On Fri, 17 Feb 2023 at 09:56, Thomas Huth <th...@redhat.com> wrote: > > On 17/02/2023 10.06, Markus Armbruster wrote: > > Thomas Huth <th...@redhat.com> writes: > ... > > My view on all this is a bit more pragmatic. > > > > For a human developer, the difference between "dnf install > > python-sphinx" and "pip install sphinx" is, in my opinion, close to > > negligible. Really no comparison to "git-clone GCC and bootstap it". > > You seem to disagree with that. > > Honestly, being a Python ignorant, I completely messed up my system with > "pip" already a couple of times, especially if the instructions forgot to > tell me to use the "--user" switch. So yes, I tend to disagree ;-)
Seconded. I trust my distro package manager and I know how it works, and I know how to uninstall a package later if I want to revert what I've done. I do not know or trust what the heck pip is doing or where it's trying to install anything, because it's not a tool I habitually use. I can't remember if I've managed to mess up the system with it, but I've definitely had the experience of "install stuff with pip, do a distro upgrade later, the pip installed stuff is all busted". > > For automated builds in general, and distro packaging in particular, the > > difference is real, and could even be a show stopper. But who's > > packaging bleeding edge QEMU on CentOS 8? I suspect the only automated > > builds are our own CI, where the difference is real, but hardly a show > > stopper. > > If we've got the feeling that nobody out there really builds QEMU on older > long-term distros anymore, then why the heck are we still trying to support > this according to our support statement? I don't think anybody is *packaging* new QEMU on an old distro. I do think we have users who do ad-hoc from-source builds. thanks -- PMM
