1 |
2010/1/12 Brian Harring <ferringb@×××××.com>: |
2 |
>> There's no discussion because Brian refuses to address any comments |
3 |
>> on the proposal and just says "we should do it anyway, and if you |
4 |
>> want it done properly instead, do it yourself". |
5 |
> |
6 |
> This is a bit of bullshit, per the norm. There is plenty of |
7 |
> discussion- the problem is you don't like the direction it's gone. |
8 |
> You want a whole new vdb- I don't oppose that. However I'm not |
9 |
> interested in trying to standardize a new vdb format into PMS, at |
10 |
> least not yet. |
11 |
|
12 |
No, I want a decent cache proposal that lets package managers know |
13 |
what's changed, not one that sometimes (but not always) might let |
14 |
package managers know when some things have changed, but not what's |
15 |
changed and not what they can still assume. |
16 |
|
17 |
> Your argument can basically be summed up as "don't do the minimal |
18 |
> tweak, do the whole new vdb with defined caches that all can share". |
19 |
|
20 |
No, I want the well defined caches that all can share. |
21 |
|
22 |
> The daft thing about this is that you're ignoring one core transition |
23 |
> issue w/ vdb2- if someone did create a vdb2, they still would need a |
24 |
> synchronization mechanism (one quite similar to what I'm proposing). |
25 |
|
26 |
If you replace VDB, you need a well defined cache mechanism. So let's |
27 |
do that bit now. |
28 |
|
29 |
> 1) portage/pkgcore support the PMS defined vdb2 while paludis doesn't |
30 |
> 2) portage/pkgcore are invoked modifying the livefs; vdb1, vdb2 is |
31 |
> updated. |
32 |
> 3) paludis is invoked. vdb1 is updated, vdb2 is not |
33 |
> 4) portage and pkgcore now cannot rely upon vdb2, since vdb1 now |
34 |
> contains extra modifications due to paludis not supporting vdb2. |
35 |
|
36 |
No, we'd not do it that way. If we're ditching VDB, the only sane way |
37 |
to do it is to ditch it with an rm -fr when creating the new layout. |
38 |
Keeping two sets of data around is going to lead to breakage no matter |
39 |
how well we do things. |
40 |
|
41 |
> Summarizing; the synchronization primitive is needed for any future |
42 |
> vdb2 |
43 |
|
44 |
No. A *proper* cache validation mechanism is needed. What you're |
45 |
suggesting isn't enough to use for anything at all. |
46 |
|
47 |
> Summing it up; what ciaran wants is reliant on what I'm proposing, |
48 |
|
49 |
No, what I want in the long term is reliant upon implementing a decent |
50 |
cache setup in the short term. |
51 |
|
52 |
-- |
53 |
Ciaran McCreesh |