1 |
On 8/2/21 11:03 PM, n952162 wrote: |
2 |
> On 8/3/21 7:37 AM, cal wrote: |
3 |
>> On 8/2/21 10:26 PM, n952162 wrote: |
4 |
>>> On 8/3/21 7:20 AM, n952162 wrote: |
5 |
>>>> On 8/2/21 10:10 PM, David Haller wrote: |
6 |
>>>>> Hello, |
7 |
>>>>> |
8 |
>>>>> On Mon, 02 Aug 2021, n952162 wrote: |
9 |
>>>>>> !!! All ebuilds that could satisfy |
10 |
>>>>>> "dev-python/isodate[python_targets_python3_8(-)?,python_targets_python3_9(-)?]" |
11 |
>>>>>> |
12 |
>>>>>> |
13 |
>>>>>> have been masked. |
14 |
>>>>>> !!! One of the following masked packages is required to complete your |
15 |
>>>>>> request: |
16 |
>>>>>> - dev-python/isodate-0.6.0-r2::gentoo (masked by: EAPI 8) |
17 |
>>>>>> |
18 |
>>>>>> The current version of portage supports EAPI '7'. You must upgrade |
19 |
>>>>>> to a |
20 |
>>>>> ^^^^^^^^^^^^^^^^^^^^^^^^^^ |
21 |
>>>>>> newer version of portage before EAPI masked packages can be |
22 |
>>>>>> installed. |
23 |
>>>>> You should read a bit more carefully... It seems your sys-apps/portage |
24 |
>>>>> is a bit dated. |
25 |
>>>>> |
26 |
>>>>> Upgrade sys-apps/portage first, then it should work. |
27 |
>>>>> |
28 |
>>>>> HTH, |
29 |
>>>>> -dnh |
30 |
>>>> |
31 |
>>>> I read that, but I last updated 2 months ago. So, this update breaks |
32 |
>>>> because portage was updated and new ebuilds using that are already |
33 |
>>>> being pushed out? Seems strict to me ... but /isodate/ being the only |
34 |
>>>> one? I still have a EAPI 4 on my system (1). Plenty of EAPI 5. Was |
35 |
>>>> there something in isodate that needed to be urgently updated - |
36 |
>>>> utilizing the the cutting edge EAPI? |
37 |
>>>> |
38 |
>>> We just had to update portage not so long ago. I admittedly presumed at |
39 |
>>> first that something else must be wrong, because it didn't occur to me |
40 |
>>> that portage would be so volatile. |
41 |
>>> |
42 |
>>> Everything in gentoo is so nicely configurable, but I think another |
43 |
>>> dimension should be add: configurable volatility - i.e. a configurable |
44 |
>>> hysteresis for upstream updates. |
45 |
>>> |
46 |
>>> |
47 |
>>> |
48 |
>> It sounds like you would be more satisfied with a distribution that has |
49 |
>> releases. You are fighting a losing battle to use a rolling-release |
50 |
>> distribution on a machine you intend to update infrequently. |
51 |
>> |
52 |
>> Keeping old software working in a rolling release ecosystem is a pain, |
53 |
>> doubly so if you have to maintain the newer version in parallel. What |
54 |
>> you ask for is more difficult than you think. |
55 |
>> |
56 |
> |
57 |
> Well, what you say is likely true, but does "old software" really need |
58 |
> to be kept working? Couldn't problems necessarily only be dealt with |
59 |
> in the newest versions? |
60 |
Old software does not exist in a vacuum. Old software has old |
61 |
dependencies; those dependencies get updated in ways that are |
62 |
incompatible, or stop working, or a new version of Python is released |
63 |
with which the old version of the software is incompatible; etc. Or as |
64 |
you've noticed with Portage, the old version of the software may not |
65 |
work compatibly with newer packages, so now you have to maintain older |
66 |
versions of those packages as well... |
67 |
> |
68 |
> Yes, a release distribution would be one end of the configuration |
69 |
> spectrum I'm thinking of. That works for other distibutions. |
70 |
> |
71 |
> My problem is, I want a source distribution. Maybe there's something |
72 |
> about source distributions that dictates a rolling strategy? I don't |
73 |
> see it, but it might be there. |
74 |
I don't think there is any particular reason this must be dictated; it |
75 |
is probably a coincidence of the Venn diagram of people who prefer |
76 |
source distributions and people who prefer rolling release. |
77 |
|
78 |
Many binary distributions do also have tools for building packages from |
79 |
source, which could be a solution if there are only specific packages |
80 |
you wish to customize, but the experience is not as pleasant as Portage. |
81 |
> |
82 |
> I've raised this issue before and someone mentioned a derivative of |
83 |
> gentoo, which sounds promising, but I'm not convinced by that one |
84 |
> implementation. |