On Thursday, September 03, 2015 9:10:26 AM walt wrote: > On Wed, 2 Sep 2015 15:34:43 +0200 > Alan McKinnon <[email protected]> wrote: > > > On 02/09/2015 15:04, Neil Bothwick wrote: > > > On Wed, 2 Sep 2015 05:24:33 -0700, walt wrote: > > > > > >> If the devs can't explain slots to their > > >> users then they don't understand it themselves. (Hm. That phrase > > >> sounds familiar. Where did I get that?) > > > > > > I think it is an Einstein quote that says something like "if you > > > can't explain it in simple terms, you don't understand it". He was > > > probably having a pop at Niels Bohr and quantum theory at the time. > > > > > > Bohr said something like "if thinking about quantum theory doesn't > > > give you a headache, you don't understand it". > > > > > > > > > > And Feynmann said something along the lines of "Anyone who claims to > > understand quantum mechanics, doesn't". > > > > Back to subslots and not replying to Neil directly: They aren't that > > hard to grasp, they look like this: > > > > cat/pkg/pkg-1.2:3/4 > > > > The SLOT is 3 and the subslot is 4. As usual, different versions of > > the same package in different SLOTs can co-exist. Subslots are a > > different matter, and it's an unfortunate choice of name, as they are > > *not* a subset of a SLOT. Look at ncurses: > > > > [I] sys-libs/ncurses > > Available versions: > > (0) 5.9-r3 (~)5.9-r4 5.9-r5(0/5) (~)6.0-r1(0/6) > > (5) 5.9-r99(5/5) (~)5.9-r101(5/5) (~)6.0(5/6) > > > > There's 2 SLOTs (0 and 5), and both have versions of subslot 5 and 6. > > Subslots are most useful for things like api/abi versions where > > upstream breaks these but don't increment the major version, this is > > why we had endless issues in the past where emerge world broke stuff > > horribly and it only got fixed much later when we could run > > revdep-rebuild. Nowadays we have better tools, if the subslot changes > > for a consumed library, then all consuming packages need to be > > rebuilt. > > > > Describing and defining subslots is not hard, neither are the > > operators. The problem with subslots is the usual one - you have to > > deal with real life, and in real life upstreams sometimes do peculiar > > things to their code that doesn't exactly match the effect of a > > subslot operation. > > > > Or put another way: subslot docs describe the effect you should end up > > with, it's not always the same thing as what you *do* end up with. > > Finding that out means testing every possible circumstances and seeing > > the results, but there's an infinite variety of those. > > I just updated my virtualbox ~amd64 guest and all went well when I ran > emerge ncurses:5/5, so I'm encouraged but not fearless about doing the > same on my real machine. I'm going to be quickpkged to the max before > I try it. > > BTW, emerge world on the vbox guest did not offer to touch ncurses in > any way. I had to do it manually as I just said. > > Leveraging Neil's quote: thinking about slots (and their misnamed > subslots) gives me a 4-dimensional headache.
I don't think they're misnamed, the problem is in our heads. As end users we've come to accept that they can be co-installed as the defining property of slots. But they're about the same in every other way AFAIU. -- Fernando Rodriguez
