1 |
On 01/28/2018 09:35 AM, Alec Warner wrote: |
2 |
> |
3 |
> |
4 |
> On Sun, Jan 28, 2018 at 9:51 AM, Zac Medico <zmedico@g.o |
5 |
> <mailto:zmedico@g.o>> wrote: |
6 |
> |
7 |
> The --dynamic-deps=n default causes confusion for users that are |
8 |
> accustomed to dynamic deps, therefore add a --changed-deps-report |
9 |
> option that is enabled by default (if --usepkgonly is not enabled). |
10 |
> |
11 |
> The --quiet option will suppress the report if none of the packages |
12 |
> having changed dependencies are in the dependency graph, since they |
13 |
> are harmless in that case. If any of these packages *are* in the |
14 |
> dependency graph, then --quiet suppresses the big NOTE section of |
15 |
> the report, but the HINT section is still displayed since it might |
16 |
> help users resolve problems that are solved by --changed-deps. |
17 |
> |
18 |
> Example output is as follows: |
19 |
> |
20 |
> !!! Detected ebuild dependency change(s) without revision bump: |
21 |
> |
22 |
> net-misc/openssh-7.5_p1-r3::gentoo |
23 |
> sys-fs/udisks-2.7.5::gentoo |
24 |
> |
25 |
> NOTE: For the ebuild(s) listed above, a new revision may be |
26 |
> warranted if there |
27 |
> has been a dependency change with significant consequences. |
28 |
> Refer to the |
29 |
> following page of the Gentoo Development Guide for examples of |
30 |
> circumstances that may qualify: |
31 |
> |
32 |
> |
33 |
> https://devmanual.gentoo.org/general-concepts/ebuild-revisions/ |
34 |
> <https://devmanual.gentoo.org/general-concepts/ebuild-revisions/> |
35 |
> |
36 |
> If circumstances qualify, please report a bug which specifies |
37 |
> the current |
38 |
> version of the ebuild listed above. Changes to ebuilds from |
39 |
> the 'gentoo' |
40 |
> repository (ending with '::gentoo') can be browsed in GitWeb: |
41 |
> |
42 |
> https://gitweb.gentoo.org/repo/gentoo.git/ |
43 |
> <https://gitweb.gentoo.org/repo/gentoo.git/> |
44 |
> |
45 |
> Use Gentoo's Bugzilla to report bugs only for the 'gentoo' |
46 |
> repository: |
47 |
> |
48 |
> https://bugs.gentoo.org/ |
49 |
> |
50 |
> In order to suppress reports about dependency changes, add |
51 |
> --changed-deps-report=n to the EMERGE_DEFAULT_OPTS variable in |
52 |
> '/etc/portage/make.conf'. |
53 |
> |
54 |
> HINT: In order to avoid problems involving changed dependencies, use the |
55 |
> --changed-deps option to automatically trigger rebuilds when |
56 |
> changed |
57 |
> dependencies are detected. Refer to the emerge man page for more |
58 |
> information about this option. |
59 |
> |
60 |
> Bug: https://bugs.gentoo.org/645780 |
61 |
> |
62 |
> |
63 |
> I can't really support sending this large report to users. |
64 |
> |
65 |
> 1) Its fairly common practice today. |
66 |
> 2) All users will get the report. |
67 |
> 3) A subset of them will file bugs about the report. |
68 |
> 4) Devs make a decision about revbumping vs not; there doesn't seem to |
69 |
> be a way for devs to say "no this change is intentional, stop nagging |
70 |
> users." |
71 |
|
72 |
I think in practice we need to revbump for most changes. If the changes |
73 |
weren't worth propagating, then we wouldn't make them. |
74 |
|
75 |
> Ultimately I'm unsure what we are trying to accomplish here. Do we think |
76 |
> Devs are doing 4 on accident? |
77 |
|
78 |
It doesn't matter whether it's intentional or an accident. The problem |
79 |
is that users need to learn to react appropriately, and I think |
80 |
--change-deps-report is probably the most efficient way to train them. |
81 |
|
82 |
> If so, why can't we give them tools to find it ahead of time (repoman |
83 |
> et. al.) or tools to find it post-commit (tinderbox or similar; but |
84 |
> these are single-sourced reports.) |
85 |
|
86 |
Sure that can be done, but it will take some work. Honestly I think the |
87 |
devs can get it right without all that, they just need some |
88 |
reinforcement that has been absent up until now. |
89 |
|
90 |
Meanwhile, users need to be trained to cope with whatever comes their way. |
91 |
|
92 |
> I also feel like we are pushing action onto users here. If we agree with |
93 |
> the premise that this is a bug (and devs should always bump) then we |
94 |
> don't need users to take any action; |
95 |
|
96 |
I really do think that the devs can learn to get it right with a little |
97 |
reinforcement. |
98 |
|
99 |
> we could build automation that sends these 'reports'. Its not like the |
100 |
> user is going to add anything interesting to the report. Making them do |
101 |
> it by hand just introduces transcription errors in filing. We just need |
102 |
> to get their permission to file the reports automatically when portage |
103 |
> finds them. |
104 |
|
105 |
The thing is, we let commits flow directly from git to rsync. Good luck |
106 |
with finding volunteers to be the gatekeepers for this. |
107 |
-- |
108 |
Thanks, |
109 |
Zac |