Planet Arch Hurd

Spork Bomb: Converting grammars to automata

Michael Walker — 2011-06-09T11:42:56+00:00

Wow, ages since I last wrote a blog post. April, in fact. Well, I have today been working on something that has the potential for fun! I have been making a little Python script using pygraphviz (which is wonderful) to convert formal grammars into automata. Currently it can only do regular (in right-regular form) and context-free (in CNF) grammars, but I want to add support for more.

Here is an example of the sort of glorious pictures you can generate:

right-regular
S → aA | bB
A → aC | bA
B → aS | bA
C → aB | λ

context-free
S → AC | BD | λ
A → 0
B → 1
C → SA
D → SB

Pushdown automata generated accept by empty stack.

You can see the code, and my to-do feature list, on github.

Spork Bomb: More information on the Arch Hurd Netbook Remix

Michael Walker — 2011-04-01T00:00:36+00:00

If you have not done so already, first read the official announcement of the Arch Hurd Netbook Remix. I shall now explain some of the motivations for this remix.

The Myth of Wireless

Firstly, there seems to be some conventional wisdom that netbook users require wireless support to be able to use their netbooks efficiently when not next to a wired network connection. This is false! What is a netbook? A computer. What is the purpose of a computer? To perform mathematical computations. The presence, or lack thereof, of a network connection has no bearing on the efficiency of the performance of these calculations!

Constant network availability is merely a mental crutch for those who are unprepared! A real developer will do all of their research when a network connection is available, and so not need one at other times!

Bluetooth? Just as bad.

Almost as bad as the 'need' for wireless connectivity is bluetooth. Ladies and gentlemen, I beg you, do not be fooled by the different name. Bluetooth is just wireless networking under another guise, and so I shall not grace it with any further explanation.

USB? Overrated.

What is USB most commonly used for on a netbook? External storage. What does this external storage do? Allow you to store more than you can on your HDD, and to exchange files with others. I shall now proceed to explain why both of these need not be done with USB.

Firstly, storing extra files. If you cannot fit all of your files on the internal HDD, clearly you did not plan your purchase properly, and so you have nobody but yourself to blame for your lack of space. Shame on you!

Secondly, exchanging files. There is no good reason why this cannot be done by the following process:

Shut down netbook
Insert bootable USB drive
Boot to live environment
Mount any appropriate partitions and transfer data
Reboot back into Arch Hurd

This is, in fact, a procedure I use daily.

Release Date

Arch Hurd Netbook Remix shall be released after Hell freezes over, but before the heat-death of the universe.

By the way, have you noticed the date today?

Spork Bomb: 8-bit computer

Michael Walker — 2011-03-01T18:48:00+00:00

Over the past week or so I've been designing a simple 8-bit computer on paper and have recently implemented an assembler and emulator for it, so I've decided to share it with the world. It's not yet finished, as I haven't designed the memory (though I know what I'm going to design, so it's implemented in the emulator). It's a simple little thing with only 14 instructions, 2 (programmer-accessible) registers, and no support for indirect addressing (which is a pain).

Instructions are a byte wide, the lower four bits of which are the opcode, and the upper 4 bits of which are the operand. The address and data buses are eight bits wide, and so memory access is performed using offsets, with an instruction to adjust the base offset. This makes programs particularly susceptible to label-related fails, as a label is assembled to an offset from the start of the segment it is in. Below is a program to calculate fibonacci numbers and save them to memory location 0xFF (which will be an output location in the final design) until an overflow occurs:

        ;; Calculate all fibonacci values that will fit in 8 bits (ie: 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233)
        ;; and write them to the output location.
        ;; This code assumes that output is written to 0xFF (which it will be, in the final design).

start:  setmdev tmp             ; Write first fib number (saved in tmp/one) to output
        load one                ;
        setmdev outh            ;
        save    outl            ;

fib:    setmdev tmp             ; Write to the output location
        load one                ;
        setmdev outh            ;
        save    outl            ;

        setmdev tmp             ; Calculate the next fib value, and if there was an overflow terminate
        load    one             ;
        sum                     ;
        save    two             ;
        load    one             ;
        copya                   ;
        load    two             ;
        jc      end             ; Carry is set if an addition of A and B *as the previous instruction* would cause a carry (regardless
                                ; of if the previous instruction was a sum). Thus, merely by having the values
                                ; in the registers, we can see if we've reached the end of our computation

        save    one             ; Save values to the temporary location

        setmdev 0x0             ; Calculate the next
        jmp fib                 ;

end:    setmdev outh            ; Print final output
        save    outl            ;

trap:   setmdev 0x1             ; Halt
        jmp trap                ;

one:    DATA[0x01]              ; Temporary location 1
two:    DATA[0x00]              ; Temporary location 2

        CONSTANT[tmp,0x1]
        CONSTANT[outh,0xF]
        CONSTANT[outl,0xF]

And here is the assembled version of that (in hex):

10 71 F0 F2 10 71 F0 F2 10 71 07 82 71 03 81 3C 72 00 4D F0 F2 10 5D 01 00

Not bad.

The documentation for the CPU, ALU, and memory currently resides in a org file in this git repository. There is a rather nice emulator which presents an ncurses interface so you can run programs you build with the supplied assembler. Here is a screenshot of the emulator running the above program.

All I have left to design (down to the level of logic gates) is the memory, and I already have a high-level view of how that will function. After that, I might start a second, improved, design. I've learnt a lot about implementing computers since I started this, and want to put this new knowledge to good use.

Also, there is the possibility of a compiler for some high level language in the future, as keeping track of memory offsets by hand and not having any indirect addressing is frustrating.

Spork Bomb: New mirror in France

Michael Walker — 2011-02-02T17:55:44+00:00

We have a new mirror in France; an abundance of European mirrors! This time it's hosted by the Institut de Génétique Humaine, many thanks to them!

Mirror URLs

Spork Bomb: m3u Playlists on the Cowon J3 in MSC Mode

Michael Walker — 2011-01-22T12:08:10+00:00

On Thursday my Cowon J3 arrived, it has epic sound quality (makes my headphones sound better than my speakers, and the equaliser is brilliant) and battery life (64 hours of music playback) and, importantly for me supports Vorbis and Flac, and has MSC (UMS) mode, so I can use it with GNU/Linux.

In short, it's brilliant.

Less brilliantly, the player itself has no functionality to make new playlists, this has to be done with a computer and the playlist copied over to the device. I'm told PLA playlists work great in MTP mode, though I'd rather use MSC so I set about investigating how to make m3u playlists. Turns out it's not so simple, they have to be in a specific format, so I don't think Cowon can even claim to support m3u (properly, at least).

In short, the playlists have to have certain metadata in comments, and use relative paths (I'm told absolute paths work, but I couldn't get it to behave). Thus, assuming your playlists are stored in \Playlists\, this format should work:

#EXTM3U
#EXTINF:539,Die Verbannten Kinder Evas - 05 - Beneath The Veil Of Ocean
..\Music\Music\Die Verbannten Kinder Evas\1995 - Die Verbannten Kinder Evas\Beneath The Veil Of Ocean.ogg
#EXTINF:175,Die Verbannten Kinder Evas - 11 - Craving Dreams
..\Music\Music\Die Verbannten Kinder Evas\1995 - Die Verbannten Kinder Evas\Craving Dreams.ogg
#EXTINF:344,Die Verbannten Kinder Evas - 03 - Darkened Skies
..\Music\Music\Die Verbannten Kinder Evas\1995 - Die Verbannten Kinder Evas\Darkened Skies.ogg
#EXTINF:92,Die Verbannten Kinder Evas - 14 - Das Letze Kapitel
 ..\Music\Music\Die Verbannten Kinder Evas\1995 - Die Verbannten Kinder Evas\Das Letze Kapitel.ogg

And another example:

#EXTM3U
#EXTINF:93,Rozen Maiden OST - 01 - Kinjirareta Asobi (TV SIZE)
..\Music\Soundtracks\Rozen Maiden\Rozen Maiden\Rozen Maiden OST\01. ALI PROJECT - Kinjirareta Asobi (TV SIZE).ogg
#EXTINF:140,Rozen Maiden OST - 02 - Battle of Rose
..\Music\Soundtracks\Rozen Maiden\Rozen Maiden\Rozen Maiden OST\02. Mitsumune Shinkichi - Battle of Rose.ogg
#EXTINF:90,Rozen Maiden OST - 03 - Komatta Shumi
..\Music\Soundtracks\Rozen Maiden\Rozen Maiden\Rozen Maiden OST\03. Mitsumune Shinkichi - Komatta Shumi.ogg
#EXTINF:133,Rozen Maiden OST - 04 - Atatakana Kokoro
..\Music\Soundtracks\Rozen Maiden\Rozen Maiden\Rozen Maiden OST\04. Mitsumune Shinkichi - Atatakana Kokoro.ogg

That is, it has to start with an EXTM3U comment, telling it that there is metadata. Each file is then specified with two lines, one containing the play length in seconds and the album/track/name. The second line is a relative path to the file. I currently have loads of playlists for most combinations of artists/albums I'll want to listen to, as I don't know how to do that with the J3 itself (yet), though I'm sure that's just me failing to figure out the player. However, I shall continue hoping for a firmware update which makes playlists easier (well, possible) to manage with the player itself, as that would be awesome and I would then have no complaints with the player.

Spork Bomb: Year of Arch Hurd

Michael Walker — 2011-01-14T15:33:56+00:00

Earlier this year, Arch Hurd became one year old! A momentous achievement for something I never intended to start in the first place. In this past year we've had people come and go, more and more packages ported, overcome problems, released LiveCDs and, most importantly, progressed from merely an idea to a reality. It's hard to think now that Arch Hurd had such humble beginnings.

Some notable achievements over the past year have included:

Getting Arch Hurd to be self-hosting

When we first started with Arch Hurd, we cross-compiled it from Arch Linux with a variety of scripts initially provided by Allan. From that base we scripted and compiled, 'til we had a native toolchain running within Arch Hurd along with pacman! We started to package all the cross compiled stuff and now all Arch Hurd packaging is done from within Arch Hurd.
Porting the Arch Linux initscripts

Once upon a time every Arch Hurd user booted to single user mode and didn't have niceties such as a VGA console or passwords. Melpo decreed that this was not to be! And so set about the task of porting the Arch Linux initscripts, with all their goodness, to Arch Hurd. Admittedly, there was probably more rewriting than porting, and they still don't work perfectly, but we have Arch Linux-like goodness, complete with rc.conf—which has to be a good thing.
Producing a LiveCD

In May, the first ever Arch Hurd LiveCD saw public release after much confusion on my part after researching how LiveCDs actually work, though it was only an early prototype and didn't feature a setup program; however, in June an updated LiveCD with installer was made available for download. There have now been three LiveCDs, with another on the way once a few outdated packages have been updated. Additionally, the next LiveCD might be the first ever bug free one (though I'm not holding my breath).
Packaging Xorg

Giselher, in May, succeeded in getting all the packages required for Xorg and Openbox in extra. This was a big step forward as the Hurd terminal tends to freeze after a time, a fate to which X seems immune.
Attended Software Freedom Day

I and hayashi attended a Software Freedom Day event in Manchester in September. Though not as fun as I thought it would be, it was still nice to have an Arch Hurd box running for people to have a look at (if many of the looks were slightly amused).
Packaging everything for GHAMP (GNU/Hurd, Apache, MySQL, and PHP)

Again done by giselher (our packaging overlord), in August everything was packaged which we need for GHAMP. A brilliant achievement, as the Hurd website runs on a Hurd (Debian, I assume) server, though I don't know of many servers running Arch Hurd in the wild (I think our mirror on hackthehardware.net does).

Of course, that's only the things I consider most notable. We have also got a lot of packages (mostly owned by giselher), a (what I consider) nice website, users hanging out in IRC, the same users caring enough to host mirrors… but most of all, we've had fun. The Hurd isn't the easiest system to work with, and a lot of stuff needs patching to run properly (or even at all), it's a complete pain sometimes but, fundamentally, we do this because it's fun.

Spork Bomb: First Term at University

Michael Walker — 2010-12-25T01:38:01+00:00

Well, as I have now (and have for a few days) finished my first term at university, I think it's apt to reflect upon what I've learned, achieved, and done during that time, and my plans for the next term.

Firstly, obviously, there is the course; the reason I went to university. During this past term I've improved my rather basic knowledge of Java to a pretty good one, as alas Java is used as the teaching language. I've learned quite a bit about the maths behind comp sci (well, mostly logic thus far) and algorithms, as well a bit about the history of computers and how they work. Next term I intend to continue learning, and just see what happens.

Secondly, there's the whole living-away-from-home thing, something I've never done before. The first week or so felt pretty surreal, but now it just feels something like staying in a hotel, and perfectly normal. I've got on well with my flatmates, learned that I have zero inspiration when it comes to cooking (beyond chilli and incredibly bizarre meals), and that ironing is boring and completely unnecessary if you just fold things or hang them up. I can proudly say I've only ironed once, and intend to keep it that way :P

Thirdly, there's the social aspect of university. I've joined a number of societies, including the Kendo club—Kendo being a rather fun martial art that involves screaming at people and hitting them with big sticks—and also plan to join more, such as the Anime and Manga society, next term.

It's all been quite fun.

Spork Bomb: Arch Hurd Isn't Dead.

Michael Walker — 2010-11-12T21:39:36+00:00

Despite appearances, Arch Hurd isn't dead. I know there hasn't been any news for months, and barely any updates in that time either, but we've all been busy. Back in late September / early October many of us went off to our respective universities, which cut into Hurd-time. In short, we've all been very busy, please forgive us that. However, that doesn't forgive the prolonged lack of progress. That happened months ago, we've had plenty of time to adjust and resume development, but we haven't. I take the blame for that; as project founder I should have sooner seen fit to try to organise people into working again. It's our birthday in a couple of months, so I want us to be able to point to Arch Hurd and say "Look at that, look what we've made."

There's a meeting in IRC on Sunday (probably at 15:00 GMT) for bug-squashing and update-building. If you want to drop by and say how awesome we are, please feel free :)

Spork Bomb: New Arch Hurd Mirror in the USA

Michael Walker — 2010-09-25T17:13:12+00:00

Arch Hurd now has a mirror hosted by Parabola GNU/Linux who, as per usual, we are very grateful to. In addition to being another mirror, this is our first mirror in the United States—so it should be a bit faster for our American users (assuming we have any).

Parabola GNU/Linux is a completely free distribution based on Arch Linux, and remains compatible with Arch so changing to Parabola is as simple as changing your mirrorlist and updating. It is a very good project for any Archers who wish to stick to Free Software.

Spork Bomb: DreamHost to Linode

Michael Walker — 2010-09-16T22:13:28+00:00

Today I switched from DreamHost to Linode. Well, almost, I still have one website to move over, and that's only because I can't get in touch with the friend I'm hosting it for. The whole thing went pretty smoothly, well, other than Postfix which I fought with for over hour hours before discovering that it was actually gmail that was misbehaving. But I got that sorted out.

Admittedly, a Linode is more expensive than DreamHost. But I think I can cope, especially if I start charging a few people who've been using my webspace a small amount to cover the extra cost to me. And, really, the benefits of a VPS more than make up for the additional cost anyway.

I have had a good experience with DreamHost, and would recommend them to anyone looking for shared hosting, but my needs have outgrown what they can provide. I do know about DreamHost PS, but if I'm going to pay extra for a VPS, I may as well go for a VPS company I've used before (for archhurd.org) and which I know are very good. The only downtime archhurd.org has had, is downtime I've caused. I hope that record will continue with my new Linode (aptly named Yuggoth).

Spork Bomb: mkahurd: Make Arch Hurd LiveCDs and QEMU images easily.

Michael Walker — 2010-09-11T23:14:58+00:00

Download

Over the past couple of days I've been working on a script to produce Arch Hurd LiveCDs, and then decided to expand it to support QEMU images (and soon HDD installs). It is called, unsurprisingly, mkahurd, using my usual talent for names.

% mkahurd --help
USAGE: mkahurd [-v] (--livecd | --qemu) /path/to/working/directory

Options:
    -v:       Verbose output (must be first argument)
    --livecd: Generate a livecd
    --qemu:   Generate a 10GB qemu image

mkahurd is written and maintained by Michael Walker.

Currently it's a bit picky about the order of the parameters (in that -v must be the first parameter, if specified), though I plan to work on that. Producing both ISOs and images is pretty simple:

ISO

% time mkahurd --livecd ~/hurd/tmp/live
==> Downloading contents of core...
==> Installing packages...
==> Installing setup script
==> Building LiveCD...
==> Done. ISO built as /home/barrucadu/hurd/tmp/live/ahurd.iso.
> For official livecds, do the following:
>     * Add documentation (installation guide & translator intro)
>     * Set up helpful /etc/motd file
Real: 876.35s User: 12.69s System: 3.64s Percent: 1% Cmd: mkahurd --livecd ~/hurd/tmp/live

QEMU

% time mkahurd --qemu ~/hurd/tmp/qemu
==> Creating image...
Password: 
==> Installing 'base'...
==> Installing 'base-devel'...
==> Finishing off...
==> Done. Image built as /home/barrucadu/hurd/tmp/qemu/ahurd.img.
> You will need to install GRUB yourself using a GRUB boot image.
> No configuration has been performed.
> First boot will fail. You will have to MAKEDEV hd0 and hd0s1 when dropped to a recovery console.
Real: 421.67s User: 14.60s System: 2.24s Percent: 3% Cmd: mkahurd --qemu ~/hurd/tmp/qemu

The commands do take a little while to run, as a lot of packages have to be downloaded. In addition, to update a LiveCD, you can just run mkahurd --livecd again with the same working directory, and it will run much faster (as not so many packages need to be downloaded).

Hayashi's Blog: Emacs! EMACS!

2010-09-04T22:04:34+00:00

emacs-nox, thanks to a patch contributed by diegonc, is now in [extra]. Do your duty to your favourite GNU distribution and use/test it!

Quite frankly, now we have a form of emacs, I fail to see why any more userland packages should be ported. With emacs, we already have a complete userland! >_>

Spork Bomb: Arch Hurd does Software Freedom Day

Michael Walker — 2010-09-04T15:08:47+00:00

On Saturday the 18th of September, I and Matt (hayashi) will be at the Software Freedom Day event in Manchester; located in Madlab, lasting from 1000 to 1600. If you're going to the event, or will just be in the area (or city), drop by and say hi. Plans include:

Demonstration Machine: We plan to bring a computer running Arch Hurd for people to have a look at. Haven't quite figured out the logistics yet (or which of us will be bringing the computer), but we'll definitely have one.
LiveCDs: I'll bring a stack of (working) live cds to give away to anyone who wants one. I'm not sure how many to bring, but if I run out I'll just distribute copies of the ISO (if you have some form of portable media with you).
Meeting People: This'll probably be what we'll spend most of the time doing, and it should also be the most fun. If you're in the area on the day, come and say hi :)
Keysigning: if you have proof of identity, I'll make a note and sign your GPG public key when I get home after the event.

I've never gone to a free software event before, so this'll be a nice gentle start to our plan of showing up and giving a talk at ArchCon 2011 next year. It should be a fun day :)

Hayashi's Blog: Namcap in testing; Meritous delayed indefinitely.

2010-09-02T17:00:02+00:00

Due to Meritous requiring sdl_mixer (a sound library with lots of sound-related dependencies), its current nearly unplayable performance on the Hurd, and on the advice of barrucadu regarding the first issue, I’ve decided to indefinitely postpone pushing Meritous to repos until sound packages migrate to [extra]. Having already done a bit of blind hacking on the Meritous source, I might try to fork a version with less sound hardcoding and push that to repos.

In other news, the version of namcap (PKGBUILD and pkg.tar.* analyser) that was previously in AUR has been moved to [testing], with what I perceive to be issues regarding dependency checking and permission checking. If I get any input regarding this, I’ll be happy to try and fix any issues.

In other news, corewars, gtk(1) and glib(1), dfm (file manager I used for testing gtk1) and lynx are all now in [extra].

Peace out, ~Matt.

Hayashi's Blog: Hi there

2010-09-02T12:43:13+00:00

Hello, everyone reading this blog and the Arch Hurd planet.

I’m Matt Windsor, or (captain)hayashi, and I’m a new Arch Hurd developer (or, at the moment, package maintainer). Over the forseeable future I’ll be handling more or less trivial ports of mostly game and multimedia-related packages from Arch Linux.

I feel that, with enough love and attention, the GNU system can be a gaming platform too. I’ve already had some degree of success with my port of scummvm [extra] which, using sdl [extra] and when coupled with old adventure games such as the now freeware Beneath a Steel Sky (bass [AUR]), provides a very playable (if silent) experience.

Beneath a Steel Sky on GNU

From Wikipedia,

Beneath a Steel Sky is a British 1994 science-fiction point-and-click adventure game in the cyberpunk genre… [it] takes place at an unknown point in a dystopian future in Australia, where the Earth has been significantly damaged by pollution or nuclear fallout. The game’s backstory is introduced via a comic book (shown on-screen in the introduction sequence of the CD release), drawn by well-known comic artist Dave Gibbons, that tells the story of a young boy called Robert who is the sole survivor of a plane crash in “the Gap” (the name applied to the Australian Outback at the time of the game). Too young to fend for himself, Robert is adopted by a local group of Indigenous Australians, who teach him the skills he needs to survive in this harsh new environment; they name him Robert Foster, partly due to him being fostered by them and also because of the discovery of an empty can of Foster’s Lager, an Australian beer, found near him at the crash site. Foster even learns engineering and technology and builds a talking, sentient robot called Joey.

After Foster has reached adulthood, he is kidnapped and his tribe annihilated by storm-troopers sent from Union City by its all-powerful computer LINC. Interestingly, Union City mentions prominent suburbs and train stations found within Australia’s largest city, Sydney, leading some to speculate that Union City was once Sydney. This was confirmed in a 2005 interview with the Australian gaming magazine, PC PowerPlay.

Foster manages to escape from his captors as the helicopter transporting him back to Union City crashes just after entering the dome, leaving him and his robot friend, Joey, to find out why they were brought there and where to go next, while security continues to search for him.

tl; dr it’s an old-school point-and-click game that I personally am quite fond of. Due to Revolution making the game resources freeware (though possibly not strictly free software – one of the reasons the game files are in the AUR) and contributing code to ScummVM, a free software meta-engine for adventure games, the game (originally for DOS and Amiga OCS) is playable on such platforms as modern Windows, GNU/Linux, the PSP (ick)… and now, thanks to a trivial PKGBUILD port, Arch Hurd.

I don’t have any screenshots (yet!) but, using ScummVM with scaling turned off, Beneath a Steel Sky runs on the Hurd at a quite playable framerate. The main problem (which seems to stand for all SDL programs at the moment) is that the longer the system has been running, the less likely SDL programs are to work. Plus, SDL seems to be singularly slow.

Other game experiments

So far, I’ve also built openttd [extra], a free software clone of the simulation game Transport Tycoon Deluxe, and meritous, a 2D action game with a somewhat Japanese atmosphere. Both run but, being based on SDL, are almost unplayably slow. Hopefully this will be resolved eventually.

In testing GTK1, I ported corewars [extra], but do not know how to play it and have thus not properly tested it. Let me know if it works for you!

In closing

Hopefully, meritous should be in the repos this week. For now, I’m going to experiment with building SDL with nasm support, though I’m sceptical of the benefits.

Giselher's Blog: What’s next?

2010-08-31T11:22:48+00:00

I finished PHP yesterday, so what’s next? Well, to be honest I like to convince more people to maintain packages for Archhurd, because It gets harder every day to maintain the packages we have and starting packaging new things are more time consuming than update packages you know already.

If you are interested in helping us you can start with dropping by in our IRC channel #archhurd on chat.freenode.net. There is always someone activ in this channel and fear not we are always friendly and do not eat your cat.

procfs

I wrote about procfs earlier and that it lacks some important feartures. You may be pleased to hear that Jeremie König a GSOC student that is working for Hurd, completly rewrote procfs. I only have to wait for melpo to return from his vacation and that he will rebuild hurd with the new procfs code. If it offers all features that we know from linux than we can iron out a lot of bugs and I can start building top, htop and other tools.

Giselher's Blog: Status & storeio – My new favorite translator

2010-08-25T23:54:32+00:00

Hy guise,

I didn’t worked much on PHP and other server technologies, because the live cd is on its way and I did update some more or less essential packages, which are used for the live cd or are popular by the users. By the way! I am so happy that I finally fixed the python package. Python worked for me, because I accidently install it to root and not to the pkgdir and I thought that I uploaded the fixed package long ago. But then a friend of barrucadu, who wanted to build namcap to test his packages, told me that python still doesn’t work. After dowloading some patches from debian we now have both, a working python build and namcap in the AUR for now.

Package Status

There is also a new syslog-ng package in testing, which does not depend on pidof. Due to the lack of a working /proc implementation pidof does not work. But fear not, work is going on the source code of procfs. We sooner or later got a working /proc and all the nice things that are depending on it, like top.

To be more developer friendly for future hurders, I added cvs and mercurial. We now cover almost every VCS, except for Subversion, which will surely come someday (if I have enough time).

Wiki Status

I also change some wiki pages, especialy the translator page, which I will quote below. This pages gives an overview about the most important ones for end user, if there are end users at the moment. Please have a look if you like. http://wiki.archhurd.org/wiki/Translators

I will make this post a bit longer than, because I didn’t post yesterday

storeio – A flexible block device translator

I’d like to present my new favorite translator, by the name storeio.

storeio is a translator which creates a block device from the given file. It also supports compressed files with gunzip and bunzip2 but I recommend not to use any compression, didn’t work as expected and takes forever to mount. You sure are asking yourself what you can do with a block device? Well, once created you can created a filesystem on it and mount it on a directory and put data into it. Because of the fact that the file is now a block device, you’re able to work with filesystem tools, like mkfs and fsck, on it. You can also create a ramdisk witk it.

Enough wtih the talk, lets show some examples. From here on I will quote the wiki page about translators. Sorry for the format. Look at our wike page if you like to see it in a better format.

storeio

storeio is like losetup on linux, but with more advantages. It also supports compression with gzip and bzib2 or normal files.

It creates a block device with the given file or arguments.
 settrans -ca loop0 /hurd/storeio -T file FILE
 settrans -ca loop1 /hurd/storeio -T gunzip FILE.gz
ATTENTION: storeio can not size the given file. A possibility to create a file with a given size is to use dd.
 dd if=/dev/null of=FILE bs=42MB
This will create an empty file with the size of 42MB. You can now work with the new block device. You can use it as temporary filesystem (ramdisk) or as portable filesystem.

ramdisk
 settrans -ca ramdisk0 /hurd/storeio -T copy zero:32M
 mkfs.ext2 -F -b 4096 ramdisk0
 settrans -ao ramdisk0 /hurd/ext2fs.static ramdisk0
First we create a buffer with copy and getting the size with zero. (Just like dd, but without invoking a file) Then we create the filesystem on our blockdevice and mounting it with ext2fs. The argument -o (orphan) replaces the previous translator with a new one without forcing it away.

portable vfs

The following example shows how to create virtual filesystems in a file.
 dd if=/dev/zero of=your.img bs=SIZE
 settrans -ca loop0 /hurd/storio -T file your.img
 mkfs.ext2 -F -b 4096 loop0
 settrans -ca mount /hurd/ext2fs.static loop0
Fist create a file with any size and create the block device with storeio then create the filesystem and mount it.

You can also lay the translators on top of each other.
 settrans -a your.img /hurd/storeio -T file your.img
 settrans -ao your.img /hurd/ext2fs your.img
The file your.img is now a directory and you can use it as such.
 cp your things your.img/
 ls your.img/things/
To unmount the filesystem.
 settrans -g your.img
ATTENTION: I noticed that if you lay one translator onto another you can only remove the one on top. I don’t know if this is a feature or bug, but after that procedure you only need to use ext2fs to mount the image, which is pretty handy after all.

Giselher's Blog: HAMP (Hurd Apache MySQL PHP)!

2010-08-23T07:14:33+00:00

After hours and 100 hacks later I finally got apache and its dependencies built, except for libsasl and libldap, which I thought are impossible to build for Hurd. But then diegonc, a archhurd user that works on the xkb keyboard drivers, told me that he used a patch to build libldap. Now I can rebuild apr-util and apache again and thats not easy, because apache compiles itself three times in the archlinux PKGBUILD. I don’t know why, but I do it too.

Apache PKGBUILD

     for mpm in prefork worker itk; do
                if [ "${mpm}" = "itk" ]; then
                        <---snip--->
                        autoconf
                fi
                mkdir build-${mpm}
                pushd build-${mpm}
                ../configure --enable-layout=Arch \
                        <---snip--->
                make
                if [ "${mpm}" = "prefork" ]; then
                        make DESTDIR="${pkgdir}" install
                else
                        install -m755 httpd "${pkgdir}/usr/sbin/httpd.${mpm}"
                fi
                popd
        done

The good news is, that packages like IMAP and MySQL are building without any major problems. I hope that we soon can setup a HAMP system to host a website or something similar

Giselher's Blog: First Commit

2010-08-22T18:36:06+00:00

I finally got a blog. Well, I did have one for a python game I wrote, but I did not a single post on it.

I will write about the status of my Arch Hurd packages, the progress on my devtools for Arch and general Linux and GNU/Hurd stuff. I might also write about other kernels like Minix and other stuff I feel like writing.

To also add usefull information to this post, I shall announce that I am currently working on packaging apache and PHP. This is rather big task because they have a lot of dependencies I have to build first in order the build apache and PHP.

For the current status of this packages look at my wiki page (http://wiki.archhurd.org/wiki/User:Giselher#Apache.2FPHP).

I also have some ideas for my devtools, like combine my addpkg and addgit script into one script, which does also a git commit.

See About for further information about me.

lnz's posterous: My personal state of the hurd

2010-08-20T15:07:00+00:00

I've been a developer for Arch Hurd for over 6 months now and I like to think that I'm finally getting the hang of things. In these last few months I've seen a few new articles on the Hurd here and there, however, all of them focus on historical developments and RMS. Not that the history of the Hurd isn't an interesting one but I feel like there is very little information out there about the actual current state of affairs. Ok, thats not quite right, there are great sources of information like the Hurd open issues page and the bug-hurd ML (especially this recent thread). But today, for the first time in my life, I felt like blogging so I'm going to share my perspective.

Real World Usability

The big points are all nicely discussed in the aforementioned bug-hurd thread.

A quick summary of the major things missing:

SATA
USB
Sound (Not really that serious but many people, me included, really want to be able to play music :D )

These are the things I consider the worst problems. Not only are they somewhat important, they will also require a lot of work to implement. Many other issues are, at least imho, fixable with some time devoted to them. This includes things like modern browsers and many of the open issues. The long lasting problem of bad NIC support is starting to look brighter too, I don't dare say the issue is solved but current progress is looking great.

I feel like I should explain in more detail why I find the points listed above to be such serious problems. Of course you can still use IDE,PS/2 and maybe you can even live without music. I think we can all agree on the fact that IDE and PS/2 are slowly dying, accordingly it becomes rarer for people to have Hurd-ready hardware at home. Please don't misunderstand this, I'm not seeing the normal home-user as a future Hurder but I am of the opinion that many interested people, who want to contribute, have a hard time doing so because they simply can't get it to run on the hardware they have.

That said, I did recently build myself a hurd box from old parts I had and everything works fine so I guess things don't look too dim yet.

A short word on my experiences with my hardware install, which is pictured somewhere in this post. I've been running it for a little over a week now and things actually work pretty well. I mainly use it for developing and package building. Occasionally the whole machine just freezes, sadly I haven't had any success in debugging this yet but I'll keep trying. X worked effortless for me(others are reporting problems though), we haven't got a lot of X programs in our package repos yet, so I don't really want to say too much about it, maybe some other time.

All in all I have to say that the Hurd certainly isn't as unusable as many make it out to be. However, the system of course isn't flawless yet, it does have stability issues and individual programs don't work quite right.

tl;dr In my opionion the Hurd is pretty much usable for motivated enthusiasts.

The Mess

Now that we got the usability covered I can go to things that suit me better, like ranting. In my opinion maintaining packages for a Hurd system is a terrible mess. I may be alone with this opinion and it may just be a lack of skill on my side, anyway this is how I see it:

Building the same software that builds perfectly on Linux may cause you terrible suffering on Hurd. Sometimes the configure scripts need a little sed magic, sometimes you'll have to find patches hidden in the deepest depths of the internet (or if you're lucky the Debian patch-tracker), sometimes you'll have to patch things yourself and sometimes everything works fine. And this is after the situation has already gotten a lot better thanks to Debian/Hurd.

The best example I can think of is glibc. It starts with the fact that upstream releases will only build after severe patching. There is a
hurd glibc repository that builds correctly but there are a lot(~20) of additional patches that should be applied to make glibc as complete as possible. So in the end, to have a decent glibc, you have to get the source from a non-upstream repository, apply a large amount of patches and maybe even fiddle with the configure flags a little. And it's not as straight-forward as it sounds either.

For me, having such a central part of the system such a state, seems like a very bad situation. I have to admit though, I don't know anything about the reasons for upstream glibc being like this.

Luckily no other packages are as bad as glibc in this respect. However, in the end package maintaining requires a lot more time than it would on a Linux based system, which in turn leaves us with a lot less time that we could use for development and/or fixing real issues.

In retrospect I remember why I don't blog. It looks like a lot of stupid rambling now.

But I guess (and hope) that the internet has seen worse.

Permalink | Leave a comment »