Gentoo Archives: gentoo-dev

From: Ryan Hill <dirtyepic@g.o>
To: gentoo-dev@l.g.o
Subject: [gentoo-dev] Re: RFC: changing the developer profile: FEATURES="test" -> FEATURES="test-fail-continue"
Date: Sat, 12 Jun 2010 03:57:00
Message-Id: 20100611215927.61574c95@gentoo.org
In Reply to: Re: [gentoo-dev] Re: RFC: changing the developer profile: FEATURES="test" -> FEATURES="test-fail-continue" by Jeroen Roovers
1 On Mon, 7 Jun 2010 14:02:50 +0200
2 Jeroen Roovers <jer@g.o> wrote:
3
4 > I see more and more calls for either 1) "fixing the test suite", as if
5 > that is suddenly not an UPSTREAM issue but the ebuilds' maintainers'
6
7 > When instead a test suite should do a SKIP but erroneously does a FAIL,
8 > then RESTRICT=test is not the solution. Fixing the test suite is, but
9 > then that's not the maintainer's problem, but upstream's. Oddly enough
10 > we have QA checks in place (for ICEs, for instance) that direct users
11 > directly to upstream (through the HOMEPAGE variable), when it's
12 > suddenly the maintainer's problem if a package fails its test suite
13 > (because of FEATURES=userpriv or a missing kernel feature, perhaps -
14 > nothing the maintainer can prepare the user's system for).
15
16 I'm having trouble understanding how you can say fixing build failures is
17 part of a maintainer's duties while fixing test failures is upstream's
18 problem. A test failure _is_ a build failure. Yes, we should get it fixed
19 upstream, just like any other bug. Packages can fail to compile with
20 userpriv or missing kernel features too. Should we also send users directly
21 to upstream in these cases? Can you explain the difference?
22
23 I agree fully with all your other points.
24
25
26 --
27 fonts, there's a hole in my neighbourhood
28 gcc-porting, down which of late i cannot help but fall
29 wxwidgets @ gentoo EFFD 380E 047A 4B51 D2BD C64F 8AA8 8346 F9A4 0662

Attachments

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