August Penguin 2008

In the last two months I’ve been less active regarding free software as life keep me busy with other stuff. Nerver the less I’ve been able to help with orginizing August Penguin - the annual Israeli free software conference (happening the 7th time in a row).

The conference takes half a day and occurs on the first Friday of August. This year we’re back to a two track conference with a tehcnical and non-technical tracks. Each track has 5 lectures, 30 min. each.

Non technical track:

• “Open source by Microsfot” -  Microsot Israel
• W3C in Israel - Ori Idan
• Building a wordpress community - Blogli staff
• Openoffice.org, the community side + Update on free software in education - Lior Kaplan & Nadav Kavalerchik
• Open Moko - Doron Ofek

Technical Track:

• Bidi in free software - Shoshana Forbes
• Django - Meir Kriheli
• Blender - Shlomi Israel
• Hocr - Kobi Zamir
• Drupal - Lior Kesos

This year with also have ISOC-IL, Microsoft (yes, you’re reading correctly) and Sun as sponsors (and a few smaller companies). Cool to have the “big boys” helping.

As I found out more and more open source people from around the world in Israel, I thought it will be a good idea to write about the conference and invite them. Please be aware that the lectures are in Hebrew. But you can also come to meet the community and talk with the people or take part in our GPG signing party. More info (in Hebrew) at http://august.penguin.org.il and in Facebook’s event info.

Hosting Debian people in Israel

This week was the third time this year I got to host people from Debian in Israel. I find it very interesting to meet people from the project in a totaly different environment from the usuall confrences.

Having a relax talk (and maybe some good food) is a better way to know people than the quick talks I’m used to from more formal meeting. And having the chance to show them a city or the country is of course something I can’t do abroad (:

So, if you’re in the area - let me know. I can’t provide a place to crash, but can still help…

p.s.

The invetation is open to any free software related people… don’t be shy.

How many projects can you identify ?

(taken from http://www.freebsdnews.net/2008/03/11/open-source-community-united/)

To separate or not to separate (the file systems)?

I have to handle some machines of commercial places which has a different standards about thier installations. The main difference is about creating a lot of file system instread of the normal few familiar in most distributions (/, /usr, /var, /tmp, /home).

While they create a FS for each product installed on the machine (a habit taken from UNIX when they didn’t have any meaning of installation other than copying files), they also separate the variale files (e.g. logs or other very active files) of each product.

I’m trying to change these standards to be more close to the FHS (and use the advantenges of RPM/DEB), but one of the main questions I get is what will happen when the FS reach 100%. When everything is separated one product can’t affect another, but that costs with a lot of sysadmin overhead. Leaving everhing variale in /var makes things easy, but hold some risk.

I’d be happy to hear what other sysadmins chose to do…

The 2nd Debian Openoffice.org bug triage has ended - 300+ bugs processed

The 2nd bug triage began with 285 bugs, which grow to around 310 bugs, since during the 4 months of the triage more bugs were reported.

So, I’m glad to announce the beginning of (my) 2nd openoffice.org bug triage. This triage targets 285 bugs out of 330 open oo.org bugs. The selected bug are ones which were reported against versions lower the 2.3.1 or have no version information at all.

At first, the triage advanced quickly with 21% of the bugs processed during the first 24 hours and 40% during the first 10 days. This left a long trail of bugs to handle manually.

The triage began for version 2.3.1, but was delayed to let version 2.4.0 to propagate into Lenny, as that was Rene’s preferable version to release with. When the triage resumed in late April, I had a 2.4.0 available in all of our repositories (sid, testing and backprots). Which meant that everyone could verify their bugs against the same oo.org version. Bug triaging heaven (:

Two thirds (66%) of the bugs were closed in this triage, most of them after a check from the submitter or myself and a few after the submitter was unresponsive and I couldn’t verify the bug myself.

Verifying the bugs myself is the most demanding part of the triage. Reading each report and understanding the problem takes a lot of time. Reproducing the problem usually takes less time than understand it. That’s way have test cases attached to bug reports are so important.

In several cases the instructions were so simple I could just open the file, check something and verify the bug. That was done in a robotic manner or in a monkey like one. These kind of attachments enables a lot of people to help with triaging, even if the lake the technical background of the program in subject.

So after closing 200 bugs and verifying around 100+, the oo.org bug count is set to 140 outstanding bugs, and 70 bug being forwarded to upstream. Except from about 10 bugs, all of the remaining 130 bugs are relevant for version 2.4.0. No need to wonder if the bug still occurs before reading the report…

I’m not going to DebConf 8

After 3 years in a row, I’ll miss DebConf this year. Seems that I shouldn’t have filled the reimbrusment form too late in the night, as I mixed the the “unable to pay” amount with “able to pay”. That’s a 1500 USD difference…

Flights across the planet (Israel -> Europe -> South America) are expensive. Further more, their costs went up, meaning that even if I got the all the reimbursement I asked for, I would still need to add money (which is a bit of a problem as I’m looking for a job).

I hope I could virtually attend the conference thought the lectures’ video streams, but that can’t be compared to actually attend the conference.

Oracle and ncompress

I’ve been approached by someone installing Oracle Enterprise Manager with a problem of the installation fails due to not finding the compress binary. It seems that some of the Oracle products relay on the ncompress package for compressing.

While compress might be the UNIX standard for compressing, I’m curious why didn’t they changed the product to use gzip. Or at least to check whether ncompress is available and than use it. But failing because of that seems to much to me.

This isn’t the first time I’ve challenged Oracle’s dependencies, which I think sometimes aren’t minimal as they should be or complaint with common distribution standards.

Why aren’t our users familiar with backports.org ?

As part of the openoffice.org bug triage I asked users to test their bugs with recent versions of oo.org, and referred the users of the stable branch to backports.org.

I got a lot of responses from users saying they didn’t know about the backports repository. For most of them that’s a time saver in the process of getting newer software. Some were glad to upgrade and to confirm their bugs with the recent version.

I’m wondering how should we publish this (unofficial) service? Having results from backports.org appear in packages.d.o was a good start. Maybe adding a link from the PTS as a reminder to the maintainer? (but this don’t help the users).

Any ideas ?

Ubuntu’s reportbug-ng sends bug reports to Debian’s BTS

I checked the Debian firefox bugs page and noticed a new bug was reported against firefox. Checking the bug report showed the bug is new as it reported against version 3.0b5. As there’s now firefox package in debian testing/unstable that was weired enough.

Reading the bug report revealed these details:

Package: firefox
Version: 3.0~b5+nobinonly-0ubuntu3

and

Debian Release: lenny/sid
500 hardy-updates us.archive.ubuntu.com
500 hardy-security security.ubuntu.com
500 hardy us.archive.ubuntu.com

I mailed the user to see why does he reports Ubuntu’s bugs to Debian. He replied that he’s using the reportbug-ng program and wasn’t aware that his bug have been sent to Debian.

I check the reportbug and reportbug-ng in Ubuntu and found that reportbug has the needed patch for Ubuntu:

diff -pruN 3.39/reportbug.conf 3.39ubuntu3/reportbug.conf
— 3.39/reportbug.conf 2004-12-06 13:59:04.000000000 +0000
+++ 3.39ubuntu3/reportbug.conf 2007-10-25 07:22:47.000000000 +0100
@@ -6,7 +6,7 @@
# severity normal

# BTS to use
-#bts debian
+bts ubuntu
# See ‘reportbug –bts help’ for a current list of supported BTSes

# Submission address: default is ’submit’

but since hardy reportbug-ng package in Ubuntu is identical to the one in Debian.

$ rmadison -u ubuntu reportbug-ng
reportbug-ng | 0.2007.10.30~feisty1 | feisty-backports/universe | source, all
reportbug-ng | 0.2007.06.27 | gutsy/universe | source, amd64, i386, powerpc
reportbug-ng | 0.2007.10.30~gutsy1 | gutsy-backports/universe | source, all
reportbug-ng | 0.2007.10.30 | hardy/universe | source, all
reportbug-ng | 0.2008.03.28 | intrepid/universe | source, all

Doing more checks reveals that this is a known problem in Ubuntu and is already reported at Bug #175508. This was also discussed three weeks ago on “Ubuntu Open Week - Reporting Bugs“.

A quick look at the reportbug-ng code makes it clear that reportbug-ng will require more work to support the Ubuntu BTS. There were also some comments about the configuration for Ubuntu in reportbug, as it just send the reports to ubuntu-users mailing list and not to the real BTS at https://bugs.launchpad.net/ubuntu/.

I’m still not sure what to do about the bug. On one hand, that package originates in Debian, and the bug is probably relevant. On the other hand, the cause might be a change from Ubuntu. In any case, I
don’t see the user following our BTS, and I don’t way to fixed package (if and when created) to propagate to Ubuntu.

What can Debian learn from Fedora about recruiting new people?

I’ve been visiting the fedora website in the last couple of days. One thing caught my interest, and that’s the “join Fedora” link just bellow the “get Fedora” one.

I clicked on it and got to this very simple “Join Fedora” page. The appealing part is the large icons which roughly lists the main ways to contribute to Fedora:

• Content Writer
• Designer
• People Person
• OS Developer
• Translator
• Web Developer or Administrator

Clicking etch icon gives you a description of relevant skills, related teams and typical tasks of this role. This is very useful for people not sure about what can they do or where exactly their skills are needed.

On the Debian’s website we have a “help debian” page which lists very similar functions the user can help with. The difference is that we list them at text which is less appealing than the Fedora’s icons. Fedora does have similar text to Debian, but it is organized into roles instead of suggesting everything to everyone.

I also think there’s a semantic difference with the term help and join. To me joining a project sounds more strong than helping it. Probably because joining something makes you a part of it, while helping does not. Although in the end both term have the exact same meaning in for the project themselves - users getting involved.

Openoffice.org has a big “I want to participate in openoffice.org” text in their font page, which like Fedora leads to a set of defined roles. Same thing in Ubuntu with their “Get Involved” page. It is important that each role page will have links the to tools people need in order to start contribute.

I don’t have the required graphical skills to do such icons, but I’m willing to create/edit the pages on the Debian website. I’ll be happy to hear comments before I approach the debian-www people.