Older blog entries for joey (starting at number 533)

experimental multiuser git-annex repository

On the last day of DebConf 13, RichiH and Ganneff and I set up a git-annex repository that all Debian developers have automatic commit access to, and everyone can clone and browse. See the announcement email.

This is an interesting repository for me for several reasons. Selfishly, I have always wanted an easy way to download lots of DebConf photos for offline viewing, and never seem to get around to downloading everything from gallery.debconf.org when I have bandwidth. I've also wanted an example repository that shows how git-annex can be used by a large group for collaboration. Finally, the way this repository is set up with an incoming queue is fairly unique.

With 430 files in the repository, totaling over 3.5 gigabytes (which doesn't include all the talk videos that are #included into it), and at least 18 people having cloned the repository so far, the debconf-share repository is well on its way to being a sort of large git-annex repository.

Just running git annex whereis is interesting; many of the files already have 8 copies. Some talk videos are more popular than others and you can see when they're downloaded too. But enough snooping.. ;)

So far people have uploaded mostly photos and talk slides. Other places exist to store those things in the DebConf infrastructure, but it's nice to have them all available in one tree. I particularly like today's addition of chrysn's files which include the raw photos and hugin files used to produce panoramas, and then pull those together into a postcard which has all its sources available.

In my corner of the debconf-share repository, I'm collecting together files regarding the possibly-historic dpkg-source-git-re-re-redesign process that would have otherwise been scattered around various places and probably not all published. This includes an hour long recording of the main design session (recorded with permissions) made by my laptop's mic, which, surprisingly, turned out to be pretty listenable. I will probably have more to say about this process later, once Ian announces dgit.

So, we're still seeing how usage develops. I hope that having this available during the next DebConf, and other Debian meetings, rather than only at the end, will further facilitate file sharing and storage. Especially if a fast clone is available right on the DebConf LAN. ;)

The technical details of how the repository is put together are:

  • There's a repository on git.debian.org, which piggy-backs on the collab-maint group, so most Debian people have commit access to it.
  • git-annex is used to upload files to that repository, as an incoming queue.
  • A git-annex-shell annex-content hook is run whenever someone uploads a file to there. It moves all annexed content over to annex.debconf.org for publication. This involves some ugly but safe stuff to do with publically readable restricted use ssh private keys. It was the hardest piece to get working, and is only necessary because we don't want to bloat git.debian.org with this stuff and it's not practical to give everyone logins to annex.debconf.org.
  • As an additional guard against accidental bloat, the git.debian.org repository will refuse to accept uploads when there is less than 5 gb free disk.
  • annex.debconf.org was set up following the git-annex public repository on a web site howto.

I've found at least 4 bugs in git-annex as a result of this repository, which is a rather unusual use case. And fixed 2 of them so far..

Syndicated 2013-08-21 20:20:18 from see shy jo

Debian at 20

RichiH used flint and steel to light the bonfire. We carefully fed it up from those sparks to a blaze. Put on the biggest logs we could find to make it last. Now I'm sitting on the hill above it watching folks gathered around. A poem is being read, in Hindi, then translated for us to English. Others have shared songs and poems in a dozen languages, both classics and their own. Out below the darkness of the lake. This is Debian at 20.

At the start of this DebConf, I gave a talk on "Debian Cosmology. In that (and a later "dh_busfactor" talk) I shared my hopes and my fears. I was conflicted about giving the this talk, worked on it for weeks, felt it might not work, or be depressing. I've had nothing but good comments about it.

20 years is ages in internet time and technical projects ossify over time. The last session I was in this afternoon was a presentation of a new tool, which I hope & feel has the potential to fundamentally change an important and suboptimal part of Debian. Then I walked outside to a rainbow over crystal clear Swiss alps on the horizon. How encouraging, and what a nice story that will be around some future campfire.

Syndicated 2013-08-16 22:16:42 from see shy jo

please build a haskell to perl compiler

The Fay compiler is a simple way to build fairly comprehensible javascript code from Haskell source.

It occurs to me that it should be rather easy to modify Fay to emit perl code rather than javascript. This would allow contributing things like plugins to various perl programs, without writing perl.

Of course, the same idea could probably be used to compile Haskell to other languages like python, but perl seems particularly well suited as a second Fay target, since javascript and it have quite similar syntax and similar support for features like closures which Fay relies on.

I do not have time to work on this idea myself. It would be a good project for a beginning Haskell programmer. You probably don't even need to fully understand monads to do it! Essentially, look at Fay output examples, translate them from javascript to perl, and then much of the code changes in Fay would probably be in simple string generation code.

I will forward any bitcoins sent to the address 149eBtWS6i8cwQdPQJJ8hAGpDuEqNidyTj to whoever makes this. If it doesn't happen in 1 year, any donations will be forwarded to the EFF instead.

Syndicated 2013-08-12 12:40:54 from see shy jo

good morning

panorama of sunrise

Woken at 3 am by fil singing "Join Us Now And Share The Software" (all verses!) I could not get back to sleep and spent 3 hours thinking up a new take on the hopelessly blocked dpkg-source v3 (git) format.

A diagram of the new plan, which should meet all ftpmaster requirements, is posted in Hacklab 1. I am looking for reviewers.

A rather hard to read photo (DebConf needs mandatory whiteboards!) is available here:

Fil has paid me back in full for his drunken carousing by gifting me a Rhombus-Tech system on a chip on a PCMCIA card. I've checked this new computer, which features a modern multicore ARM CPU, into my wallet.

And that's 10% of what went on today at DebConf for me, and we've not even gotten to the cheese and wine party tonight.

Syndicated 2013-08-12 12:21:19 from see shy jo

DIY crowdfunding and bitcoin

Well, my git-annex crowdfunding campaign is half way to its August 15th conclusion. So far it's raised more than five times what I hoped it would. I wish I could say I'm like some canny NASA engineer who intentionally sets low expectations for their Mars rover, but in both the previous kickstarter and this campaign I've really had no idea how far it'd go. I'm glad that I'll be working on git-annex for another year.

I was particularly unsure if it'd be successful to move off Kickstarter. During the git-annex assistant Kickstarter campaign, I saw many small contributions from people who learned of it due to it being a successfully funded project, a staff pick, etc. Losing that easy network effect is a gamble.

So far I've had only half the number of contributors that I got on Kickstarter. I've basically missed out entirely on the $5 level casual contributors. On the other hand, my backers have generally been more generous (and some have been exceedingly generous). And I've avoided rewards that will cost much money, so I may end up in the same ballpark funding level in the end!

Incidentially, I'm really enjoying getting in touch to let people know when I make their sponsored commits. There's still time to sponsor one of your own ;)


I also was curious to experiment with Bitcoin in this campaign. Partly because Paypal isn't available everywhere internationally, and takes really obnoxious percentages of transactions (though probably not as bad as Kickstarter taking its percentage followed by Amazon payments taking its percentage..) and partly because there seem to be interesting possibilities for supporting free software with Bitcoin. (Especially if any of the microtransactions on top of Bitcoin take off.)

So far 5% of backers have used Bitcoin. It's been quite strange to actually have significant amounts of bitcoins in my wallet. Wordpress has had 94 bitcoin payments over 9 months since starting accepting them. I've had 47 payments in the two weeks my campaign has run so far. Wow!

Most of the bitcoin payments have come in via Coinbase (a few people have found my direct payment address), but of those very few were using bitcoin purchased on Coinbase. Most are probably transfers of bitcoin they already had, or perhaps bitcoin purchased on other sites.

The one technical issue I've had with using bitcoin is that Coinbase has not provided details about who sent most of the donations. Probably some of them are intentionally anonymous, but I suspect Coinbase's interface to claim incoming bitcoin transactions failed for some of them. (If you donated bitcoin and want to actually get a reward, please email me.)

By the way, I'm converting most of the bitcoins back to USD pretty quickly. I'm not interested in speculating on currency exchange rates with money that has been donated so I can accomplish a particular task..


I put up the campaign website without any means in place to handle updating it. This is because I never automate anything until I've done it at least 10 times by hand. ;) After the first trickle of donations became a flood, I quickly realized I needed at least something to handle keeping the numbers straight.

What I whipped up in an hour of coding is a system where I enter incoming payments into a hledger file and a small haskell program parses that and writes out various files that are included into the website. Amusingly the percentage calculation and display code was copied from git-annex, so part of git-annex is helping run its own fundraising campaign. The campaign video is itself hosted in a public git-annex repository, come to think of it.

The rest of the site is built using ikiwiki. Given that it's hosted at Branchable, this is a high level of dogfooding and DIY. There are certianly better crowdfunding platforms, but all I miss in this one is automated transaction entry. And I have total flexability, double entry accounting, and a powerful static website generator that handled being on the top of Hacker News without a sweat. Oh, and some money. What's not to like?

Syndicated 2013-08-02 06:22:09 from see shy jo

git-annex as a podcatcher

As a Sunday diversion, I wrote 150 lines of code and turned git-annex into a podcatcher!

I've been using hpodder, a podcatcher written in Haskell. But John Goerzen hasn't had time to maintain it, and it fell out of Debian a while ago. John suggested I maintain it, but I have not found the time, and it'd be another mass of code for me to learn and worry about.

Also, hpodder has some misfeatures common to the "podcatcher" genre:

  • It has some kind of database of feeds and what files have been downloaded from them. And this requires an interface around adding feeds, removing feeds, changing urls, etc.
  • Due to it using a database, there's no particularly good way to run it on the same feeds on multiple computers and sync the results in some way.
  • It doesn't use git annex addurl to register the url where a file came from, so when I check files in with git-annex after the fact they're missing that useful metadata and I can't just git annex get them to re-download them from the podcast.

So, here's a rethink of the podcatcher genre:

  cd annex; git annex importfeed http://url/to/podcast http://another/podcast

There is no database of feeds at all. Although of course you can check a list of them right into the same git repository, next to the files it adds. git-annex already keeps track of urls associated with content, so it reuses that to know which urls it's already downloaded. So when you're done with a podcast file and delete it, it won't download it again.

This is a podcatcher that doesn't need to actually download podcast files! With --fast, it only records the existence of files in git, so git annex get will download them from the web (or perhaps from a nearer location that git-annex knows about).

Took just 3 hours to write, and that's including full control over the filenames it uses (--template='${feedtitle)/${itemtitle}${extension}'), and automatic resuming of interrupted downloads. Most of what I needed was already available in git-annex's utility libraries or Hackage.

Technically, the only part of this that was hard at all was efficiently querying the git repository for a list of all known urls. I found a pretty fast way to do it, but might add a local cache file later on.

Syndicated 2013-07-28 21:03:02 from see shy jo

524 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!