1 |
-----BEGIN PGP SIGNED MESSAGE----- |
2 |
Hash: SHA256 |
3 |
|
4 |
On 07/09/12 03:25 PM, Fabian Groffen wrote: |
5 |
> On 07-09-2012 12:03:16 -0700, Zac Medico wrote: |
6 |
>> On 09/07/2012 11:17 AM, Fabian Groffen wrote: |
7 |
>>> I guess real-life examples, more extensively described than you |
8 |
>>> did before, with exactly where it goes wrong, and how the |
9 |
>>> situation is improved would help. |
10 |
>> |
11 |
>> Perhaps some of the greatest frustrations for Gentoo users stem |
12 |
>> from the lack of support for automatic rebuild of packages when |
13 |
>> necessary. Imagine how nice it would be if necessary rebuilds |
14 |
>> would automatically occur when appropriate, so that you wouldn't |
15 |
>> experience build failures that require you to manually intervene |
16 |
>> by running revdep-rebuild, perl-cleaner, or something like that. |
17 |
>> And there are other kinds of necessary rebuilds that don't |
18 |
>> trigger build failures, but lead to runtime failures that are |
19 |
>> noticed much later (like xorg driver failures after a major |
20 |
>> xorg-server update). Sub-slots can be used to solve the bulk of |
21 |
>> problems like these that our users have had to deal with |
22 |
>> manually. |
23 |
> |
24 |
> I like that! Kudos for making it work! |
25 |
> |
26 |
> I just wonder what the heck that has to do with SLOT. This |
27 |
> discussion has been done before in this thread, and it somehow |
28 |
> settled. |
29 |
> |
30 |
>> ... sub-slots are a relatively simple extension to slot-operator |
31 |
>> deps, and they are poised to greatly improve user experience (via |
32 |
>> automatic rebuilds) if they are included in EAPI 5. |
33 |
> |
34 |
> And we want it. But is it a good idea to add some feature that |
35 |
> feels like just a hack? |
36 |
> |
37 |
> |
38 |
|
39 |
Originally the sub-slot idea came about because one of the ways |
40 |
"around" all of this broken-and-requiring-afterthefact-rebuilding was |
41 |
to just make everything slotted -- so there would always be multiple |
42 |
slots of everything installed -- and use slot-operators to indicate |
43 |
when things should be re-emerged |
44 |
|
45 |
Although this would work, the end result would (imo at least) be |
46 |
horrible on-disk. |
47 |
|
48 |
Sub-slots allow the main part of SLOT to still specify what's |
49 |
installed on disk, while allowing PMS to identify and trigger rebuilds |
50 |
for SLOT changes based on slot-operators. |
51 |
|
52 |
I see it akin to the '-r' portion of ${PV} -- Used by portage to |
53 |
trigger updates but having very little meaning to the actual version |
54 |
of the package that gets installed. (ok i might be stretching it with |
55 |
this) |
56 |
|
57 |
-----BEGIN PGP SIGNATURE----- |
58 |
Version: GnuPG v2.0.19 (GNU/Linux) |
59 |
|
60 |
iF4EAREIAAYFAlBKUK4ACgkQ2ugaI38ACPDbCAEAiG+7hQch043se8ZfDE4qC52w |
61 |
79ZImWn5jazqGQDN3zsA/3B1AJR+SWxUFDHZF1LArX0r0Gd7J2madTqP0m+llxuG |
62 |
=7IEF |
63 |
-----END PGP SIGNATURE----- |