Today's reports of apt-cron on both my etch home server and sid laptop
were interesting.
Because of the very recent point release for Etch, I had more packages
to update on my etch server than I had on my sid laptop.
Probably the first time in years that this happens. It apparently
means that developers have, this time, been very careful avoiding
uploads to unstable during the deep freeze.
It has been a quite long time since I blogged about D-I.
As everybody now probably knows, D-I RC2 for lenny was released on January 31st, which now allows releasing Lenny.
While the last work for Lenny is going on, I noticed that, during this week, one of our "new" translators completed the first two parts of Debian Installer translations...which is the key for a language to be "activated" (with translations flowing to packages).
So, this morning I activated Kazakh in Debian Installer, which means that we will be painting a significant part of the world in red (dear readers, were'nt you missing "my" neat maps?).
Of course, this is way too late for lenny, but given the high activity of that team now, I don't doubt that Kazakh will be part of Squeeze D-I.
So, the count is now 64 and I really have to move my a** to shake those 16 translators and teams that are listed in "work in progress". Most of them are indeed silent, unresponsive, or just sleeping..:-)
We're apparently starting an "update thy hackergotchi" meme with Andrew
Pollock and Daniel
Burrows updating their faces smiling to the world (both being much better than their previous version, indeed).
I really should update mine, too, as it is:
Christmas holidays in France are not a good season for runners: too much champagne, foie gras, cheese, wine and food in general.
So, while visiting my mom and sister in the region of St-Etienne last week, I wanted to keep my sports pace to a sufficient level as we're now 3 months ahead from this marathon I've been crazy enough to apply for.
That went good: 3 days after my first run over 30km, my brother in law convinced me to enjoy the big snow fall in Le Bessat, the ski station of my young years, 30 minutes away from the city of St-Etienne.
That was the second time ever I'm doing croos-country skiing in skating technique. But, being supposedly used to skiing in general (20 years of ski are supposed to help), we went on the longest and hardest track: 16km and over 200m positive slope....which I completed in less than 2 hours, so a lot more than the same distance by running. Cross-country skiing in skating technique is awfully hard, particularly when you have no technique at all..:-))
To complete this, I went running for about 12 km on New Year, 3 hours after partying for New Year's Eve. Interesting way to eliminate the weight surplus..:-)
And, finally, today, I went out for a 19km run in the frozen country around my place (-8 degrees) for a wonderful discovery of woods covered with ice and the climbing of the highest "mountain" in the Paris area (280m height, still !) and a great view of the neighbourhood in white, under the sun. These are the joys of running in winter, really.
Nicola Benaglia reported bug
#510000 on Sunday December 28th.
As bug #500000 was reported as of Sep 24th 2008, this marks a
considerable slowdown of our bug reporting rate: between bugs 400000
and 500000, it always took us nearly exactly two months to have 10000
bugs reported. Here, we needed more than 3 months! This is a 50% drop
in the bug reporting speed. I think we can conclude this is a very
good illustration of the considerable drop in the development of
Debian, while everybody is awaiting for the release of Lenny.
You still have a few days to correct your bet for bug 600000 date.
Ah, and the very same day, the result for the very controversial vote
about firmwares has been published. Even though it is very hard to interprete, I think that the very very low rating of option 1 (worded as "Reaffirm the Social Contract" but indeed meaning "let's be Free Software Zealots and nitpick every single bit of firmwares so that we can be sure to never release Lenny") clearly shows that Debian developers still have their feet in reality.
So, now let's release the bloody thing...and prove the world we still exist...
Last year, on December 24th, my Christmas gift to me was running my first half-marathon (alone in the woods, just as I like).
This year, I couldn't do less and I offered myself another wonderful gift: my longest distance ever: 33,8km.
That was a wonderful run on a sunny winter morning, around zero degrees, around the "Etangs de Hollande", a very well known spot in the Rambouillet Forest, then north to Montfort l'Amaury (the run is the red line on the map...the blue line is more or less my 2007 first half-marathon).
33,8km in 3h20 minutes, by the way, for those of you who are marathoners. Just above 10km/h (the end was very slow as I was dead tired!).
Really the best gift I could do to myself. Great Christmas.
Samba in unstable is 3.2.5 currently, while upstream published 3.2.6 a few days ago.
This is mostly because 3.2.6 has too many changes that make it unsuitable for lenny, which will have 3.2.5. As we want to keep the "update through unstable" path, we can't upload 3.2.6 to unstable.
Simultaneously, we are using experimental for 3.3.* releases. Upstream recently published 3.3.0-rc2 which I uploaded yesterday to experimental.
So, we don't have any easy way to make post-3.2.5 releases available to our users.
I briefly imagined using backports.org for this, but:
deb http://pkg-samba.alioth.debian.org/packages/ ./ deb-src http://pkg-samba.alioth.debian.org/packages/ ./The packages are signed with my own GPG key, which you need to add to "trusted" keys for APT" (be sure to read http://wiki.debian.org/SecureApt first). They are numbered the same way they would be if they were on backports.org. These packages are maintained in the Samba packaging team SVN. Please take this as a commitment to try providing up-to-date packages for lenny during its release cycle. Of course, the security support will be done on a best effort basis but should at least not be worse than upstream's (which is known to be good and very reactive). I hope that samba users will enjoy....and if some of you could join the team and help triaging bugs, that would be appreciated..:-)
Samba in unstable is 3.2.5 currently, while upstream published 3.2.6 a few days ago.
This is mostly because 3.2.6 has too many changes that make it unsuitable for lenny, which will have 3.2.5. As we want to keep the "update through unstable" path, we can't upload 3.2.6 to unstable.
Simultaneously, we are using experimental for 3.3.* releases. Upstream recently published 3.3.0-rc2 which I uploaded yesterday to experimental.
So, we don't have any easy way to make post-3.2.5 releases available to our users.
I briefly imagined using backports.org for this, but:
deb http://pkg-samba.alioth.debian.org/packages/ ./ deb-src http://pkg-samba.alioth.debian.org/packages/ ./The packages are signed with my own GPG key, which you need to add to "trusted" keys for APT" (be sure to read http://wiki.debian.org/SecureApt first). They are numbered the same way they would be if they were on backports.org. These packages are maintained in the Samba packaging team SVN. Please take this as a commitment to try providing up-to-date packages for lenny during its release cycle. Of course, the security support will be done on a best effort basis but should at least not be worse than upstream's (which is known to be good and very reactive). I hope that samba users will enjoy....and if some of you could join the team and help triaging bugs, that would be appreciated..:-)
Dato points me that the way I mentioned my vote is confusing and he's right.
By writing that I voted 5462371, I wanted to say that I ranked option 5 as first, option 4 as second, etc.
However, one could understand that I ranked option 1 as 5th, option 2 as 4th, etc.
...which is indeed the most common way we talk about our votes....
So, let's make it clearer. I voted: 7452136.
So the first option (freeness bigotry) is ranked below Further discussion.
I'll re-consider it in terms of strategy but I consider it honest: this is how I rank the various options altogether. I did not try to mke this vote "strategical" but I agree that this needs to be deeply thought. The issue is really important and can even be seen as a trap, yes.
I am somewhat under the feeling that the Debian project is currently
facing a critical moment of its history.
The development momentum has slowed down a lot during last months
(approximately since Debconf and even more since October.
I feel just like we're facing a moment where either we release Lenny
(which is technically a great progress over Etch and, being a little
bit more than 18 months after Etch, well suited wrt "modernity")....or
we delay it for several months and will suffer more and more from the
long freeze process (remember sarge). The momentum is then very likely to continue moving elsewhere (should I say Ubuntu or did everybody understand already?).
And now we have to vote about firmware blobs, a proposal that popped up because a few (*very* few) vocal folks who (in my opinion) live on a planet that has no connection with the real world, think that we should triple check the freeness of each and every bit that is likely to be run on a Debian machine.
Moreover, the proposal is worded and was pushed in a way that all (but one) options that would allow us to release lenny will need a 3:1 majority, which is obviously very hard to achieve (and as, there are several such proposals, the vote will be spread out among all of them).
So, those folks who make the release of lenny their priority are now facing an interesting tactical choice: how the ..... should we vote for this to happen?
I personnally did choose to vote this way: 5462371. Yes, I ranked
"reaffirm the Social Contract" after everything, which makes me
sad....and I blame the play on words by the zealots who pushed that
proposal for forcing me to do this. I think this is the best "tactical" choice for those (and I hope we are a strong majority) who still make the *release* of the best possible and most possibly free operating system their priority.
Believe me or not: if we don't release lenny soon, we will die slowly and rot peacefully. We will rot being 150% free and proud to be 150% free on our little planet, outside the solar system and the real world. *This* is the choice we are facing.
Ah, and before I forget: film at 11, of course.
PS: no comment system on my blog (I don't want to deal with spam and put complicated systems there). Please comment in public in your own blog. I very much prefer this over private comments which will be mostly ignored.
Waking up this morning, I noticed a low traffic in my incoming mail in
"Debian work" mailboxes (PTS mailboxes, personal email, bugs, -release, etc.).
Then I read -devel and understood. Everything is there. We have a new flamewar (apparently about NEW processing).
It's interesting to see that the closer we get from a release, the biggest flamewars we are doing. Sounds like everybody is boring (have you notived the strong drop in the bug report rate?).
Crap, there's a lot of work left to do, to release the damn Lenny:
Interestingly, the just finished Extremadura QA + ftpmaster + i18n meeting also turned into a quite strong Bug "Smashing" party.
Several participants were digging around the list of RC bugs affecting both lenny and unstable. Most often, it turned out to many bug downgrades (for many of which I'll take the blame).
It's "funny" to see how sometimes our bug reporters tend to inflate their reports, most often because they feel that "bug should absolutely be fixed", but completely forgetting about the definition of 'serious' or 'grave'.
I also feel that, in some way, several maintainers begin to be quite demotivated by the length of the process (which I personnally consider as being caused by several factors: decreasing (wo)manpower, haircutting about freeness of every single byte on Debian CD's or even shyness wrt correct actions needed to deal with bugs). And several of these bugs just needed a very simple action by the maintainer.
And, therefore, just because we were about 15 people with a lot of time to devote to Debian for 3 days, we ended up digging in that pile of mud....removing most of the dust and sometimes very debatable RC bug reports, and finally having that count of bugs affecting both lenny and unstable go down again.
This comforts me in the feeling that an aggressive approach is a good key to bug squashing. After all, if I'm sometimes too aggressive and inappropriately downgrade an RC bug, I'm sure that someone will correct me (and, anyway, I'm not that mad: sometimes I seek for advices). Ah, and "aggressive" also sometimes mean proposing to tag some bugs "lenny-ignore" (and, that needs to be very convincing with our release managers, believe me).
So, what about you? Have you already smashed an RC bug today? If not, do it. Kill those nasty beasts, they deserve it. No mercy.
It's interesting to be here in Casar de Caceres for the joined
i18n+QA+FTPmasters Extremadura meeting.
Today, I've been sitting all day long in the biblioteca, which is used as hacklab, and just watched all great people around playing around with the concept of "tdebs".
That weird beast can be described in two words as a concept of deb packages that would be able to distribute translation data in a more independent way then "regular" deb packages.
Don't smash me as the concept is currently being twisted around in all
directions by the nice folks here, to bring it to a great concept well
suited for "side" material of our packages.
To get an idea, you can read the original concept that popped up during the 2006 i18n Extremadura meeting, but got transformed drastically by ideas exchanged at Debconf 8
All this is made possible because the concept is now a very important issue for Emdebian and is strongly pushed by Neil Williams, whose nickname shall now be "MisterT", by general agreement of the folks here.
What's really interesting right now is to see live discussions happening all the way round and see the concept slowly emerging from this. Something we could probably never do in regular mailing lists discussions or on IRC. We should really grant tons of rewards to the Extremadura region, and particularly out tireless C sar who again organised that meeting.
It's still hard to say what will emerge from this and we willprobably need some more such live meetings to have real implementations of these tdeb beasts, but this is really how I like to see free software development going on.
And now people begin to think about coding stuff and prepare patches, which is a good sign, in general.
I just registered for the Paris Marathon. There were only 1350 seats left out of 37,000.
Finally, I decided that it's more likely that I would regret it if I don't try, so I decided to go for it. We'll see.
Again, my offer to host Debian marathoners for that week-end still stands.
Last week-end, I went very close to the 2/3 marathon mark by running 27 km again after I did one such run on Nov. 11th. So, indeed, I ran 27km twice with only 4 days in between (and I ran also 7km during two of these four days).
That last 27km run was like a November dream: over 20km in woods, with
muddy paths full of dead leaves, about 200 meters positive slope (in four climbing sections), and
I didn't finish completely dead. The rate was 5'55"/km (9'37"/mile,
Dirk). Tuesday, I didn't go under the 6'/km mark, so that's an
improvement. Actually, this is mostly because this time, I wasn't
originally imagining I would go for so long, but more an
half-marathon.
Ralf Treinen and No l K the (who ran together at Cologne marathon the
week-end before) are trying to convince me to participate to the Paris
marathon next April. It seems indeed more and more likely that I
will. Other Debian runners, what about meeting next April in Paris and
do it? I can host folks in my house, of course.
All this, the very same week-end Bdale got
his level 3 in rocketry. Great success for the 61 folks.... Cheers!
The D-I team finally released Debian Installer RC1 for lenny. That wasn't
an easy process and it revealed some miscommunication and rampant
conflicts inside the team.
That will needed to be solved for squeeze development but, as of now, we (the very few of us still working on D-I development...) focus on preparing an RC2 release.
That release is meant to be out ASAP. It is mostly meant to deal with recent kernel updates and definitely not add new features. Even bug fixing will probably be limited to what's considered major blockers.
That's another step towards the overall release of lenny, I think. Now
we should probably all focus on remaining RC bugs, which count seems
to have stopped going down during last weeks, which is IMHO
worrying. But, of course, let our good release team deal with all this
first (which probably involves dropping some crappy packages, imho).
Disclaimer: this post talks about Debian *Installer*....not overall Debian.
The D-I team and the CD team have created images of Debian Installer Release Candidate 1 for Lenny. They are available at http://cdimage.debian.org/cdimage/.lenny_rc1/.
Please test them out and report problems through installation reports as described from the Debian Installer web page. This is really important as this is very probably how the real released Lenny installation media will look like.
Debian Installer is short of testing. So, yes, we need *you*.
And, again, if you spread the word, don't talk about "Debian Lenny RC1". This is no such beast (even though this will look like a Release Candidate of Debian as the rest of the distribution is of course begin prepared for the release).
The 'official' name is "Debian Installer for Lenny, Release Candidate 1".
...for maintaining bacula and maintaining it well.
I adopted Bacula as backup solution for my local machines, at home, after I went very close to a complete disaster on my home server, last summer.
As of now, my home server and my laptop are now backed up with it. The backup is done on a dedicated disk of the server (until I setup a dedicated server, probably in the house's ceiling).
And things are just...working. Of course, Bacula is a fairly compilcated beast but, well, as I managed to make it work, I consider that anybody can do it.
This morning, I finally setup the bacula graphical console on my laptop. As the laptop is running unstable, with Bacula 2 packages, I indeed had to upgrade bacula packages on the server (which runs Etch, where Bacula version is 1.*).
I just went to backports.org, found backported packages there, kindly prepared by John (and Rhonda, apparently).
One "apt-get install -t etch-backports" after, I'm just running Bacula 2.
John, you deserve a full tank of beers... I'll think about it next time I happen to come in the middle of Kansas..:-)
... as people at work found this one.
Next.
