1 |
On Fri, 08 Jul 2011 16:46:38 -0400 |
2 |
DJ Cozatt <ygdrasil@×××××××.net> wrote: |
3 |
|
4 |
> |
5 |
> Is a discussion/flame about the report upstream qa messages. |
6 |
> Help me out here guys and weigh in. (dons flame suit) |
7 |
> |
8 |
|
9 |
To be honest, I am still quite surprised at the fact that distribution |
10 |
maintainers even waste time fixing obvious bugs, let alone QA |
11 |
issues. If anything, outright bugs should be the responsibility |
12 |
of the upstream developers and only of the upstream developers. |
13 |
Yet there seems to be this many-tiered approach to reporting and |
14 |
fixing bugs, with each distribution maintaining its own independent |
15 |
set of reports and patches. This makes little sense. For example, |
16 |
there are patches available on Gentoo (and other distributions as well) |
17 |
that are needed to fix certain software bugs but these same patches |
18 |
are not included in the original source code. The way I see it, |
19 |
there is much manpower being wasted by having all of this duplicated |
20 |
effort. |
21 |
|
22 |
Upstream developers should be very accommodating when it comes |
23 |
to bug reports. After all, the software is *their* creation and |
24 |
their sense of pride -- if nothing else -- should impel them |
25 |
to release the best possible code. |
26 |
|
27 |
But upstream developers are known to sometimes be less than |
28 |
enthusiastic about bugs. I experienced an issue recently with |
29 |
the login program and thought it best to make a report to |
30 |
upstream only (http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=600755). |
31 |
After a long period with no significant response I decided to file |
32 |
a report with Gentoo ( https://bugs.gentoo.org/show_bug.cgi?id=324419). |
33 |
The Gentoo developers tackled the problem and found the source |
34 |
of the error. I reported these results back to upstream where |
35 |
hopefully the bug will be fixed, but there should not have been |
36 |
the need to approach both parties. |
37 |
|
38 |
As far as QA issues, this should definitely be only the concern |
39 |
of upstream developers. The only way to improve code for every user |
40 |
is to put pressure, in the form of constructive criticism, on upstream |
41 |
to adhere to good coding practice. |
42 |
|
43 |
|
44 |
> Further on the conversations wandered-to in the topic 'optimization' |
45 |
> in the kernel config menu under the heading 'General Setup' lies |
46 |
> [*] Optimize trace point call sites |
47 |
> |
48 |
|
49 |
Is This option for debugging purposes? If so then there is no |
50 |
need for it with ordinary user builds. |
51 |
|
52 |
Frank Peters |