This (extended) weekend features: pkg-phototools

• Update libgphoto2 which is used (surprise!) by gphoto2 and others to access digital cameras. Some hours after 2.4.1-1 got uploaded, 2.4.2 got released. Got to get back to it soon
• Prepare an SVN snapshot of hugin, which is a panorama sticher, using libpano13 from panotools. People are so active with their GSOC students that the long-awaited 0.7.0 version isn t released yet, hence the snapshot. Testing is very welcome: binaries for i386 and amd64 will be available on people.d.o until the upload. Thanks to Julien Valroff and Johan Kiviniemi for their patches, and to the NMUers who took care of hugin in the meanwhile.
• Bug triaging for both of them has been quite missing lately, but hopefully that ll get fixed soon.

collab-maint

• Fix blender, which could have migrated, if ftgl wouldn t be still missing on alpha being optional:uncompiled doesn t help. It started to FTBFS: bug #490317. Against crappy upstream build system, strace -v to the rescue (-v is needed to get all variables, so that one can figure out why that execve call failed). Here s the fixing-an-RC-bug dance for Stephen Frost, as promised: \o_ /o/ \o\ _o/. The Add > Text crash (bug #487890) got fixed too, although I didn t get any ACK from upstream. Oh, and I know about this tiny bug (which appears during the byte-compilation phase). I m not very keen on trying to get it fixed myself and possibly breaking stuff, especially when upstream doesn t care, and when the following commits are looking even more messy.
• Update sunflow, a rendering system for photo-realistic image synthesis, it can now happily depend on better Java packages now that openjdk-6 made it through NEW.

More to come

• A new graphviz upstream bugfix release is available.
• As written above, a new libgphoto2 upstream release is available.
• A new module, python-ssl (bug #453101), might be nice. Python 2.6 got a real ssl module and this python-ssl module would be a backport of this new ssl module to Python versions 2.4 and 2.5. Interested people can compare the current ssl module API against the one available in the new ssl module.
• And the sponsor queue is growing, too.

I have received a brand new USB card reader. At first sight, everything was working well “out of the box”, but when I have checked the log files, I could see the following lines repeated very frequently: I repeat everything seems to work OK, but I can’t stand this kind of messages, especially when it deals with storage (by experience, everything works well during the tests, you simply ignore error messages, and you loose your data the first time you really use your hardware!).
After some googling, I have found this issue is hardware related. A lot of people seem to get hit by this. Some have even reverted to USB 1 (UHCI)! Further googling pointed out this article. I have tried to disable USB auto-suspend, and changing the idle-delay time (default is 2 seconds). But it doesn’t help. Just for reference, this can be done like this:
echo -1 > /sys/module/usbcore/parameters/autosuspend Note the following part of the documentation:
The idle-delay values for already existing devices will not be affected To make the change permanent, you need to append usbcore.autosuspend=-1 to the boot parameters (USB support is compiled into Debian kernels).
More information on: http://www.mjmwired.net/kernel/Documentation/usb/power-management.txt I can also see the following message at boot time: According to this recent discussion, it seems it is harmless. I have however noticed that unplugging the card reader and re-plugging it after the system has booted make these log entries disappear. Does someone have an idea? Note that I use the latest Debian 2.6.25 kernel:

A small memo in case it happens again… Maybe this can also help others… I have made a typing error when incrementing a zone serial number this morning (I though we were in year 2009!). I realized this error after the transfer was made to my secondary server, and I couldn’t figure out how I could fix this: if the serial number on the slave server is lower than the serial number on the master, the slave server will attempt to update its copy of the zone. I have read some websites explaining one should add 2147483647 (2^31-1) to the number, reload the zone, let the slave get updated and revert to the appropriate serial number. Unfortunately, this solution hasn’t worked for me. I have read in RFC 1912 (section 2.2 - SOA records), that the system was planned until year 4294. I have henced updated the number of my master server to year 4294123199 (ie. the last possible serial number), reloaded the zone. After the transfer was propagated to the slave, I could simply revert to a standard serial number.

I am in the process of moving all my services (including this website and my mail server) to a new host (actually a virtual server). Until now, almost all of them were (hum, are… see below) hosted on my home server on a DSL line. The aim is to ease maintenance, reduce power usage, and avoid any risk of fire when not at home. Most of the websites and the mail server were moved to the new virtual server at the beginning of the week. I had to postpone my backup configuration to deal with the OpenSSL vulnerability. Unfortunately, one of the filer of my new hosting company crashed on Thursday and all the data of my virtual host have been lost. The server is still unreachable, I had to move back to my home server on Friday. This lead to some outages on all my websites in the meantime. Sorry for any inconvenience. I have also lost a few e-mail (which I have received on the new host, but not yet read, ie. not yet in my MUA cache). If you sent something I haven’t yet answered to, please send it again. I hope I will be able to work on the virtual server again next week-end.

As some of you might have noticed, I have uploaded a bunch of new packages to my experimental repository, all of them related to digital photographic workflow. Most of them are based on existing packages, but built from a VCS snapshot to benefit from newly supported hardware and new features.

• ufraw was updated to support my Sony DSLR-A700 camera - packaged based on the latest release available in sid.
• hugin package in unstable is outdated - I have based my package on Debian PhotoTools member Cyril Brulebois’s work. It seems that some legal questions have been raised, preventing the package to be updated in the official archive.
• autopano-sift fixes FTBS and other bugs in Christian Marillat’s package and proves working well with hugin. I will e-mail him a patch with the amendments I have worked on.
• enblend needs to be built from VCS to work with current hugin snapshots. The package is fully based on the package available in sid.

Though I use these packages on a regular basis, they haven’t been thoroughly tested as official packages would be, and might still contain bugs or errors. Please report any problem you might have. I will try and update these packages when time allows it. VCS snapshots might be unstable, or even non-usable. As usual, use these non-official packages on your own risk. Also note that these packages are currently only available for amd64. I will eventually build i386 packages if requested.

I have just upgraded one of my system from Etch to Lenny. The process was very fast and I had no problem, except hal which refused to start because of SELinux. This is now fixed thanks to audit2allow. I have also lost some time due to the fact I had forgotten to install the headers of the upgraded kernel, preventing me from building the module of my wireless card.

It seems like my application to become Debian Maintainer has been accepted. This means that in the near future, I will be able to upload myself the updated packages I currently (co-)maintain.
All of them still need to be uploaded by my current sponsors with the DM-Upload-Allowed field. Again, many thanks to Micah Anderson, Christoph Haas and Lo c Minier for their advocacies and, in a more general way, for the help they bring on a day-to-day basis.

Some time ago, I had explained why I thought would never use network-manager, especially because of the gnome-keyring password which should be typed in at every new session. The last evolution-data-server upload introduced a gnome-keyring binding which forced me to inquire again on how to avoid this annoying gnome-keyring password. This led to the libpam-gnome-keyring which allows the keyring to be unlocked automatically at login from the display manager. This is great! I am now using network-manager on my laptop! I am still puzzled about the fact that there is no network connection until someone logs in, but I guess I will live with this, especially with a desktop machine.

I have brought some work with me to be able to catch up during the week-end. Usually, I take a laptop running Windows which allows to connect to our network through a VPN the credentials of which I am not in possession of. This time, I just want to make some analyzes on spreadsheets (created on MS Excel). I have tried to use OOoCalc for the first time in “work conditions” - ie. I expect a little bit more than the basic features, which was source of disappointment. Autofiltering feature does not allow wilcards in the conditions. Even advanced filters do not seem to allow them. I want to select columns which contents end with ‘00?’ (ie ‘00′ followed by any unique character), and I can’t. I have read some posts talking about regular expressions in filters in the old 1.x series, but can’t find anything related in the documentation of the current series - regular expressions seem to be used only for the ‘find & replace’ feature. Moreover, you cannot set more than 3 conditions in autofilter, which is not a lot (I am not sure that previous releases of MS Excel were better). I am fully aware of the advanced filters which do allow up to 8 conditions, but I must say they are not easy to use on a day to day basis. Filters are mostly considered as visibility filters, which means if you apply a filter and use the ‘Fill down’ feature, hidden cells are also filled in. At least it is not the case with copying & pasting: only visible cells are copied. These limitations (bugs?) as well as the old existing bug regarding the decimal separator make it impossible for me to use OOoCalc for my work. I guess I could easily achieve to do my work thanks to OOoBase, but the data I want to extract need to be used and amended by collaboarators who exclusively use MS Excel. Are there any advanced users able to work around these issues?

I have had some fun during these 2 rainy days thanks to VirtualBox. I have first tried the latest Fedora 8 (upgrading from the previous release). Though the upgrade was painless, I was stucked with the Linux Guest Additions which weren’t working, due to SELinux being enforced (by default or by mistake, I am not sure). I had even opened a bug for this issue… Then I have tried OpenSuse, which also needs tweaking so that the additions run ok - the solution was on the VirtualBox forums (disabling concurrency for init scripts by changing RUN_PARALLEL=”yes” to “no” in /etc/sysconfig/boot file). Well, I haven’t extensively tested these distributions, but I particularly like the new Nodoka theme used by default in Fedora. I have even begun to work on Debian packages for both the GTK+ and the Metacity themes (I know there is already someone working on this, I will e-mail him to propose a co-maintenance if he accepts). On the other hand, I was very surprised that OpenSuse uses GNOME as default WM (maybe it isn’t new, but I haven’t seen a Suse desktop for years). Their gnome-main-menu is really nice, but really unusable for me! I guess I will try more distros in the coming weeks to compare with Debian - for now, I get the feeling I have made the good choice, except maybe for the default artwork, which is hopefully improving…

I had been using ndiswrapper with a Netgear WG511 PCMCIA card^[1] without any problems for a while, then have begun to get kernel crashes, which were more and more frequent. I have thus backported to Etch the 1.47 version available in Sid, which *does* fix my kernel crashes. However, this newer version has brought new problems: with a “high” network load, the module just crashes silently, and has to be reloaded so that the network connection can be enabled again. I have packaged ndiswrapper 1.48 as the release notes let me think my issues were fixed (”Disassocation with wpa_supplicant is fixed”), but the problem is still there with this latest release! I have finally managed to get things work correctly thanks to this post: I use Debian linux-image, ie. without preemption enabled, and I can now confirm the still-to-be-released 1.49 version fixes this issue. I have uploaded 1.49-rc3 packages to my unofficial repository just in case someone has the same issue.

I have worked almost all the day to update my mail server configuration to use Dovecot LDA and the Sieve mail filtering language instead of maildrop. This was very easy thanks to Christoph’s ISP-style Email Server with Debian-Etch and Postfix 2.3 how-to.
(more…)

Thanks to Philipp Hug and Patrick Winnertz, virtualbox is now available in Debian. I have just installed my first virtual machines, and I am now able to run Windows 2000 or Ubuntu from my Debian box. I was using Qemu until now, but the performances (despite the Kqemu accelerator) were quite bad, and I simply couldn’t run any Linux 2.6 based distributions with the kqemu active. Setting up virtualbox is really easy, just run aptitude install virtualbox virtualbox-source, compile the vboxdrv module (I have used module-assistant), add your user to the vboxusers group, and that’s it! I only had to download manually the guest additions iso image from http://www.virtualbox.org/download/1.4.0/, as the VirtualBox manager wasn’t able to find it automatically. You can launch the VirtualBox manager from the graphical menu and set up your first virtual machine.

Despite my previous choice to stick with an i386 system, I couldn’t resist any longer when I have seen that both openoffice.org and the proprietary Flash player were easily available in the amd64 port (not a fresh piece of news). I have thus just re-installed my main desktop machine which is now running a complete amd64 system. I have added this architecture to my unofficial packages repository, and all the arch-dependent packages in the unstable branch were re-built. Note that I do not plan to re-build packages available for the current stable distribution (namely etch). This might however change in the future if people show interest for
this. I have been able to build i386 packages in a chroot using pbuilder and cowdancer, which I will explain later in an article.

I have just received my new mobile phone: a Sony Ericsson K750i which I have chosen mainly for its price and multimedia features, but also because it has an internal modem, accessible via both bluetooth and irda. So far, I was already able to connect to the internet in GSM, will try GPRS tomorrow as I need to change my SIM card. libbtctl and gnome-bluetooth 0.9.0 were also released today. Updated packages have been uploaded to my unofficial repository (please read carefully the release announcement which explains these packages have been deprecated). I have also tried wammu for the first time, it will help me update my address book.

I decided to buy a Livebox on Saturday. The livebox is a modem-router used by my French ISP Orange. It allows me to profit from Orange VoIP service (and thus stop my standard phone line), as well as from the TV over ADSL (which I haven’t tried yet). My only issue is that I can’t disable to router mode of the Livebox (which is a Sagem F@st 3202 router) - I would have liked to keep my current configuration with a simple modem, in bridge mode. However, I have managed to re-configure shorewall on my internal router and redirect everything from the livebox to this router. This setup means double NAT, but it seems to work like a charm.

VirginMega.fr is a one of the French music download service which offers MP3 files (without DRM). They previously prevented all the browsers except MS Internet Explorer on MS Windows platform from accessing their website. I (we) used to change the browser id thanks to the FF User Agent Switcher plugin.
It now seems they have changed their mind under the pressure of Linux users. However, it seems there is still a problem for me (us) to access their website as they do not use the engine (Gecko in our case), but the user agent name… No, Iceweasel is not Firefox! Nor is Galeon! But why blocking access to a website which works perfectly with these browsers which share the same engine?

I received this week my new bridge camera: a Sony DSC-H9. It is meant to replace my Konica Minolta Z5, which had some defaults for the use I have. Along with its 8.1MP CCD and its 15x optical zoom, the H9 has some nice features, such as a big 3″ screen, a NightShot infrared system, a high sensibility mode (up to 3200 ISO with some expected noise) etc. Following to my first tests, I particularly like the menus, which are clear and allow very quick fine setup. The AF is very reactive, the built-in flash quite effective. The camera is delivered with an infrared remote control, as well as a lenscap. The only drawback for the moment is the usb cable which is not easy to connect to the camera. I guess I should use an external smart card reader in the future.

My unofficial Debian packages repository is now powered by lighttpd. This website is hosted on a small vserver I rent and I hope the switch will allow better performance in the future. The configuration was very easy, using fastcgi for PHP5 and mod_simplevhost. The next step is to allow encrypted sessions through OpenSSL which should be painless. I will stick with apache2 on my main (home hosted) server, but will use lighttpd by now on all the vservers I own. I only regret not to have done it earlier!

Thanks to RMS’s call, I have finally sent my subscription fees to become a member of APRIL (the French association for promotion and research in libre computing). I should have done this earlier but postponed my decision without any reason. I hope this will allow me to meet nice people with the same interests.