| 1 |
On Sat, 16 Jun 2012 18:41:51 +0200 |
| 2 |
Pacho Ramos <pacho@g.o> wrote: |
| 3 |
> > The :*/:= feature was designed to solve one specific problem: if a |
| 4 |
> > user has foo installed, and foo deps upon bar, and bar:1 is |
| 5 |
> > installed, and the user wants to install bar:2 and then uninstall |
| 6 |
> > bar:1, will foo break? :* means no, := means yes. |
| 7 |
> |
| 8 |
> And, wouldn't it be covered simply making that package not depend on |
| 9 |
> any slot specifically? |
| 10 |
|
| 11 |
Some people use "no slot" to mean "and it's fixed at build time", and |
| 12 |
some use it to mean "and I don't care". We *could* just omit :*, and |
| 13 |
use := for locking, but an explicit :* means someone has checked their |
| 14 |
work (and can be verified by repoman) whereas no slot probably means |
| 15 |
laziness. |
| 16 |
|
| 17 |
> > I'm pretty sure the route Exherbo is going to take with this is very |
| 18 |
> > different, and will involve souped-up USE flags that allow "parts" |
| 19 |
> > of a package (such as its libraries) to be kept around, possibly |
| 20 |
> > together with a special form of blocker that acts only upon |
| 21 |
> > installed packages, with a strict post ordering. It's not going to |
| 22 |
> > involve sub-slots, in any case. |
| 23 |
> |
| 24 |
> Well, probably the problem is to predict when will that be really |
| 25 |
> solved there :( |
| 26 |
|
| 27 |
Naah. This is one of those things that requires developers to put quite |
| 28 |
a lot of exta effort in to their packages in order to improve the |
| 29 |
quality of experience for users, which means it's not going to be |
| 30 |
suitable for Gentoo's development model. |
| 31 |
|
| 32 |
-- |
| 33 |
Ciaran McCreesh |
Archives