| 1 |
On Fri, 23 Nov 2012 12:45:56 +0800 |
| 2 |
Patrick Lauer <patrick@g.o> wrote: |
| 3 |
|
| 4 |
> On 11/20/12 21:57, Alexis Ballier wrote: |
| 5 |
> > On Fri, 16 Nov 2012 09:10:51 +0000 (UTC) |
| 6 |
> > "Patrick Lauer (patrick)" <patrick@g.o> wrote: |
| 7 |
> > |
| 8 |
> >> patrick 12/11/16 09:10:51 |
| 9 |
> >> |
| 10 |
> >> Modified: ChangeLog |
| 11 |
> >> Added: lyx-2.0.5.ebuild |
| 12 |
> >> Log: |
| 13 |
> >> Bump |
| 14 |
> >> |
| 15 |
> > |
| 16 |
> > |
| 17 |
> > |
| 18 |
> > While the bump was fine, please read the damn metadata.xml when you |
| 19 |
> > touch a package you're not used to. Pavel has been doing a very good |
| 20 |
> > job in (proxy) maintaining lyx since years and you do not seem to have |
| 21 |
> > contacted him before doing the bump, which is a bit disrespectful for |
| 22 |
> > him IMHO. |
| 23 |
> |
| 24 |
> I disagree. A fix is a fix, a bump is a bump, no ego involved. |
| 25 |
|
| 26 |
And how much do you know about the particular package in question? |
| 27 |
Did you grep through open bugs before bumping it? How many |
| 28 |
configurations did you test? How many considerations did you make? Did |
| 29 |
you add yourself to maintainers or grepped bugzie for the next few days? |
| 30 |
|
| 31 |
> > If you want to help in having things done quicker because I'm not |
| 32 |
> > always responsive enough, then please do it correctly and ask Pavel to |
| 33 |
> > CC you when he sends me instructions for lyx. |
| 34 |
> I dislike this territorialism. Why add a single point of failure to |
| 35 |
> package maintenance? (What if you or Pavel "disappear" for any reason?) |
| 36 |
|
| 37 |
Are you saying that multiple points of failure are better? |
| 38 |
|
| 39 |
I believe in package maintenance and *responsibility*. What |
| 40 |
responsibility are you taking when you take someone's package |
| 41 |
and silently modify it? People aren't really required to keep track |
| 42 |
of your actions and check whether you just didn't introduce something |
| 43 |
awful to our users. |
| 44 |
|
| 45 |
That said, maintainers usually know more about the package in question |
| 46 |
than you do. The maintainers may be aware of awful bugs which you |
| 47 |
missed and which are the reason for not bumping the package. If that's |
| 48 |
the case, your 'trivial bump' may have just unleashed destructive |
| 49 |
issues for our users. |
| 50 |
|
| 51 |
Moreover, the maintainers may have a few changes stashed for the next |
| 52 |
bump to avoid people rebuilding the package unnecessarily. If that's |
| 53 |
the case, you are either forcing a second rebuild for our users |
| 54 |
(through requiring the maintainer to go with a revbump) or delaying |
| 55 |
those changes even more. One way or the other, our users lose thanks |
| 56 |
to you. |
| 57 |
|
| 58 |
That said, I believe that a dev disappearing and delaying the bump |
| 59 |
for a few days is not something tragic. Of course, unless security |
| 60 |
issues show up but these can't be solved cleanly with a bump anyway |
| 61 |
if it's a stable package. |
| 62 |
|
| 63 |
What is much worse, a single impatient developer bumping a package |
| 64 |
and taking no responsibility for it. Now imagine that your bump could |
| 65 |
have caused serious issues. These issues were reported quickly |
| 66 |
to bugzilla but since the actual maintainer was away, nobody noticed |
| 67 |
them. |
| 68 |
|
| 69 |
And I think I've said something similar already. Simply said, these |
| 70 |
are a few Gentoo developers who take their work seriously. Try to |
| 71 |
respect that. |
| 72 |
|
| 73 |
-- |
| 74 |
Best regards, |
| 75 |
Michał Górny |
Archives