1 |
On 12/07/17 15:36, William L. Thomson Jr. wrote: |
2 |
> On Wed, 12 Jul 2017 15:19:32 +1000 |
3 |
> "Sam Jorna (wraeth)" <wraeth@g.o> wrote: |
4 |
> |
5 |
>> On 12/07/17 15:14, William L. Thomson Jr. wrote: |
6 |
>>> Is it in system? |
7 |
>>> Is it in a set? |
8 |
>>> Is it in world? |
9 |
>>> If no to all, its a dep, warn! |
10 |
>> |
11 |
>> All this says is whether the package was explicitly installed and |
12 |
>> recorded in world, or is a member of @system. The target package may |
13 |
>> or may not be a dependency, may or may not have been --oneshot'd. |
14 |
> |
15 |
> Then when the user sees the warning they can discard knowing they |
16 |
> merged the package with --oneshot. |
17 |
|
18 |
I have trouble remembering what I ate for dinner last night, let alone |
19 |
what I may or may not have merged a week, month or year ago, or what |
20 |
options I used when merging it. |
21 |
|
22 |
> What harm does a warning do? |
23 |
|
24 |
Depends on the user, which can't really be avoided, but means that |
25 |
warnings should be clear and meaningful, otherwise they become |
26 |
background noise. |
27 |
|
28 |
>> It may have been installed as a dependency but the requiring package |
29 |
>> was removed, |
30 |
> |
31 |
> Then the person failed to run --depclean and maintain their system. |
32 |
> Either way, did the person install the package directly? |
33 |
> |
34 |
> If the package was not installed by the user. Should they not be warned |
35 |
> about removing something they did not install? |
36 |
|
37 |
Such as: |
38 |
|
39 |
emerge --unmerge dev-python/keyring |
40 |
* This action can remove important packages! In order to be safer, use |
41 |
* `emerge -pv --depclean <atom>` to check for reverse dependencies before |
42 |
* removing packages. |
43 |
|
44 |
>> or may have been installed as an orphan but is now a |
45 |
>> dependency. |
46 |
> |
47 |
> Now being a dependency the warning would be valid. |
48 |
|
49 |
"Sometimes being accurate" is not the most noble of goals. |
50 |
|
51 |
>> Assuming that if it's not in a set it must be a dependency is |
52 |
>> incorrect and misleading. |
53 |
> |
54 |
> Again there are various ways. There cannot be that much overhead to |
55 |
> check if a single package depends on one being removed. If you cannot |
56 |
> use simpler methods as mentioned. |
57 |
> |
58 |
> Clearly people have objections to warnings about packages users did not |
59 |
> install themselves.... |
60 |
> |
61 |
|
62 |
So the idea is to duplicate the functionality of '--depclean <package>' |
63 |
without actually checking to see if the package is a dependency, only |
64 |
whether it is listed in a set; or to check if it's a dependency of |
65 |
/something/ and, if so, redirect the user to the command they should be |
66 |
using anyway? |
67 |
|
68 |
-- |
69 |
Sam Jorna (wraeth) |
70 |
GnuPG ID: D6180C26 |