1 |
On Fri, Jun 26, 2020 at 01:48:31PM -0400, Aaron Bauman wrote: |
2 |
> So, thanks for proving my point that all the tooling changes, notices, |
3 |
> ML posts, etc don't matter. Someone *will* find something to complain |
4 |
> about. |
5 |
They will also complain about the status quo, you won't win there! |
6 |
|
7 |
> >2. The report does not list maintainers, which means nobody is likely |
8 |
> >to know they have a package on the list. |
9 |
> > |
10 |
> |
11 |
> Do you argue just to argue? Sad. If someone like Robin (who at one |
12 |
> point had like 5% of the tree under his maintainer ship) complained |
13 |
> about that I may see it worthwhile. |
14 |
Some of the packages shown on the py27 list I had LONG forgotten that I |
15 |
maintained: I'll try and get to fixing them now. |
16 |
|
17 |
I did q-text-as-data this week, because I actually needed it for a quick |
18 |
project and I haven't used it since I changed my default python away |
19 |
from py27 last month. |
20 |
|
21 |
> Just another red herring... |
22 |
I'm mostly speaking to the QA team here, and Python team indirectly: |
23 |
Indirectly because the Python team is one of the few teams to step up |
24 |
and provide QA checks outside of the QA team directly. This thread |
25 |
however has shown that output of those checks however needs to become |
26 |
easier to consume. |
27 |
|
28 |
TL;DR: Please make it easier to search on the QA reports site for |
29 |
issues, and only show things directly relevant to the search. |
30 |
|
31 |
A long time ago, there was blizzy's site that listed packages that were |
32 |
stabilization candidates, and you could filter by developer. It really |
33 |
helped making it easier to detect and progress. |
34 |
|
35 |
At a bare minimum, having an on-site way that already expands the data |
36 |
and makes it searchable by developer. |
37 |
|
38 |
Have files like https://qa-reports.gentoo.org/output/gpyutils/py2.txt |
39 |
directly loaded into qa-reports and expanded out (which is cachable) and |
40 |
then let devs search w/ their browser. |
41 |
|
42 |
cat/package:slot (reason) (all-direct-maintainers),(expanded-projects) |
43 |
|
44 |
get-git-file.sh tries very hard to get there for the gentoo-ci output, |
45 |
needs performance and usability improvements, but it's vastly better |
46 |
than the py2.txt file already. |
47 |
Some issues: |
48 |
- Make it more visible! Right now you have to have a link to it from |
49 |
somewhere else, and it doesn't accept a branch name |
50 |
(https://qa-reports.gentoo.org/output/gentoo-ci/HEAD/output.html |
51 |
returns 404) |
52 |
- Why does it take 15 seconds to load? |
53 |
- Add filter by developers (by direct maintainer OR membership in alias) |
54 |
- Add filter by package/cat |
55 |
- Add filter by check name |
56 |
- Machine-readable format should be the same data as human-readable (I |
57 |
can't just take the .xml and grep it, it doesn't have maintainers at |
58 |
all) |
59 |
|
60 |
> >> See above. Qa-reports will output a very nice list (even a graphic!) |
61 |
> >of such things. Anyway, yes, I do expect devs to understand their |
62 |
> >packages state if they maintain it. Don't be so myopic. |
63 |
> > |
64 |
> >Well, you can expect whatever you want, and then you can be frustrated |
65 |
> >out of your mind when 95% of devs fail to meet your expectations. |
66 |
> > |
67 |
> I am not frustrated. I will continue to perform the same in intervals to drive the removal of Py2. |
68 |
Can we have that graphic in a searchable text format? <3. |
69 |
|
70 |
-- |
71 |
Robin Hugh Johnson |
72 |
Gentoo Linux: Dev, Infra Lead, Foundation Treasurer |
73 |
E-Mail : robbat2@g.o |
74 |
GnuPG FP : 11ACBA4F 4778E3F6 E4EDF38E B27B944E 34884E85 |
75 |
GnuPG FP : 7D0B3CEB E9B85B1F 825BCECF EE05E6F6 A48F6136 |