| 1 |
On 11/08/2014 02:05 PM, hasufell wrote: |
| 2 |
> On 11/08/2014 10:52 PM, Jauhien Piatlicki wrote: |
| 3 |
>> 08.11.14 22:47, hasufell написав(ла): |
| 4 |
>>> On 11/08/2014 10:30 PM, Rich Freeman wrote: |
| 5 |
>>>> On Sat, Nov 8, 2014 at 2:48 PM, hasufell <hasufell@g.o> wrote: |
| 6 |
>>>>> On 11/08/2014 08:32 PM, hasufell wrote: |
| 7 |
>>>>>>> Sorry to chime in like that but if you don't mind, I'd like to ask for a |
| 8 |
>>>>>>> real-life example for badly declared dependencies with a few words why |
| 9 |
>>>>>>> those are bad and how to make them actually better? |
| 10 |
>>>>>>> |
| 11 |
>>>>>> |
| 12 |
>>>>>> from dev-haskell/hashtables (note "hashtables" != "hashable"): |
| 13 |
>>>>>> || ( ( >=dev-haskell/hashable-1.1:=[profile?] |
| 14 |
>>>>>> <dev-haskell/hashable-1.2:=[profile?] ) |
| 15 |
>>>>>> ( >=dev-haskell/hashable-1.2.1:=[profile?] |
| 16 |
>>>>>> <dev-haskell/hashable-1.3:=[profile?] ) |
| 17 |
>>>>>> ) |
| 18 |
>>>>>> |
| 19 |
>>>>>> Latest stable version of dev-haskell/hashable is 1.2.1.0. |
| 20 |
>>>>>> On a stable system (arch) the paludis dep-solver will try to match the |
| 21 |
>>>>>> first group first and realize that there is also a stable version |
| 22 |
>>>>>> 1.1.2.5 that matches that group. At that point there is a correct |
| 23 |
>>>>>> solution, but since that involves downgrading a package, it will require |
| 24 |
>>>>>> user-intervention, because it may not be what the user wants. |
| 25 |
>>>>>> (this is the easy scenario... if downgrading causes blockers, you get |
| 26 |
>>>>>> much more interesting output) |
| 27 |
>>>>>> |
| 28 |
>>>>> |
| 29 |
>>>>> To be more specific... it is assumed that hashable-1.2.1.0 is already |
| 30 |
>>>>> installed. Every time the dep solver runs through those packages without |
| 31 |
>>>>> specifying what you want, you will hit the downgrade-problem. |
| 32 |
>>>> |
| 33 |
>>>> I'm missing the problem. The package requires one of two ranges of |
| 34 |
>>>> hashable versions. One of them is already installed. The dependency |
| 35 |
>>>> is satisfied. |
| 36 |
>>>> |
| 37 |
>>> |
| 38 |
>>> The one that is installed (1.2.1.0) is *excluded* by the first group, |
| 39 |
>>> but there is a valid version that fits instead (1.1.2.5). |
| 40 |
>>> |
| 41 |
>>> That's the point where the assumptions start about what the depstring |
| 42 |
>>> means and what the user wants. |
| 43 |
>>> |
| 44 |
>> |
| 45 |
>> So the problem is only with intervals? First of all, maintainer would specify higher interval first here and it would solve a problem with possible downgrading. |
| 46 |
> |
| 47 |
> I have a feeling that this is an assumption as well. PMS just says this |
| 48 |
> is an 'any-of' group. There is not a single word about the processing |
| 49 |
> order of these specs or which one to prefer, in which case some is |
| 50 |
> better than the other and so on. |
| 51 |
|
| 52 |
I think the two obvious algorithms are: |
| 53 |
|
| 54 |
A) If the user's resolver parameters request maximum upgrades, then the |
| 55 |
resolver should choose the choice that results the most upgrades. |
| 56 |
|
| 57 |
B) If the user's resolver parameters request minimum change, then the |
| 58 |
resolver should choose the choice which results in keeping the most |
| 59 |
installed packages in place. |
| 60 |
|
| 61 |
For the dev-haskell/hashtables scenario, the choice on the right wins |
| 62 |
regardless of whether you're using algorithm A or algorithm B. |
| 63 |
-- |
| 64 |
Thanks, |
| 65 |
Zac |
Archives