RCBW - #22

With a mini-rush in the week-end, I'm now back on track to the weekly schedule of RCBW; here are this week's squashes:

About this week highlights:

  • In his latest RC-bug statistics, Tolimar has pointed to the RCBW description as inspiring. That's good, I hope more and more people can find it inspiring (in fact, that's why I wrote it). A handful of people doing RCBW is pointless, unless as a side-effect we manage to change the culture of our project in order to be more collaborative on how we reach a release.

  • I've been amazed by how many RC bugs are "stuck" just because no one took care of establishing communication among the appropriate people. That of course happens when the maintainer of the bug has been unresponsive, but some investigations of the bug is usually enough to make the appropriate people (upstream, porters, related package maintainers, ...) aware of it ... and then the bug will "mature" all alone.

  • Josselin has announced the Debian/GNOME bug weekend for the next weekend (27-28/02/2010): an initiative to collaboratively triage GNOME-related bugs in Debian. For large-enough packages (and GNOME fits that definition for quite some different notions of "large"), such efforts are as badly needed as RC squashing: if you liked RCBW, you can't miss the GNOME bug weekend :-)

The sad truth about doc-central.

Hi Stefano,

I just figured it out this morning:

  • Because doc-central did not the respect FHS it got a RC bug,

  • because we lack manpower to spend enough time on this package, the RC bug was solved in a way that interrupts functionality upon upgrade,

  • the positive outcome may be that we discovered that doc-central is not actively maintained; users are suggested to use dwww instead,

  • and that's the sad part: guess what ? dwww does not respect the FHS either, but the DD who opened a RC bug on doc-central did not notice because instead of realy tackling the issue, he just relied on Lintian.

Shall we laugh or cry ? I will not report the RC bug on dwww, I am too disgusted. We are just heating air.

The morale of the story is that Debian as a project has absolutely no management of the priority of issues. Next time somebody makes a mass RC bug filing, I will oppose stronger unless there is a real plan behind, like a release goal, or at least some evidence that there has been some effort to reduce the false negative rate.

Cheers,

-- Charles

Comment by Charles Plessy Tue 23 Feb 2010 01:06:02 AM CET
Re: The sad truth about doc-central.

Charles, let me first of all observe that your comment is pretty much off-topic. It is basically an anti-manoj and anti-"dir-or-file-in-var-www" rant, which has quite nothing to do with RCBW (beside the fact that I did the doc-central NMU: not surprising given that I've performed almost 150 of them now). So, the best short answer for your comment would be "take that to the appropriate place":

  • are you against the MBF? complain on -devel, not here (you did that and your complains were not listened to? it is not my fault, sorry)

  • are you against the proposed solution to that MBF, since it makes things that were working out of the box before no longer working out of the box? propose a solution and work on it. Back then, I've discussed a bit with the webapps apps guys and there were a proposed framework to automatically enable webapps to both respect FHS and work out of the box; guess what, there is a lack of manpower to implement it. How about contributing some instead of ranting?

  • are you against the solution I've uploaded for doc-central? Even better: as you might have observed if you followed the bug log, I've proposed to the maintainer (which replied to my NMU thanking for it) to phase doc-central out in favor of dwww. How about contributing some work there?

And how about the fact that the dwww issue has been overlooked, so what? Please, report a bug against it: it is as much an RC bug as it were for doc-central, and it should be equally fixed.

Finally, let me say that I deplore the style of your post. With RCBW we are trying to show that collaboratively working together towards a release by the means of NMUs is feasible. ... and you take a single issue as a pretest to throw mud at all the project? I am annoyed by that attitude, which is just plainly destructive (a feature way worse than "having no management of the priority of issues").

In a project as big of Debian, everyone can always find a single glitch which bother her and which she cannot stand. What good do you think come from amplifying it the way you just did?

kthxbye

Comment by zack Tue 23 Feb 2010 09:48:05 AM CET
Re: The sad truth about doc-central.

Hi Stefano,

I worked on the doc-central bug: some packages were recommending it, and would have become buggy by its removal. I searched for them, reported the problem, and they got fixed, opening the door to a smooth exit for doc-central.

Then Mats proposed a patch and I helped him to improve its quality. Then a discussion started on restarting or reloading apache (or not). Then you uploaded the patch. I am slower than you but please do not suggest that I did not work on this bug, that is not true.

So far for collaboration on fixing RC bugs...

Cheers,

-- Charles

Comment by Charles Plessy Tue 23 Feb 2010 12:58:40 PM CET
Re: The sad truth about doc-central.

No, sorry, that's just not true either (or else the discussion did not happen in the right place).

Your last comment on 553496 is from 24/01/2010: a month ago. In that message you gave some suggestion on the appropriate file locations and, for whatever else, you say "I can give you some help if you decide to go ahead and save doc-central." Discussion has continued a posteriori, but without any further comment from you. Your suggestions have been implemented in the refined patch by Mats that I've finally uploaded.

So, your contribution is there. As long as it is there however, it has been implemented in the version that got uploaded. If you had something else to contribute, I'm sorry, but it was not visible in the right place.

Frankly speaking though, we are loosing time here: packages get uploaded continuously, if you have some more fixes to contribute you know how to upload and/or report them. You can just NMU the package again, for instance, instead of ranting after the DELAYED NMU period has expired.

FWIW, all my other remarks stand.

Cheers.

Comment by zack Wed 24 Feb 2010 08:49:47 AM CET