The pbuilder-based chroot script (empdebuild) in Emdebian SVN is now functional - that is it can run from a directory created by emsource, patch the Debian source using emdebian-patch files ( = emdebianising the package), get the build dependencies and install them, run emdebuild to cross-build the package and copy the updated patches, build log, .changes, .dsc and .deb files out to a nominal location before cleaning up the chroot. Only the basic pbuilder options and setup is supported and some values that can be configured for pbuilder are calculated automatically by empdebuild.

The embootstrap script that creates the chroot used by empdebuild needs some more work before this can be released as part of emdebian-tools 0.2.2 and there is no support (yet) for installing the cross-dependencies required for some packages. I'll be experimenting with Gyrosgeier's debian/xcontrol idea to specify these Build-Cross-Depends: using information already gathered on an ad-hoc basis. xcontrol files can then be committed to Emdebian SVN as patches. The pbuilder code is sufficiently flexible that the same code that parses debian/control for Build-Depends can be adapted to parse debian/xcontrol for Build-Cross-Depends.

In theory, empdebuild can become the basis of a cross-buildd, significantly improving the rate at which packages can be emdebianised.

The script does, currently, have to use apt pinning to retain the Emdebian toolchains because of problems with gcc-4.2 on arm.

(And yes, I do realise that emdebuild and empdebuild are very similar names but then so are debuild and pdebuild upon which the respective Emdebian scripts are based.)


http://buildd.emdebian.org/svn/browser/current/host/trunk/emdebian-tools/trunk/pbuilder

Dear layweb, I know that somebody from the kernel team (iirc) posted an article/mail about how to make a customized version of the Debian linux-image package, but I can't find it anymore.

For short, I want to produce a slightly modified linux-image-2.6.18-4-amd64 package. Which is the right source I want?

Is

apt-get source linux-image-2.6.18-4-amd64

what I want or is it

apt-get install linux-source-2.6.18
?

I am confused. Help would be appreciated.

I forgot to say, I payed for it yesterday. Now I have to buy another suitcase since the one I bought 2 years ago was ruined by the people which do the luggage handling on the Otopeni port, when I returned from Spain in September last year.

Those who program C are used to seeing function / subroutine definitions in header files. Not so in Fortran, there people use the C pre-processor to put Header files in subroutine definitions. Don't try this at home, folks. Again, names changed to protect the afflicted ...

subroutine  funmame( param1,
	      param2, param3,
#ifdef FOO32
	     foo, bar,
#endif
	      param7, param8, 
#include "basarg.h"
	     )

Remarkably, basarg.h included:

	C Comment.
	param5, param6,
#ifdef SOMETHING
        param10, param20,
#endif

There were in fact dozens of parameters in basarg.h.. And even a higher.h that included basarg.h but had more parameters ... Tags programming, bizarre, work, fortran

Often looking through old code you find yourself wishing for more comments, or even more correct comments. Rarely, less comments. But looking at some code I'm debugging at the moment at work I found the following bizarre idiom (details changed to protect the afflicted) ...

#ifdef DOC
C  Some fortran comments.
C  Yes, ordinary comments, nothing fancy like preprocessor directives, but if you want
C  your code without a paragraph explaining what it does, you can just preprocess it
C  and remove it.
C 
C Perhaps they'd expected that their fellow developers would write essays in the
C source code that would make it hard to follow, but then, you'd still need to edit
C the full source code to debug and develop it, right?
C
#endif

Tags programming, fortran, bizarre, work

I'm currently quite busy and I don't have time to manage a vote for the DeFuBu contest. I am already 2 months late for this. I am hereby requesting some help to organise the vote. I have a list of people/bug number and you should only put this formatted and find people to vote for, and then calculate the ranking. I'll do the rest. So please contact me if you have some spare time.

A number of recent events reminded me of how much time has passed since I started using Linux. Among these, the retirement of several 32-bit CPU architectures from the list of supported platforms on Debian and Ubuntu. The first architecture to fall was m68k. Release Managers at Debian proposed to drop support for m68k after Sarge, which indeed happened: Etch was released earlier this month and, for the very first time in Debian history, without m68k support. Granted, a number of dedicated developers are working on porting the GNU toolchain to Coldfire (an embedded platform that supports about 80% of the m68k instruction set) and good chances are that they will succeed but, let's not fool ourselves: m68k is gone. Around the same time, Ubuntu dropped support for PowerPC, just before releasing Feisty. Granted, there is still some PPC64 hardware being manufactured, but this is on the server side and thus getting farther away from the desktop market that constitutes Ubuntu's core business. Of course, given the plethora of 32-bit PPC hardware in circulation, in the form of second-hand Power Macs, the platform still has some lifespan left but, again, let's not fool ourselves: PowerPC is gone too. And now, a message sent to Debian mailing lists suggested that, as of Linux kernel 2.6.21, support for 32-bit SPARC was broken and, due to an insufficient developer interest for maintaining the GNU toolchain, 32-bit SPARC was likely to be dropped for the Lenny release. Yup, we really cannot fool ourselves: support for 32-bit architectures really is gone. Not surprisingly, my own computer collection has followed a similar trend: this morning, my beloved PA-RISC host left for the home of a happy computer hobbyist who had won the online auction on it. My Atari TT030 is likely to follow a similar route fairly soon. Ditto for the exotic Atari Stacy sitting besides it ...and let's not mention the huge stack of Pentium II and K6 desktop hardware sitting in the corner. Looking back, I'm glad to have shared so many brilliant years with 32-bit hardware. The variety of CPU architectures and hardware designs was quite a learning experience. This being said, life goes on and my current needs are fully covered by my Linutop and by my aging ThinkPad. Thank you for all those years, 32-bit computers. Enjoy a well-deserved rest in bit heaven.

I'm rather surprised to find that Liz RADFORD comments:

"As my first year as organiser of this ride, I would like to thank you and everyone who came out for us. It was a great day and looks like we raised over 10,000! As Leukemia is the most common form of cancer in Childen I get a bit over-enthusiastic sometimes when I see so many people joining in doing so much good and prehaps see the event through rose tinted specs. I hope you enjoyed it and the legs are better now, the hill is a real challenge. One question for you. Was it something you enjoyed enough to do again or do you think a flat route along the front Sand Bay and Wick would serve us better at bringing riders back each year? Your comments appreciated. Liz X Also check out www.UKeventphotos.com"

10k. That's a great result! I think I'd do it again and I think the initial hill has its uses - it's a bit of an unusual challenge and strings the field out. I didn't have any congestion problems until the first checkpoint after we merged with the short route. If you're looking for changes, I'd put that Somerset Avenue checkpoint somewhere with more space and reroute a quieter way around Worle, maybe up Walford Avenue with the split to the Ebdon loop at Riverbank. UKeventphotos.com only shows me the first page of the gallery - clicking "Next results" dumps me back at the search front page. I think they need to learn how to make accessible web sites (my work offers advice...). I'm glad I took my own camera and I'm sorry if their bad web site limits the amount LRF gets from photographs.

I’ve been struggling for a tagline for Webconverger. The previous tagline “Webconverger is a free, up-to-date Web browser kiosk live system”, was a bit of a mouth full. Now, I’ve decided that a “Web operating system” best describes Webconverger. Webconverger isn’t just a “kiosk”... The reason why I didn’t call Webconverger a Web operating system initially, was because I didn’t want to confuse anyone. You see I’ve read blogs which talk about Web desktops or rather Web applications that emulate desktops, “operating systems”. I think for such Web applications, I’ll call them Web desktops, to be called an operating system is just abuse. Calling Web desktops “operating systems” tarnishes the good name of operating systems, when silly Web applications emulate broken dead WIMP win32 environments. I can’t be the only one that hates Windows. I’m not talking about Microsoft Windows. I am talking about overlapping Windows windows. :) So I have a little fight on my hands with Wikipedia and the entry on Web operating systems. Hopefully someone agrees with me that Webconverger is a Web OS. Sidenote: I’m in the process of releasing the Debian-live etch based Webconverger 2.15 (download) which features a new textless progress bar and reset feature.

I was reading some of the debate regarding this guy and his business: The man who owns the Internet and got to thinking about the comparison between domain names and real estate. One of the things that irritates me about domain name squatters is that they can essentially sit on a good name forever, at very, very little cost. You can't do that in real estate, because you have to pay property taxes, which acts as an incentive to actually do something productive with the property, rather than simply let it sit there until someone comes and offers to buy it from you for even more than you paid. So you can't let the space go completely to waste - you have a big incentive to at the very least rent it out to someone who can make money with it. You can't really rent out a domain name, though. Since it's an integral part of your brand and identity, it would be insane to borrow one for a few years, whereas with a physical business renting space, people will still be able to find you if you move. Perhaps an annual tax based on the purchase price of domain names would encourage people to either do something valuable with them or pass them on to someone who could? I don't think anyone would be sad to see pure speculators gone, but on the other hand, I have a strong suspicion that all the extra taxes and red tape would make the idea a net negative, and taxes on 'real' domains like google or ebay, would simply be passed on to consumers, and so wouldn't be desireable. In any case, it's something interesting to think about.

I just finished reading Ben Yoskowitz's Top 10 reasons to join a startup, and I feel like I have to say something. As an experienced programmer, I think of a startup as an abusive relationship. Most technology startups I've been in have been run by charismatic sociopaths with no actual management or technology experience. They make wild, ignorant, unfounded promises to employees, investors and potential customers, then get increasingly anxious and unpleasant as their unsupportable predictions of schedule and market share don't come through. Employees end up working 80-hour weeks, typically for low or no pay and worthless stock options that never pay out. Guilty that living the "startup lifestyle" (work from home, come in late, have fun) has caused schedule or feature slips, they push themselves too hard at work, to the detriment of their health and home life. It doesn't help you much get another job later. Most startups don't succeed, which means that you end up with 1, 2 or 3 years of space on your r sum consumed by a made-up company name that no one has ever heard of and whose phone line has been disconnected. If people have heard of the name, they'll associate it with failure ("Oh, yeah, wasn't that the company that spent all that money and went nowhere?"). On top of all that, the tech experience of working in a startup is usually pretty bad. The engineers are typically unexperienced -- just out of school, or close to it -- and they waste a lot of time and effort trying to reinvent the wheel. The idea that you're changing the world encourages this kind of bad engineering. You end up cutting a lot of corners and developing a lot of bad habits that you have to unlearn later on. Schedule pressures make you drop important engineering steps like analysis, design, unit tests, documentation. That is, if you even knew you were supposed to do them in the first place. It takes a lot of experience in this kind of pressure-cooker environment to come out of a startup sane, healthy, and wealthy. The vast majority of people end up with broken relationships, poor mental and physical health, and not much in the way of money to show for their efforts. tags: startup benyoskovitz work

Wikiclock in the news So, Niko, Hugh, Mike and Leoben Richardson all blogged about the Wikiclock. As far as I know, no MSM coverage yet, but it should be coming down the pipe any day. I talked to Niko tonight, and he said that he saw Hugh at La ka, and Hugh was updating the wikiclock. They traded off updating all afternoon. Sigh. It's actually quite addictive. tags: wikiclock niko hughmacguire leonardrichardson mikelinksvayer

Transit strike We've got a transit strike on in Montreal (see the CBC for deetz). It's a pretty heinous problem -- the maintenance workers are asking for X% pay increases over Y years with blah blah blah pension and fie fo fum and... Whatever. The city is refusing to budge, which makes our mayor, G rald Tremblay, officially a hypocritical idiot. Tremblay has been wagging fingers at provincial, federal and international figures for years about sticking to the Kyoto protocol, but when the hard decisions had to be made, he was willing to shut down Montr al's transit system... in summer... after one of the warmest winters in history. You can blame both sides on this issue, but when it comes down to it the union's job is to get the best deal it can for its workers, and Tremblay's job is to keep the Montreal M tro running. Someone's not doing his job. The city spends millions trying to get commuters out of their cars and into the public transit system. And all that work's out the window, now. Despite the public call for more people to rollerblade to work (riiiiiiiight), traffic is way up and auto use is way up. That means air quality is going to go down, and our city can go from being an environmental leader to an environmental mistake. I'm glad to hear that the Quebec government is stepping in to break up the fight, but I hope it doesn't take too long. tags: montreal quebec metro strike

Calling graphic designers: Universal edit button One of the interesting movements to come out of RoCoCoCamp this weekend is the idea of a universal wiki edit button. This would be an icon similar to the RSS radio-wave icon that's become ubiquitous for indicating an RSS feed. The UWEB would symbolize that the current page can be edited. Folks at AboutUs.org have taken the lead on this effort, which could be really interesting. Getting a universally-understood sign for "edit this page" would greatly help with acceptance of the wiki way by the general public. I don't think it'd have to be limited to wikis; any page that could be edited (e.g., on a CMS) could also have the UWEB. I highly encourage skilled graphic designers who want to take a crack at designing the Next Big Icon to look at the UWEB effort and give it a try. tags: graphicdesign universalwikieditbutton wiki rocococamp aboutus.org icon

At my Institute, we get the El Financiero newspaper - I like it quite a bit. It's very well balanced, with opinions from all of the political and economical points of view. Of course, it has huge business and market sections which I don't usually even care on looking at, as I won't understand much anyway. Those sections usually include advertisements for potential courses and businesses.
Today, Nadezhda showed me an ad that's just beautiful - It shows off the level of commitment and seriousness a company has. It is so beautiful that it deserves me copying it in Spanish and translating it for your further enjoyment. Of course, if somebody stumbles upon this advertisement at my personal blog - I don't know them, they might just be serious but clueless :)

GRUPO DE ASESORES PROFESIONALES
CONVOCA
A acad micos, investigadores, intelectuales, economistas y expertos en reformas estructurales para el an lisis de las siguientes:

• Econom a y finanzas
• Educaci n y empleo
• Seguridad y desarrollo social
• Pol tica y Energ a
• Reforma del estado

Para integrarse a un equipo serio y propositivo de Profesionales en estas reas que brinde asesor a puntual sobre las antes descritas.
REQUISITOS: Titulados con Maestr a y/o Doctorado. Contar con publicaciones especializadas.
Interesados enviar CV al correo electr nico:
juan_zzz_at_starmedia_dot_com
juan_3zzz_at_yahoo_dot_com_dot_mx
Bosque de Ciruelos No. 140, piso 12, oficina 1202, col. Bosques de las Lomas

And now, for your further amusement, in my hastily translated English, as faithfully as I can do it.

GROUP OF PROFESSIONAL ADVISORS
INVITES
Academics, researchers, intelectuals, economists and structural reform experts for the analysis of the following areas:

• Economics and finances
• Education and employment
• Social security and development
• Politics and Energy
• State reform

To join a serious and propositive team of Professionals in the above areas that gives punctual advice regarding the aforementioned.
REQUIREMENTS: Holders of a title, with M.Sc. or Ph.D. studies. Having specialized publications.
Interested people, send your CV by email to:
juan_zzz_at_starmedia_dot_com
juan_3zzz_at_yahoo_dot_com_dot_mx
Bosque de Ciruelos No. 140, piso 12, oficina 1202, col. Bosques de las Lomas

Yes, several of the redaction mistakes are in the original text (and several more were introduced due to my English translation, of course).
Anyway... Would you believe in the seriousness of a professional-looking group of economists, paying around US$2000 for 1/4 page in one of the leading Mexican newspapers? By the way, their offices are located at one of Mexico City's most exclusive, expensive areas. But... For $ deity 's sake... COME ON! Please tell juan_zzz to get a decent-looking mail and domain! Having their main contact addresses at two free mail providers, Starmedia and Yahoo, does no service at all to their professional image! How can people still not pay even a bit of attention to those basic details?

A while back I set up a d-i test lab, a kind of automated mechanical computer torture rack, which has been trying (and frequently failing) to automatically install Debian onto a half-dozen varied machines every day for a couple of years. Problem is that hardware sucks, even when it's tucked away in a friend's basement and so not blaring at you with its combined total of 100 fans and 50 disk drives. It has an annoying habit of dying, especially if it's oddball machines for different architectures. Cords have a habit of getting unplugged. It's hard to automate booting from installation media. And so on. A while ago I decided to go virtual. Now, thanks to Dann and Matt and HP, I've gone from this: to this: HP donated hydra, a maxed out 8-way Xeon with 4 (soon 8) disks. The idea is that it can run 7 or so virtual machines, all doing test installs, in parallel. Fast. I'm using kvm for the amd64 and i386 tests. Since Debian has mips and mipsel kernels specially built for running under qemu, I'm able to test the installer for both those arches. Hercules will cover s390, once the kernel gets fixed. This leaves some architectures still todo:

• arm should be supported by qemu, but Debian doesn't yet have a kernel for it. Please, please kernel team, add an arm qemu flavour!
• sparc's qemu support seems very rough. I haven't tried to get it working yet.
• powerpc probably has an emulator I can use. Any suggestions? I need something that can run headless. Qemu would work, if we had an appropriate special kernel for it.
• m68k I have not researched emulators for yet.
• hppa, ia64, and alpha all lack a usable emulator, AFAIK. Luckily I have machines for all of these, so I can pare the physical hardware down to just those 3 boxes.

I also still have to write enough installation test cases to keep this monster machine firing on all 8 cores all day long. It can probably handle around 300-400 installation tests per day. Since hydra has the disk space, and these emulators are generating lots of disk images with Debian pre-installed, I have also set it up to retain the disk images. I imagine that some of these disk images will eventually be useful for devleopment, since it's much easier to boot a disk image rather than sit through a slow qemu install on an arch like mips. I haven't worked out how to make these images publically available yet, but hydra is located on the name network as gluck.debian.org, if anyone wants to get access to them. Update: I also have qemu doing a 'savevm 1' before shutting down, so you can even restore the running system from the disk images. Especially useful if the install failed somewhere in d-i! :-) BTW, I can give out accounts on hydra to any Debian developers who need a fast machine to run kvm/qemu on, or to anyone who is interested in setting up their own installation tests.

The first lesson was about wrong assumptions that people do when it comes to i18n. This article is mostly targeted at developers which want to support i18n properly, in particular, plural forms.

This is my second n?getetxt lesson.
This is about making your update-po target to include all the strings you want localized in the pot file.

Probably, at some point you, as a developer have written something like:

print("You won %d point%s", points, points>1?"s":"" ) ;


Well, is nice and dandy that you want to make the proper choice of words when you have more than one point. I appreciate that, but you are doing it wrong. There is a tool for that, as I explained before.

Now, lesson 2.

You have understood what you have to do, but you just hate typing ngettext/gettext all over the place, so you probably defined something like:

#define _(X) gettext(X)
#define _n(sg,pl,n) getext(sg,pl,num)

So now you can write trhis piece of code:

print(_n("You won %d point", "You won %d points", points), points ) ;

The problem: how to convince xgettext to add those strings to the pot file?

Nothing simpler... just add _n as an additional keyword when invoking xgettext... but you have to specify that this is a definition for ngetetxt, so, you will end up having in your update-po target something like this:

xgettext ../src/*.c --keyword=_ --keyword=_n:1,2


TADA!

Thanks to Planeta Debian (yes, planeta, the Spanish version of Planet Debian), I came across this Dar o Rapisardi's post.
Sometimes poetry can be expressed in human-interface guidelines.

Has anyone seen a recent poll for Galway West? (In the Irish General Election, that is). The last poll was March 26th, which is way out-of-date : long prior to the election being called, before the Cryptosporidium Crisis, etc. The result is, I'm in the unusual position of having to make a decision. Ordinarily I'd vote Number 1 for Niall O'Brolch in (Green) and 2. Micheal D. Higgins (Labour). Ireland uses the unusual Single Transferable Vote system, with multi-member constituencies. Galway West is a 5-seat constituency, and Micheal D. is a long-sitting TD, who most would have expected to be automatically returned. Niall O'Brolch in is fairly new; he's a member of the City Council and current Mayor of Galway (thats a relatively powerless post in the Irish local council scheme of things, though). Unfortunately news from the Labour camp appears that too many people assume he will be automatically returned; the word they are getting on the street is "Yes, you'll be getting my no.2 ". But in the STV system, this could mean he wouldn't be in the running. Those votes for the Greens and others will perhaps stay with the Greens and won't transfer to him. So the choice, Vote 1. Greens, which is my natural preference, or 1. Labour, in the hope of getting both Micheal D. and Niall elected? a poll of the current situation would help. BTW, for voting geeks, following the link and reading about STV is worth it. It makes Irish election counts the most fun election counts in the World. One of the main arguments against electronic voting in Ireland has been that it would take all the fun out of the count, which would be over too soon ... Tags politics, galway

I should have known better than to even hint that life was getting back to normal. We're back on weekends and overtime at work. Personal life is also a bit hectic. So, Debian stuff is postponed in favour of sleep in most cases. I'll clear what bugs I can, but new code is at the bottom of the list. On the good side, my daughter Caoimhe, started Walking! Its good to get home and see the kids. Spring is coming. I'd like to get out and see some of it. Tags blogging, family, caoimhe

Well, theres a General Election here in Ireland on the 24 May, and the politicans are well and truly on the last lap of the campaign trail. We're a bit rural to see them all, for those who don't make it to the door, here's something to think about. In Moycullen we are on the edge of the Gaeltacht. Various institutions, such as the main primary school, have been struggling to retain that status with the encroachment of the B arla 'burbs (suburbs of the Gaeltacht where English dominates). Our children go to Baile Nua, the villages' gaelscoil. To hold standards, the school has a policy of capping its current enrollment (to 84, I think), to keep class sizes small, until the potential intake is large enough to justify another teacher. Fair enough. However this has the side-effect that, in an area supposedly set aside to protect the Irish language, students are necessarily turned away from the only Irish-language school in the Village. Any comments? Tags irish, politics, schools, galway

Pete, an old high school friend and biology-project partner, recently discovered my email address and we reconnected. It occurred to me that the next generation may never experience reunion with long-lost friends: they’re all on Facebook and Myspace from middle school on up, so how will they ever lose touch in the first place? Peter wrote to ask me the name of an Amiga game we used to play in the late 1980’s. I knew immediately which game he was asking about: Mind Walker. I remembered it as a fantastic surreal action/video game exploration of the human psyche. I think it was probably more Freudian than Jungian. Only a few games stuck with me at this level — another one was Weird Dreams. (The problem with Weird Dreams was that it was impossibly difficult to get past the level where you have to smack Dali-esque statues with flying fish. If anyone ever did, I’d like to know what happened next.) This review gives a good summary of Mind Walker:

You are a physics professor gone mad. Your course of action? Delve into your Mind, to inspire “Ideas” by tracing “Paths of Coherent Thought”, with the help of your split ego. Then through opened-up Tubes, enter your Brain, to retrieve “Shards of Sanity”. Finally, put them back together in Subconscious.

I remember Mind Walker as having amazingly spooky and captivating graphics. Then I found this screenshot: Oh well. I’m sure it actually was impressive at the time. This discovery reminded of times when I’ve rediscovered a favorite food from childhood — for example, a certain type of cookie — only to find that it really isn’t very good at all. Just kind of sugary and low-quality chocolate. It’s also like going back to watch the original Jurassic Park again. The amazement is gone. Actually, even though the Mind Walker graphics aren’t as impressive as I remember them, I’m sure it was still a great game. Anyone have screenshots from Weird Dreams?

Share This

playing german lotto update: sorry for the broken feed, I hope I ve kicked this piece of php hard enough so it works as it s supposed to work. Tomorrow I ll replace it by something maintainable, written in python. Stefano, I have no clue how the lottery in Italy works, but I ll play the numbers generated by this little program on Wednesday and Saturday:

#!/usr/bin/python
 
numbers = ["09", "F9", "11", "02", "9D", "74", "E3", "5B",
           "D8", "41", "56", "C5", "63", "56", "88", "C0"]
 
#in germany all lotto numbers x must be 1 <= x <= 49
numbers_int = [ (int(num, 16) % 49) + 1 for num in numbers ]
 
#unfortunately 16 numbers are not enough for 3 fields in
#the german lottery, so let's add the current year
#(of course, you can add 2 other numbers, if you don't like this year)
numbers_int += [20, 07]
 
print "your numbers for the german lotto:"
print numbers_int[0:6], numbers_int[6:12], numbers_int[12:18]