1 |
Hello, |
2 |
|
3 |
since I've not really been involved in the whole Sunrise discussion I'd |
4 |
like to give my view in a condensed form, instead of spreading it out |
5 |
over 20 replies in the ongoing discussion. Also I hope to summarize the |
6 |
main points a bit, but I know this mail is far from objective and as |
7 |
such not much of good summary. |
8 |
|
9 |
There are really 2 big discussion points in the whole Sunrise discussion |
10 |
as far as I can see. First there is the purpose of Sunrise and how that |
11 |
ties into Gentoo and secondly there's how it came into being. I would |
12 |
also like to discuss how I think Sunrise will influence the work of |
13 |
developers. Let's tackle them one by one. |
14 |
|
15 |
* The purpose of Sunrise |
16 |
|
17 |
The purpose of Sunrise as far as I can distill them from their goals : |
18 |
|
19 |
1. Make stale ebuilds in bugzilla accessible |
20 |
2. Provide some level of QA for user contributed ebuilds in bugzilla |
21 |
3. Lower the step to becoming a developer |
22 |
|
23 |
Let's handle them. |
24 |
|
25 |
1. Stale ebuilds are often stale for a reason, there is obviously not |
26 |
enough interest to add and maintain them. Not just on the developer |
27 |
side, but also on the user side. If someone really cared enough he/she would |
28 |
go trough the process of becoming a dev. As far as I know most |
29 |
maintainer-wanted stuff just belongs in the category WONTFIX, but the |
30 |
real problem is telling that to the user who sweated on it. I think most |
31 |
of the devs have gone trough a close-reopen cycle on some ebuilds that |
32 |
really added nothing useful to the tree and know how uncomfortable this |
33 |
can make you feel. |
34 |
Then what is a solution to these ebuilds ? I for one would like to see |
35 |
them go upstream, like rpm's and deb's . That would make it clear that |
36 |
these ebuilds are not Gentoo approved, but would provide a starting |
37 |
point for the user who would want to use such a package. I think that |
38 |
was always the main idea when overlays got introduced to portage. |
39 |
Sunrise just lowers the step to get these often mediocre ebuilds, people |
40 |
can get them right now, just not as easy. |
41 |
|
42 |
2. QA for ebuilds is not just a question of making a package build, but |
43 |
also knowing what it does and how it would tie into Gentoo. The fact |
44 |
that some ebuild is semantically correct doesn't mean it is doing the |
45 |
right thing. Very few of the newly proposed ebuilds I handled and eventually committed was actually without major |
46 |
flaws. This was because the submitters lacked specific knowledge of either the eclasses to use and |
47 |
the environment it belonged in. In my case : any gnome ebuild fits in a |
48 |
larger set of applications/libraries that got more complex as time went |
49 |
by, it is not trivial to understand all the interactions that take |
50 |
place. Even Gentoo developers not in the gnome team make serious |
51 |
mistakes in this sense in my experience. Therefore I do not believe that |
52 |
QA for a tree that is as extensive as Sunrise done by a few 'official' |
53 |
developers amounts to much real world quality. |
54 |
|
55 |
3. Although I agree Sunrise may lower the step to becoming a dev, I doubt it will have a serious positive impact on our |
56 |
developer base and as such there is no reason to support Sunrise officially. |
57 |
I think the people attracted to Sunrise development are the ones that |
58 |
fall in the category 'want to be there, but don't really have the |
59 |
time/skills'. Those people will not evolve to real developers; they |
60 |
either will stick it out in Sunrise for a short while or keep to a very |
61 |
small subset of it. |
62 |
My prediction is that Sunrise will see a high turnover of 'developers', |
63 |
either because they are there for one specific package (probably fresh |
64 |
and included in the main tree when mature) or find out they lack the |
65 |
time to really invest in learning the full extent of ebuilding. Also |
66 |
'junior' devs on Sunrise might not take that extra step towards devhood |
67 |
because they got the influence they want, as such we might lose out on |
68 |
devs that never develop beyond Sunrise contributors. |
69 |
As a developer I would not really think of Sunrise developers |
70 |
any different than someone coming 'fresh' to Gentoo developing. I |
71 |
would still require them to work on real bugs for a good while to see |
72 |
their intentions/devotion over time before I would even consider |
73 |
submitting them for real developership. In that sense Sunrise would only lengthen the time a wannabe dev has to spend in the no-mans land between active user and official developer. |
74 |
|
75 |
In conclusion these 3 points come together here : being a dev is not |
76 |
about adding new ebuilds to the tree, it is about maintaining what is |
77 |
already there. Dealing with bugs and users. That aspect of Sunrise is not at all tackled in its goals. What are the longterm |
78 |
prospects of ebuilds in the Sunrise tree ? That is what QA is about, |
79 |
providing a stable base to work on. |
80 |
I do not think that devs who mainly add ebuilds and new packages |
81 |
to the tree are good devs, the real job is maintenance and bughandling. |
82 |
In that sense Sunrise might be giving the wrong impression to wannabe |
83 |
devs. |
84 |
|
85 |
* The rise of project Sunrise |
86 |
|
87 |
Now for the second big point concerning Sunrise : how it came into |
88 |
being. |
89 |
|
90 |
I checked back on the initial announcement, where it Sunrise was made |
91 |
public as an official Gentoo project without any prior discussion. The |
92 |
announcement actually stated 'This is an announcement - No flamewars |
93 |
allowed'. I guess the creators were already aware of the feelings of |
94 |
some other developers on this issue and decided to just go ahead instead |
95 |
of going through the proper channels (GLEP?) and do things as they |
96 |
wished. As we all know this can be very effective, but this particular |
97 |
time one of the largest and longest ongoing 'discussions' in Gentoo's |
98 |
history ensued. |
99 |
If you know it's flamewar material, why do you go ahead |
100 |
so bluntly with your project ? Why not go trough the proper channels and |
101 |
discuss it beforehand in a public place ? |
102 |
|
103 |
Anyway, the project after the initial announcement got a 'temporarily removed' status |
104 |
from gentoo.org . The problem here in my opinion is not so much that the |
105 |
people who support the project needed to defend it, but that people who |
106 |
are more conscious about the project need to prove it is wrong. This had |
107 |
to happen in a mere 2 months where the project has had hardly any impact. |
108 |
If you want to properly evaluate such an extensive project it needs to be given |
109 |
much more time. The project should prove itself before it should be |
110 |
allowed to 'join' Gentoo, not the other way around. I have seen no |
111 |
tangible benefits from Sunrise so far, aside from the fact that |
112 |
developers have left over it and the developer community is seriously |
113 |
divided these days. As such Sunrise has been one big mistake, the |
114 |
possible benefits at this point in time do not outweigh the havoc it has |
115 |
caused. |
116 |
|
117 |
* The implications of sunrise |
118 |
|
119 |
What will Sunrise mean to the general developer ? |
120 |
|
121 |
Again here I can share my experiences with a similar project, the |
122 |
infamous BMG was created with similar goals and turned out to be a |
123 |
serious nightmare for the gnome team. At a certain point in time every |
124 |
bug we got had to be double checked for possible overlay problems. I |
125 |
cannot count the times I had to spend hours on an unexplainable problem |
126 |
to find out in the end that it was caused by BMG ebuilds. This is |
127 |
incredibly destructive for my mood, not to mention the time wasted which |
128 |
could've been spent on real problems. The other side of the medal is |
129 |
that there are false-positives where you think it's BMG, but it really |
130 |
isn't and I can tell you that is not a nice experience for the user and |
131 |
dev alike. BMG was mainly gnome oriented, so a lot of devs may never |
132 |
have noticed such problems, but they surely existed for the gnome team. |
133 |
Another exponent of the BMG tree were the infamous love-sources which |
134 |
also caused inexplicable problems left and right, which may ring a bell |
135 |
with much more devs. |
136 |
|
137 |
In short, from my experience Sunrise will only result in more work for |
138 |
the general developer with little benefits. This may not happen often, |
139 |
but every single time is one time too much. This is can be really |
140 |
demotivating, which is probably the worst thing about it. |
141 |
|
142 |
regards, |
143 |
Marinus |
144 |
|
145 |
-- |
146 |
gentoo-dev@g.o mailing list |