Gentoo Archives: gentoo-dev

From: Mart Raudsepp <leio@g.o>
To: gentoo-dev@l.g.o
Subject: Re: [gentoo-dev] [RFC] package.deprecated to mark packages deprecated and report dependencies
Date: Sat, 17 Aug 2019 09:24:14
Message-Id: 1669e5b040761693477c364e03ce8130afda8e58.camel@gentoo.org
In Reply to: Re: [gentoo-dev] [RFC] package.deprecated to mark packages deprecated and report dependencies by Thomas Deutschmann
1 Ühel kenal päeval, R, 16.08.2019 kell 19:58, kirjutas Thomas
2 Deutschmann:
3 > Hi,
4 >
5 > I like the idea. This will allow the following change in workflow:
6 >
7 > When you now want to last-rite app-misc/foo for example, you would
8 > schedule a CI run. I.e. create a pull request against Gentoo
9 > repository
10 > at GitHub containing your package.mask entry. When the results will
11 > be
12 > available, you will start filling bugs against packages depending on
13 > the
14 > package you want to get rid off. Once all depending packages are
15 > gone,
16 > you will commit the mask. However, this process can take some time
17 > and
18 > in theory someone could add a new dependency on your package in the
19 > meanwhile...
20 >
21 > Thanks to the new package.deprecated file we would have a check in
22 > real
23 > time against current repository. And once all CI warnings are gone
24 > you
25 > can commit the mask.
26
27 I imagined it more in terms of replacing that PR CI run to get the
28 initial list and start signaling that we want it to go away. However
29 packages shouldn't be put in there that are really still used a lot
30 (say, x11-libs/gtk+:2).
31 I don't think it should nag maintainers using repoman (or pkgcheck in
32 the future) by default (at least for pre-existing cases), but included
33 in a CI run as lower prio warning to be able to quickly search through
34 the list to see what the state of things is, if it's realistic to
35 really get rid of it by filing the bugs, etc. And it should warn for
36 completely new packages, if they add a dep on it. Bonus points if the
37 CI check can signal that a deprecated use isn't the case anymore in a
38 newer revision already - to signal that it's a matter of clean-up work
39 there.
40 But that's just my thoughts, and what you propose is also an
41 improvement. Though with that kind of approach I would instead mark it
42 up and push that to main tree, and then do the bugs from the refreshed
43 report with the low prio warnings instead though; or remove the entry
44 if it's still too much and unrealistic.
45
46
47 Mart

Attachments

File name MIME type
signature.asc application/pgp-signature