On Mon, Jan 18, 2016 at 08:46:06AM +0000, Ian Malone wrote: > On 18 January 2016 at 01:32, Zbigniew Jędrzejewski-Szmek > <zbyszek@xxxxxxxxx> wrote: > > On Mon, Jan 18, 2016 at 12:55:54AM +0100, Heiko Adams wrote: > >> But it seems to be broken since Feb 2015, which is IMHO unacceptable > >> since a default package manager and all of its features have to work > >> absolutely reliable. > > > > When was the last time you saw a program bigger then /bin/true that was > > "absolutely reliable"? Your implicit premise that yum was bug free > > is completely bogus, just look for yum bugs in bugzilla [1]. > > > xz > tar > cp > ... I'd argue that those three programs aren't that far from /bin/true in terms of complexity. And if you look in the bugzilla, all three of those *do* have occasional bugs. And they are missing features: xz still does not support parallel compression, and people have been asking for it for years. tar does not support ACLS, etc. https://bugzilla.redhat.com/buglist.cgi?classification=Fedora&component=xz&product=Fedora https://bugzilla.redhat.com/buglist.cgi?classification=Fedora&component=tar&product=Fedora https://bugzilla.redhat.com/buglist.cgi?classification=Fedora&component=coreutils&product=Fedora&short_desc=cp&short_desc_type=allwordssubstr > Being reliable might be difficult, but it is achievable, and the more > core a tool is the more important it is that it approaches > reliability. You can't really compare programs which have a single purpose and a simple user interface and for which there is one obvious way to do things with a sprawling beast of a program that has plugins, hundreds of options, deals with the network, needs various heuristics for speed, depends on other libraries, ... and is just complex. > > It seems that with dnf we are currently in the phase of fine-tuning > > user interaction. The resolver works nicely, there is a growing system > > of plugins based on a stable API, the codebase was ported to the > > current version of python, speed is decent most of the time... There > > *are* things to fix, but calling for the return of yum is a complete > > waste of the time of everbody on this list. > > > > So there's no need to fight hyperbole with hyperbole. Do you really think that there's any chance of us returning to yum? Or a reason to advocate that? There doesn't seem to be, so it *is* a complete waste of time and this is not hyperbole. Zbyszek -- devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxxx http://lists.fedoraproject.org/admin/lists/devel@xxxxxxxxxxxxxxxxxxxxxxx
