? > > Sure, we have multiple options available. Including moving this server > to a smaller (or even dedicated) virtual machine. Note that, even if we > miss a full explanation, we could still try this, but I'm really curious > to know whether the Fedora behavior can be explained. Excellent. I'm glad this can be solved for the current (4.1.6) and past versions (4.1.5, 4.1.4, etc) which people are already using. > > how are users going to be notified that any future version is available? > > Fortunately we can fix this in code, even though Brane explained well > why this is hard to do on the 4.1.x line. I meant if this could not be solved on the server side, there was no way for people on 4.1.5 or previous to be notified of any other version (including 4.1.6) > I'm just left with the big question mark of why 4.1.6 works well in > Fedora and the same binaries fail on CentOS 7 (where I built them) and > Ubuntu 16.04. If we could explain this last part we would be able to go > to Infra with more confidence. I agree. Maybe Brane (or anyone else) can suggest other tests? Unfortunately now the ball is on your side, I can't help with Fedora... However if users start to complain on forums that they can't update to 4.1.6, the server side solution should be applied sooner. Regards, Pedro --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
