1 |
On Wed, Aug 21, 2013 at 5:50 AM, Sergey Popov <pinkbyte@g.o> wrote: |
2 |
> |
3 |
> As i said earlier, we should recruit more people -> then problem will go |
4 |
> away. |
5 |
|
6 |
This is a point most of the people in this thread seem to be dancing |
7 |
around that's sort of problematic. You can talk about recruiting |
8 |
until you're blue in the face, but the simple fact is Gentoo DOESN'T |
9 |
have adequate manpower. And has it ever, really? Can you honestly |
10 |
say we've ever had a solid surplus of devs with time [0]? We've |
11 |
gotten where we are, by and large, because Gentoo works smarter. |
12 |
|
13 |
Fundamentally, I see this as a problem of tooling. |
14 |
|
15 |
Let's turn the question around; try thinking about it like this: What |
16 |
tools have historically allowed relatively few active developers |
17 |
handle stablisation and integration of upstream patch flow IN SPITE of |
18 |
not having a lot of recruits? What tools could be added to assist |
19 |
with, if not outright _remove_ steps of the process? |
20 |
|
21 |
I'd like to point out something that jumped out to me as a red flag |
22 |
earlier (not to pick on you specifically, Tom; this is just the |
23 |
cleanest example I saw), and turn it into an example: |
24 |
|
25 |
On Wed, Aug 21, 2013 at 4:51 AM, Tom Wijsman <TomWij@g.o> wrote: |
26 |
> |
27 |
> Well, they are listed there; but it's quite some work to actually go |
28 |
> through that list, that is, manually check the bugs of ~2000 packages |
29 |
> as well as file a STABLEREQ bug, takes quite a while... |
30 |
> |
31 |
This right here is a real problem. Any time you're talking about |
32 |
doing anything on this scale "manually", you've already lost the |
33 |
battle. You need a tool to minimise the overhead of time and |
34 |
cognitive load. What would that tool look like? Think about the |
35 |
steps involved and how you can reduce them to only the parts that |
36 |
absolutely require decisions on your part. |
37 |
|
38 |
>> At least in the areas I usually work, I have found a combination of |
39 |
>> the automatic stabilisation requests and imlate have definitely cut |
40 |
>> back on the bitrot. |
41 |
> |
42 |
> A single unimportant bug can prevent the automatic STABLEREQ bug from |
43 |
> getting filed; as for imlate, not everyone seems to know that tool, not |
44 |
> everyone seems to run it. Attention for some stabilizations is lost... |
45 |
> |
46 |
First off, why do developers not know about the tools? How can this |
47 |
be addressed? For a start, I'd suggest making sure the tools are at |
48 |
least mentioned in the docs. Somewhere other than the amd64 Arch |
49 |
Testing guide from 2006. [1] That's the only concrete (i.e. actually |
50 |
in the DOCS rather than the ML archives) documentation I've found so |
51 |
far, and it only references the file, rather than the tool. Maybe in |
52 |
the devmanual Tools Reference? [2] |
53 |
|
54 |
But, imlate is a good example of a tool that could ease the time cost |
55 |
of grindy crap. You showed before that it can get an ordinary count |
56 |
bounded on n days. That's handy, but only a little. Build out: |
57 |
- How many of those stablereq bugs reference versions no longer in the |
58 |
tree? Those can probably get closed. |
59 |
- How many have newer STABLE versions in the tree in the same slot? |
60 |
Probably fine to close those, too. |
61 |
- Of the remaining, how many have patches or ebuilds attached? Those |
62 |
may be solved problems waiting for closure; shortlist them. |
63 |
- How many are packages with newer versions that have been in the tree |
64 |
for >30 days? Consider changing the target version, then. |
65 |
- How many have open blockers, and what are those blockers (w/link and |
66 |
summary)? Scan for low-hanging fruit jumping out in that list. |
67 |
- Get views by category; are there categories where updates are more |
68 |
important? Things in @system, and things with security concerns |
69 |
(stuff in net-*) should probably get higher priority; games... |
70 |
probably less so. |
71 |
- Are there bugs with certain keywords in the body that should raise |
72 |
priority? Things like "security" or "overflow" might be good |
73 |
candidates. |
74 |
- Are there bugs with certain keywords in the body that indicate it'll |
75 |
be really easy to decide? e.g. "trivial" or "minor" might turn up some |
76 |
of those super-small version bumps that you pretty much know aren't |
77 |
going to affect stability. |
78 |
|
79 |
These are just examples off the top of my head, and by no means |
80 |
bulletproof, but these are in the class of improvements that have ROI |
81 |
because they reduce a task that previously took developer time to one |
82 |
that takes CPU time. CPU time is essentially free compared to the |
83 |
value of dev time. |
84 |
|
85 |
And I'm not saying more recruiting shouldn't happen, but relying on it |
86 |
is no better than hoping at $deity for bugs to close themselves. ;) |
87 |
|
88 |
Cheers, |
89 |
Wyatt |
90 |
|
91 |
[0] Okay, maybe in the "glory days" when we were higher up on |
92 |
Distrowatch and thing were really kicking. (I know, I know, "DW isn't |
93 |
representative", but really? Sabayon is doing better than we are, |
94 |
now?) |
95 |
[1] http://www.gentoo.org/proj/en/base/amd64/tests/index.xml?part=1&chap=2 |
96 |
[2] http://devmanual.gentoo.org/tools-reference/index.html |