Gentoo Archives: gentoo-dev

From: Duncan <1i5t5.duncan@×××.net>
To: gentoo-dev@l.g.o
Subject: [gentoo-dev] Re: kernel bug reporting policy
Date: Thu, 12 Aug 2004 09:49:51
Message-Id: pan.2004.08.12.09.49.44.218462@cox.net
In Reply to: Re: [gentoo-dev] kernel bug reporting policy by Maurice van der Pot
1 Maurice van der Pot posted <20040812001759.GB25017@××××××××.com>,
2 excerpted below, on Thu, 12 Aug 2004 02:17:59 +0200:
3
4 > On Wed, Aug 11, 2004 at 07:04:21PM -0500, Jason Waldhelm wrote:
5 >>
6 >> i searched using 'ALL 2.6.8-rc4' (using ALL as the example suggests) and
7 >> your bug is the only one that comes up.
8 >
9 > Maybe I'm alone in this, but in general I do not want to see bugs that
10 > have been marked RESOLVED. If I do ALL nfs (it is a problem with nfs),
11 > then I get 83 hits. 18 of those are still open (and returned by just
12 > searching for nfs) and 19 of those (18 open + mine) are still relevant. As
13 > the bug database grows, this will become even more of a pain (think glibc,
14 > gcc, portage, ...)
15
16 Note the difference in Jason's search example vs your own. Jason's search
17 includes the version, yours doesn't.
18
19 You have a point with the too many hits thing, but there are multiple ways
20 to narrow a search. You narrowed it by excluding RESOLVED bugs (by not
21 including ALL), while he narrowed it using the version number. It has
22 already been pointed out that there are times when narrowing by excluding
23 RESOLVED might not be appropriate. The same of course applies to version
24 number. However, in this case version number narrowing would be more
25 appropriate, due to the type of package and bug under discussion,
26 where the bug is known to be not only a version specific regression, but
27 an rc specific one as well! In this case, there's no reason to search all
28 kernel bugs, or even all kernel 2.6 bugs, or even all kernel 2.6.8 bugs,
29 since it's known to be 2.6.8-rc4 specific. Further, that's a very
30 efficient narrowing since it's a very narrow category. OTOH, excluding
31 RESOLVED bugs not ONLY may exclude the bug you want to see, but is far
32 less efficient as well, because as you yourself said, there are still
33 multiple bugs listed, while by version narrowing, there's only the single
34 bug listed.
35
36 I guess one lesson here is that search strategy is important. Use the
37 wrong one, and you either fail to see what you need, or get overwhelmed by
38 the returns. Use the right one, and it's likely you'll get a manageable
39 list of hits, that include /just/ what you are looking for. If it's not
40 manageable, perhaps it needs refined a bit more, but be sure not to refine
41 out the hit you are looking for.
42
43 --
44 Duncan - List replies preferred. No HTML msgs.
45 "They that can give up essential liberty to obtain a little
46 temporary safety, deserve neither liberty nor safety." --
47 Benjamin Franklin
48
49
50
51 --
52 gentoo-dev@g.o mailing list