1 |
On Wed, 29 Mar 2017 16:01:18 +0200 |
2 |
Ulrich Mueller <ulm@g.o> wrote: |
3 |
|
4 |
> UPSTREAM |
5 |
> The requested bug is considered to be out of the purview of the |
6 |
> distro and should be submitted/discussed directly with the |
7 |
> respective upstream project. This could include a number of things |
8 |
> such as changing default configuration options or behavior, adding |
9 |
> new options or functionality, or deleting support for older |
10 |
> systems. |
11 |
> |
12 |
> IMHO it would be out of proportion to remove the field, just because |
13 |
> some developers don't use it as intended. |
14 |
|
15 |
To me, this is a case of "bad choice of name leads to bad use" |
16 |
|
17 |
I doubt many people actually know that definition. |
18 |
|
19 |
But naming is hard. |
20 |
|
21 |
The RESO/UPSTREAM situation would be *much* better if bugs that were marked |
22 |
that way injected a blob of text *somewhere* on the page in a different colour |
23 |
to inform the user about what it means. |
24 |
|
25 |
"This bug has been marked `Upstream`, which means the nature of your request |
26 |
is outside the scope of things Gentoo developers consider to be things within |
27 |
their realm of responsibilities to resolve. |
28 |
|
29 |
Please contact upstream and have them implement your feature/fix your bug |
30 |
there, instead" |
31 |
|
32 |
As it stands, receiving a "RESO/UPSTREAM" resolution is almost as big a slap in the face |
33 |
as "RESO/INVALID". |
34 |
|
35 |
_Especially_ when a developer marks a ticket as that without adequate explanation as to |
36 |
why. |
37 |
|
38 |
All you see is "Piss off mate, we don't like ya" |
39 |
|
40 |
Even a class of "Resolved" is pretty much not useful here, because it implies |
41 |
to the reader "the bug is gone", even though its not really. |
42 |
|
43 |
"DEFERRED/OUTOFSCOPE" would probably be a more adequate and natural resolution |
44 |
if you had to make one in a very small amount of space that conveyed all the right information. |
45 |
|
46 |
Because then, if its merely "deferred", that implies it can be pushed back to Gentoo in |
47 |
some capacity once external effort is complete, and then the author of the bug can: |
48 |
|
49 |
1. Change the name of the bug from a request for a change, to a request for a bump to the feature-including bug |
50 |
2. Change the resolution back to open |
51 |
3. Have the bug resolved / Fixed. |
52 |
|
53 |
"DEFERRED/TESTREQ" -- Its not resolved, but action can't complete until somebody tests it |
54 |
"DEFERRED/INVALID" -- Data given so far indicates its not a bug, but perhaps better data could change that |
55 |
"DEFERRED/NEEDINFO" -- Its not "fixed" as such, but we can't continue due to lack of information. |
56 |
|
57 |
But I guess I'm reacting too much to the use of the word "Resolved", which is a bit easy to conflate. |
58 |
|
59 |
Because while it is /a resolution/, "resolved" in colloquial usage /implies/ "fixed", even though |
60 |
its use in bugzilla is no such thing. |
61 |
|
62 |
But I think a renaming/recategorizing campeign at this point is too much, and an alternative of finding a way to embed |
63 |
resolution-descriptions as helper text in the issue once a resolution state has been chosen, to smooth the path for the reader, |
64 |
and to _reinforce_ what that resolution means, would probably be more productive. |
65 |
|
66 |
But I suspect such a feature is best petitioned to bugzilla, so I could file a bug about this in Gentoo bugzilla that we |
67 |
add this feature, and somebody can mark it "RESOLVED: UPSTREAM" and really satiate my meta-recursion fetish. |