Search Results: "mattb"

17 February 2006

Simon Richter: On 2.6 and overcoming suckiness by l33t technology

Matt, since 2.6 is now managed in git, you can use git's bisect feature to find out the commit that caused the problem. Basically it boils down to git bisect bad HEAD
git bisect good <some version where it worked> then compile the tree, test it, and use "git bisect good" or "git bisect bad" if it works or it doesn't, respectively. Repeat until it tells you which commit is at fault.

25 January 2006

Matt Brown: LCA 06 Update

Finally on the third day of the LCA (or the first day of the conference proper) depending on how you look at it, I’ve had a chance to sit down and catch up on a few things . I managed to take on the task of running the network for the conference, and as we all know, the geeks need their network access. We’ve had some issues getting the network up and running which has been interesting. Most of the problems were caused by issues with incorrectly configured rate shaping on our upstream link which was limiting our outgoing capacity to about half of what it should have been. Once that got sorted out we discovered that the fibre tranceiver was not playing nicely with the NIC in the linux router we had been lent use of for the conference. This led to 20% of packets being dropped, and severely hampered usage of the network! When we swapped the dodgy NIC out for a new Soekris box it was immediately obvious that the connection “felt” better and the usage immediately spiked up over 3Mbps. By far the most prominent form of access at the conference is via wireless, we have 25 APs around the conference, built using Soekris net4526 biscuit computers and Atheros 802.11a/b/g wireless cards. We’re managed the network using the CRCnet Configuration System, which I also plan to introduce to the world during my talk tommorrow. We were a bit worried that we had overprovisioned the wireless, however the usage stats that we’re seeing seem to suggest that it’s about right. Earlier this morning we peaked with just over 210 concurrent clients on the APs and almost 5Mbps (our limit) of traffic coming in / out of the network. We’ve implemented some interesting protection strategies on the APs to try and prevent ARP spoofing attacks, although writing software like this in the week right before the conference and then trying to debug it as you build the network turns out to be not such a great idea! Currently sitting in listening to a talk by Brendan O’Dea about the high volume l2tpns software that they’re written at Optus. Very interesting.

22 January 2006

Matt Brown: At LCA06 - Busy Busy Busy

Haven’t posted for a while. At LCA06 in Dunedin since Thursday, network is finally up and running. Still lots to do!

5 January 2006

Matt Brown: New GPG Subkey / Offline for a few days

Happy new year :) One of my Christmas presents was a brand spanking new USB key. For quite a while I’d been planning to put my SSH / GPG keys off onto something like this for some extra security. I’ll write up the exact steps of how I achieved my setup in the sometime next week, but it was all fairly easy. As part of the transition I followed Adrian von Bidder’s guide to multiple subkeys such that I have now a signing key (0×4054AB08) which I keep separately from the private key material of my primary key. My new key fingerprint is below

pub 1024D/59B2D9A0 2005-02-03
Key fingerprint = 6583 A11F 1428 FCDD 65FC C7B5 E0CD 3CDC 59B2 D9A0
uid Matthew G L Brown (Default Key)
uid Matt Brown (Debian Packages)
uid Matt Brown (WLUG Key)
uid Matt Brown (CRCnet Key)
uid [jpeg image of size 4390]
uid Matt Brown (MediaLab Key)
sub 1024g/DF8A0504 2005-02-03
sub 1024D/4054AB08 2006-01-04 [expires: 2008-01-04]

I’ve uploaded the new key to the keyservers, or you can grab it from my key information page. All in all I’m quite liking the new setup, plug usb key in, password dialog pops up for SSH key, enter password, password dialog pops up for GPG signing key, enter password. I then have both keys stored in the relevant agents easily accessible for signing email / logging into machines. Meanwhile my GPG private key is sitting securely somewhere else. I plan to sign all outgoing email from now on. Next step is to investigate libpam-ssh for local authentication so that I don’t have to type 3 passwords whenever I login. Moving House
We’re moving house tommorrow, this means that my DSL connection will be down for a few days while Telecom and my ISP get everything sorted out. Hopefully it will go smoothly and I’ll be back online sometime mid next week.

27 December 2005

Matt Brown: Miredo/Toredo Patent Status

Miredo is an implementation of the Teredo - IPv6 over IPv4 tunnelling protocol which I have been playing with lately. I have filed an ITP bug to get Miredo into Debian, however this has been delayed while I try and get a handle on some of the patent issues surrounding the Teredo protocol. While Miredo itself is licensed under the GPLv2 the state of the Toredo protocol is less clear. Teredo was created by Christian Huitema, an architect with Microsoft. It is currently a Proposed Internet Standard waiting for the IETF editor to review it and assign an RFC number for it, IANA will then assign a permanent prefix for use by Teredo clients, servers and relays. According to the IETF IPR disclosure page, Microsoft has filed a statement of claim against the technologies described in the Proposed Standard. Microsoft states that they will provide a Royalty-Free, Reasonable and Non-Discriminatory License to All Implementers in their statement. Specifically the license that they offer the Teredo protocol under is called the Microsoft Royalty Free Protocol License Agreement. This is a catch-all license that Microsoft uses for all client-server protocols used by Windows. The license itself doesn’t make any claims about Microsoft’s IP rights, it simply gives you a license to use any rights Microsoft may have in the specific protocol you are using. For example at the end of the license agreement you select which protocols you wish to have covered under the agreement. Alongside Teredo you can also choose protocols such as IPv4, DNS, 6to4 and lots of other common protocols. The actual patent application for Teredo is incredibly difficult to find, the IETF IPR disclosoure doesn’t link to it, none of the proposed standards link to it and googling for “Teredo Patent Application” isn’t particularly successful either. In the end it turns out that the title is quirkily named “Allowing IPv4 clients to communicate over an IPv6 network when behind a network address translator with reduced server workload” and was published as Patent Application 20040190549 on the 30th September 2004. Incidentally if anyone knows of a nicer way to link to a US Patent Application than the ugly URL above, I would love to know. Half the reason that the Patent Application was hard to find is that it doesn’t actually mention Microsoft anywhere at all, unlike the ~5000 odd other Patent Applications currently filed by Microsoft. From what I can tell it appears that Christian Huitema has filed the application through is own personal attorneys independently of Microsoft. So where does this leave us in terms of getting Miredo into Debian? The short answer is I’m not entirely sure. I can’t seem to find a definitive statement regarding Debian’s position on patents anywhere, the best I’ve been able to come up with is a debian-legal post from Andrew Suffield from Jan 2005 which offhandedly describes the informal Debian patent policy as:
“ignore patents until somebody starts to wave them around, then drop the offending thing like a hot rock”
So maybe I’ve already shot Miredo in the foot, by investigating its Patent Status… :( My take on the situation seems to be that Debian is perfectly happy (and perhaps has no other option but) to ignore patents applicable to software in Debian until it becomes apparent that a particular patent is being enforced or has the potential to be enforced. So the question now becomes whether the Microsoft Royalty Free Patent License is acceptable enough to Debian to allow Miredo to be included in main or non-free, or whether it simply can’t be redistributed at all. The main problem I see with the license is that it does not permit you to sublicense other parties to make use of the Toredo technology used by Miredo. If this were a copyright license then that would most certainly fail to pass the Debian Free Software Guidelines, does Debian treat patents in the same manner? I think it’s time for me to post to debian-legal to seek some “official” guidance on the situation, but hopefully this post has ordered my thoughts enough that I can now write a coherent mail seeking help. If you have any comments or pointers to Debian statements on this sort of situation that I have missed they would be most welcome.

23 December 2005

Matt Brown: Firefox Flicks

If you’re a bit bored I would highly recommend heading over to Firefox Flicks. I haven’t had time to look through all of them yet, but of the 5 minutes I spent I found these, which range from: The Good through to the The Average and finally The UGLY

21 December 2005

Matt Brown: CRCnet Configuration System

CRCnet Configuration System
Now that the CRCnet Configuration System is nearing a releasable state, we’ve had approval from the University’s commercialisation arm that we are allowed to release it under the GPL. This is good, I’ve been working on the system for the last year at least and it is nice to know that it will be available to the community at large. I think what is going to be most interesting to me is to see whether there is any interest in using it from external people and particularly how easy they find it to use. The current plan is to release version 1.0 at LCA’06 in conjuction with our talk of everything that the CRCnet Project has achieved over the last few years. Debian Stuff
Haven’t made a lot of progress with my NM application over the last month. I think both Alexander (my AM) and myself have been too busy with other things. Alexander did get the latest PHPwiki packages that I had prepared uploaded and there have been no new bug reports yet, so I think we’re making progress. This package made use of the wwwconfig-common functions developed by Sean Finney which was a great help and we’re much closer to having a robust PHPwiki package (and one that is up to date) than we were a few months ago. The next step for the PHPwiki packages will be to try and migrate to the dbconfig-common structure for database creation and management. The hold-up there is that dbconfig-common does not support sqlite or sqlite3 at the moment. We use sqlite as the default database for the Debian PHPwiki package to save having to install a heavy-weight system such as mysql or postgresql. Reading the dbconfig-common documentation it seems that it will not be too difficult to add sqlite support so hopefully I will find some time for that over the Christmas break. I wrote some scripts to nicely automate VAP creation with the new madwifi-ng drivers for Atheros wireless cards and posted them to the Atheros mailing list. Kel Modderman picked them up and pointed me towards the pkg-madwifi alioth project where madwifi packaging for Debian is being co-ordinated. It will be good to get involved with that as working with madwifi is a fairly large part of my job at the moment, so it will be easy for me to justify spending time on Debian packaging tasks for it. The key issue we need to sort out here is whether udev rules or ifupdown hooks are the most appropriate way to create the VAP devices, or whether a combination of the two is required.

4 December 2005

Matt Brown: Linux Australasia?

I’m pleased to have prompted a bit of a discussion on the linux-aus mailing list about Linux AU and the relationship with NZ. James Purser kicked things off with a blog post on Friday and so far the discussion is going really well. I’ve outlined a few more of my thoughts below. I think the issue of national identity is probably the key to this whole discussion. It is important for the NZ community and LUGs to retain their own identity and not be subsumed into Australian identity. When Craig, John and I started the Linux NZ discussion, it came in part from a fair degree of envy from what we were seeing Linux AU acheiving. If you read our initial invitation it was quite clear that we were looking for support to “duplicate” Linux AU into NZ. We didn’t get there, and after taking a month or two to think about it, I believe that trying to duplicate Linux AU into NZ was the wrong approach to take. As beneficial as it would be in terms of creating a distinct identity there are far too many other things that it just doesn’t make any sense to duplicate at all. LCA, the podcasts and Advocacy preparation are the three key things that pop immediately to my mind and I’m sure there are others too. I think the biggest thing we learnt from our LinuxNZ endeavours is that we don’t (currently?) have huge numbers of people available to do things in NZ, and we are always going to suffer from problems of scale to some extent. In many ways the more duplication we can avoid the better. That allows us to free what people we do have to focus on the things that are NZ specific. This is the key reason that I’m pushing to see if there is support for Linux Australasia. It’s not that we want LinuxAU to adopt or subsume NZ LUGs, the idea is that LinuxAU would grow to include NZ and become a bigger, better and more inclusive organisation as a result. I guess the flipside for Australians is that the resulting organisation would be less nationalistic…. More on that in a minute. ISON probably fits in here too, I only just become aware them yesterday so I’m still reading their site. However my impression is that while it is definitely an organisation we should be linked with, its aims and goals are orthogonal to what I’m trying to achieve by proposing Linux Australasia. The impression I’m getting from talking to Pia, and lurking on this list is that LA is working very hard to empower the members to be the workers rather than the committee. With the committee providing the organisational oversight, legal structure, vision and authority needed to keep everything hanging together. I think it’s a great strategy, and would fit in perfectly with the bigger, broader Linux Australasia that I’m proposing. Give NZ members the opportunity to work on the press, publicity and advocacy sub-committees too and you’re pretty much there in terms of
achieving Linux Australasia! The sub-committees as a whole can work on their specific areas (avoiding unnecessary duplication) and then the
NZ/AU components of each sub-committee can do the necessary (hopefully minor) localisation work if and when required. Following from this, if there is a need for a more nationalistic component then it’s simply a matter of creating a sub-committee to do the work required to “promote national identity”… perhaps we need to have FOSS competitions across the ditch so we can beat each other in something other than sport… :) Plenty to think about and as many have pointed out the key place to resolve all this is probably going to be at LCA’06 in Dunedin. I think as James has pointed out though, it is going to be very beneficial for us all if we can do much of the talking beforehand and at least get an idea of where everyone is at, even if we are poles apart in terms of what we think.

24 November 2005

Matt Brown: Irony

I had to laugh today while reading the NZ Herald when I came across the an article entitled Using the net to get the story. The article talks about the rise of Computer Assisted Research (CAR) and prominently quotes Francis Till, the webmaster and a technology contributor for the National Business Review. Mr Till says that CAR can be used to "produce better-researched, more in-depth articles"! I wonder if perhaps Mr Till is thinking of well researched (not!) articles like his vitriolic attack on Linux and Open Source Software published in the NBR several weeks ago. Apart from the incredibly unprofessional tone of the article, it appeared to rely almost completely on facts from Microsoft press releases and made many factual assertions that have since been thoroughly refuted. The irony of the author of such a badly researched article wanting to teach other NZ journalists how to use the Internet to research their article is hillarious.

13 November 2005

Matt Brown: My LA LUG Roundup Interview

A few weeks ago I had a chat to Karin Purser of the Linux Australia Podcast Service. Karin regularly compiles the Linux Australia LUG Roundup, which does an excellent job of keeping people informed about what is going on in other LUGs around Australia, and now New Zealand. This month they decided they wanted to talk about WLUG and one way or another I ended up as the person to talk to. The interview is now online and apart from being a bit long winded and sounding a bit poncy in a couple of places I think I did an OK job of describing WLUG and the state of Linux in New Zealand. I need to remember not to talk so much in response to questions in future! Grab it from the following locations: or see the Linux Australia Podcast Service Homepage for links to torrents of both files. I think what I’ve come to realise over the past month is that our efforts to create a Linux NZ type organisation may have been misfounded. Linux Australia is already showing a great deal of support for NZ (LCA’06 in Dunedin, WLUG in the LA Lug Roundup, etc), so perhaps we should be pushing to have this support formalised and extended. Worth thinking about some time in my opinion.

6 November 2005

Matt Brown: Cool new feature in SSH 4.0, Weekend Update

Stumbled across a blog post earlier today on a cool new “connection sharing” feature in SSH 4.0. Basically it multiplexs multiple SSH sessions over a single TCP connection so that subsequent sessions don’t suffer the key exchange and round trip setup delay. Given that I use SSH extensively to a small number of hosts pretty much continuously during my work day this sounds like a very handy feature for me. Read more at: http://little.xmtp.net/blog/2005/11/04/ssh-connection-sharing-c-visibility-share-libraries-and-gcc-attributes-oh-my/ In other knews, I’ve been doing some more hacking on darpwatch as a result of some patches I received, so there will be a new release happening soon there. I’m still battling away against Glade, GTK# and F-Spot in my quest to get all my photos exported out of gallery and into a nice F-Spot + Original setup. Updated PHPwiki packages have been built and have been pending upload for a week or two now, but my sponsor is quite busy at the moment, so there is a bit of a delay at that point inthe path. If you’re desparate for the new packages you can grab them from http://www.mattb.net.nz/debian/.

26 October 2005

Matt Brown: Lightweight Debian Archive Scripts

Hello Planet Debian :) , Steve Kemp wants a solution to magically build uploaded packages. I maintain several small Debian archives and recently needed a solution like this too. The solution needs to provide semi-automated package builds and installation. The full Debian Archive Kit (DAK) was far too heavy for my needs. I ended up creating a couple of wrapper scripts around sbuild and debarchiver to achieve this. The basic flow of packages through my archives goes like this:
  1. Packages uploaded to an incoming directory (/home/sbuild/incoming/<dist_name>)
  2. process_packages script scans incoming directory for each dist every minute and spawns builds (via sbuild) as necessary
  3. Once packages are built site admin is emailed build logs
  4. Admin checks packages and runs add-packages to add the packages to the archive
  5. add-packages invokes debarchiver to do the heavy lifting and then generates the release files.
There is no support for GPG signed packages at this stage, hence the manual check before packages are added to the archive. The scripts were a quick hack a few months ago, and I’ve just hacked them again to pull all the site specific stuff out into easily modifiable variables at the top of the scripts, so hopefully I haven’t introduced any syntax errors! You will need correctly configured sbuild and debarchiver setups prior to running these scripts. If you want to use them, feel free, I’d love to know if they are useful to anyone but me. If there is some interest I’d be happy to tidy them up, package them properly and maintain them. Get them from:
http://www.mattb.net.nz/debian/misc/process_packages
http://www.mattb.net.nz/debian/misc/add-packages I run process_packages from cron with an entry like

*/1 * * * * sbuild /home/sbuild/bin/process_packages &>/dev/null
My /home/sbuild layout looks like

drwxr-xr-x 2 root sbuild 4.0K Jul 31 11:16 bin
drwxr-xr-x 2 sbuild sbuild 12K Oct 26 16:32 build
drwxr-xr-x 6 root root 4.0K Sep 14 12:38 chroots
drwxr-xr-x 4 sbuild sbuild 4.0K Jul 13 09:59 incoming
drwxr-xr-x 2 sbuild sbuild 24K Oct 26 16:31 logs

bin contains the scripts, build contains symlinks to the chroots, and is used as the directory to invoke sbuild from. The resulting .debs are also placed in build. chroots contains the actual distribution chroots, incoming holds a directory for each distribution to process and is scanned by process_packages. logs contains a dump of all the sbuild logs. Debarchiver should be configured to pick up packages out of the build directory. Let me know if I need to provide more details.

24 October 2005

Matt Brown: NBR Article Attacking Open Source

There is an article in today’s NBR attacking Open Source Software, and attacking it quite vehemently. You could almost believe that the writer was on the Microsoft payroll! If you can look past the obvious anti-OSS sentiment that the author holds then there is validity in some of the points, such as around the capability of NZ IT companies to support OSS. Overall the author just comes across as a raving loon in my opinion. Find it at http://www.nbr.co.nz/home/column_article.asp?id=13257&cid=3&cname=Technology

Matt Brown: Choosing Original names for software

I spent some time yesterday evaluating and tweaking the Original (Online Remote Image Gallery Initially Not As Lovely) photo gallery for use on my website which is currently running an old version of PHP gallery. This got me thinking about the naming of software, and particularly about the fact that while Original is a nifty name for an online gallery script, choosing an adjective to use as a noun can have some interesting effects later on! Read on for the full story. I’ve been becoming more and more unhappy with PHP gallery for quite some time now, but I hadn’t found a credible replacement until Original came along. For me, PHP gallery has become far too bloated and featureful to the point where it is excruciatingly slow to perform any task (even browsing) on my gallery. I don’t think my gallery is particularly huge, with 26k photos, spanning 113 albums, it’s certainly at the larger end of the scale, but nothing that I wouldn’t expect some well written software to handle. Couple to this the fact that PHP code like gallery is notorious for being a huge security whole in your server and I was keen to get rid of gallery as quick as possible. I should point out that gallery2 might address some of these issues as it is much more modular than gallery 1 ever was. It still wasn’t enough to tempt me to stay with gallery though. Other well-known galleries like coppermine failed the aesthically pleasing test (not that PHP gallery ever passed it either!). So that left Original. Written by Jakub ‘jimmac’ Steiner, a well known GNOME coder it fulfills most of my requirements: I took me about 2 hours to get Original set up, themed and working to my satisfaction which did involve hacking the code a bit. Mainly to change the layout of a few of the pages, if you weren’t as pedantic as me you would probably get away with just creating your own stylesheet. Now that I have Original setup and ready for photos I have to somehow get all my photos (and preferably their captions and comments!) out of Gallery and into Original! As part of this I would quite like to narrow down the selection somewhat and have my online gallery being a “best of” collection, rather than the canonical repository of my photo collection. My tool of choice to achieve all this is F-Spot, one of the new bread of Mono apps for GNOME sponsored by Novell. F-Spot already supports creating Original galleries, but doesn’t officially support importing from gallery. More on that in another post. To stay true to the title of this post I want to finish up by pointing out that choosing an adjective “Original” to use as the name (noun) of your software, isn’t the best idea in the world. The first time I tried to export an Original gallery from F-Spot I gave up and concluded that it wasn’t supported. I went to the File > Export menu, Choose “Export to Web Gallery” and stopped. This option seems to only support exporting to the very gallery program that I’m trying to rid myself of! I didn’t really expect to find support for Original anywhere else (seeing as it’s a web gallery), but just to be sure I checked the other options, Export to Flickr, no. Export to CD, no. Export to Folder… Maybe. In the Folder Export dialog there is a set of options title “Export Method”, the options are: The first (and second) times I looked at that dialog, despite the fact that I was trying to export photos to Original, and was conciously looking for the option, I completely failed to realise that “Use Original” was what I wanted. My mind was reading the caption as “Use original folder structure” rather than “create folder structure for use by the Original gallery script”. I’d like to file a bug on this to get the wording clarified, but I’m having trouble deciding what it should be. Once you know what it is it seems obvious, the O is capital, so it’s not being used as an adjective, but that is a detail easily thrown away by your mind when scanning the dialog! I’m currently favouring a longer caption such as “Use Original Gallery Folder Format” perhaps with Original in quotes. I also think that moving it into the “Web Gallery” export tab rather than the “Folder Export” tab would clear things up immensely. So, fun and games with software names. What other sofatware uses an adjective as a noun, and does it cause as much confusion as this has?

Next.

Previous.