1 |
> On 21 Feb 2021, at 04:30, Tim Harder <radhermit@g.o> wrote: |
2 |
> |
3 |
> Hi all, |
4 |
> |
5 |
> Is there interest in enforcing some basic QA for the semi-formatted |
6 |
> comments in profiles/package.mask (and possibly other profiles file |
7 |
> types)? I have code implementing the basic functionality done for |
8 |
> pkgcheck, but wondered if the format should be standardized and |
9 |
> documented more than it may be already before support is merged. |
10 |
> |
11 |
|
12 |
I definitely have interest in this. Every so often, I find a class of mistakes |
13 |
like incorrectly-ordered masks, wrong email format, and so on, and fix |
14 |
them in the area I’m touching as it’s cheap to fix whatever file I’m on - |
15 |
not necessarily so to fix every single one in profiles/. |
16 |
|
17 |
This also gives us scope to make requirements like e.g. a bug reference, |
18 |
clear date for removal if last-rites, and generally be a bit more verbose |
19 |
by making clear what the expectations are for message content. |
20 |
|
21 |
(There is a clear benefit for last-rites having at least one bug — it gives |
22 |
users a chance to object or mention alternatives.) |
23 |
|
24 |
Anyway, not trying to bikeshed re last-rites process, I just mean there’s |
25 |
definitely a collection of uses here. Thanks for working on it. |
26 |
|
27 |
Note that this was brought up last month too by jstein, so there is definitely other |
28 |
interest: https://archives.gentoo.org/gentoo-dev/message/0600146362529770aa88225b29de46ae. |
29 |
|
30 |
> Also, I'm unsure it's been noticed but `pkgcheck scan --commits` now |
31 |
> verifies any profile changes done in git commits so this support would |
32 |
> automatically get run for package.mask changes (and other enabled files) |
33 |
> when using pkgcheck locally in that fashion. |
34 |
> |
35 |
> Thanks, |
36 |
> Tim |
37 |
> |