1 |
On 01/29/2017 10:06 AM, Andreas K. Huettel wrote: |
2 |
> So I've been talking to kent\n and the conclusion is that our ideas basically |
3 |
> do and achieve the same, with a slightly different approach. (I still like |
4 |
> mine better though. :) |
5 |
> |
6 |
> However one valid point that came up in discussions is - whether an arch |
7 |
> supports stable keywords is a per-arch setting, not a per-profile setting. So |
8 |
> we can actually make things much easier (and the transition safer). |
9 |
> |
10 |
> Proposal No 2: |
11 |
> * Leave profiles.desc unmodified |
12 |
> * Introduce a new file arch.desc, which contains the "stability status" of an |
13 |
> arch; |
14 |
> |
15 |
> Syntax: 2 columns, |
16 |
> # arch status |
17 |
> amd64 stable |
18 |
> mips testing |
19 |
> sh unstable |
20 |
> |
21 |
> The meaning of the keywords "stable", "testing", "unstable" is the same as in |
22 |
> the previous proposal, |
23 |
> |
24 |
>> "Does this arch support stable keywords, and how should "arch" vs. "~arch" |
25 |
>> be treated?" |
26 |
>> - "stable": separately check consistency of ~arch and arch tree, both have |
27 |
>> to be OK. This is what repoman is doing now, and is the default if the 4th |
28 |
>> column is undefined. |
29 |
>> - "testing": treat "arch" as "~arch" when requiring consistency, do not |
30 |
>> check "arch" alone. Useful if an arch wants to prepare going stable, useful |
31 |
>> for arch teams maintaining a pseudo-stable subset for stages. repoman could |
32 |
>> have a new command line switch that temporarily upgrades from "testing" to |
33 |
>> "stable" (for arch team work). |
34 |
>> - "unstable": check "~arch" only, "arch" in an ebuild produces a fatal |
35 |
>> repoman error |
36 |
> |
37 |
> The combination of current profiles.desc and new arch.desc provides the same |
38 |
> flexibility as in the previous e-mail. |
39 |
> |
40 |
> Compatibility and transition: |
41 |
> |
42 |
> 0) PMS should be amended to allow the additional file. |
43 |
> |
44 |
> 1) Compatibility: No arch.desc and new system, or arch not listed in arch.desc |
45 |
> *Arches* are treated as "stable" by repoman (current behaviour), with profile |
46 |
> status according to profiles.desc. |
47 |
> Gentoolkit and other tools trying to determine a list of stable arches should |
48 |
> fall back to current method of scanning profiles.desc for stable profiles. |
49 |
> |
50 |
> 2) Compatibility: arch.desc and old system |
51 |
> Tools ignore the unknown file (?). |
52 |
> Repoman and other tools may emit surplus errors when profiles are checked on |
53 |
> arches that are "testing" (they check the consistency of the stable tree |
54 |
> alone, which is not OK, since "arch" is supposed to be treated like "~arch"). |
55 |
> |
56 |
> 3) On introduction of the new column, it will be set to "stable" for all |
57 |
> stable arches, "testing" for all arches where "inofficial" stable keywords |
58 |
> exist (sh, s390, ...), and "unstable" everywhere else. |
59 |
> |
60 |
> 4) Arches in "testing" or "unstable" may eventually consider re-introducing |
61 |
> stable *profiles* so their deptree in ~arch remains consistent. |
62 |
> |
63 |
> More opinions, flames, cookies? |
64 |
> |
65 |
> Cheers, Andreas |
66 |
|
67 |
I like what I have read here and elsewhere. |
68 |
|
69 |
Solutions centric around a minimized profile will allow gentooers to |
70 |
use identical (minimized) profiles for a wide variety of hardware |
71 |
types:: different uP, dsp, fpga, custom (soc, asic, etc). I.E. |
72 |
heterogeneous gentoo clusters without any systemd associated codes. I do |
73 |
not have access to Kent's posting so perhaps a reference to Kent's ideas? |
74 |
|
75 |
|
76 |
It would be very much appreciated if there is a posting on this list of |
77 |
what the consensus becomes, with some discussion as to how it will |
78 |
affect the ever expanding variety of cluster formations, particularly |
79 |
gentoo-style unikernel types of clusters and how to cluster up a variety |
80 |
of gentoo-embedded systems, which are actually quite similar to |
81 |
unikernel based clusters. |
82 |
|
83 |
|
84 |
A state-diagram of just how all of these profiles are intertwined, would |
85 |
help to clarify the details and thus be keenly appreciate, when a final |
86 |
verdict is reached. |
87 |
|
88 |
|
89 |
hth, |
90 |
James |