Note #2: yes, there are Intel i915 video issues, but both the Debian Squeeze d-i graphical and text installers work fine with this video chipset, however, the screen is blank on first boot – either a) install the SSH Server task and log in via ssh for module and xorg.conf edits, or b) boot into single user mode on first boot with “i915.modeset=0″ to make the edits.

How not to start:

The very first thing I did, after opening the box, was boot the latest Debian Squeeze net install from USB and got to work – I had a heck of a time with R&R Media creation after the fact. R&R 4.3 is quite different than older versions, and I could not boot into it from the Grub bootloader – writing Grub to the master boot record, as I usually do, hosed booting to the Rescue and Recovery partition as well as using R&R fully from within Windows 7 for some reason. Attempting to reinstall R&R 4.3 fails while trying to write the MBR (which is IBM/Lenovo’s suggested fix to MBR problems..)

I finally got the media created on an 8GB USB flash disk, after some dreaded Windows MBR hacking only to find that the R&R USB key was not bootable, and I could not create the Recovery Media a second time – one attempt is all you get.. After mounting the R&R partition (label: Lenovo_Recovery) under linux and digging around, I found the simple edit of $MOUNT/factoryrecovery/service_done.ini from “DONE=1″ to “DONE=0″ allowed the factory reset Recovery Media to be written again. Nice. The trick to getting my USB key to be bootable was to completely zero out the boot blocks of the key, plug it in while running windows and quick format NTFS on it, and then go fire up R&R Rescue Media creation – the boot and data recovery media takes up ~6.5GB on the key. Now that I finally had bootable recovery media, the X201 got a factory reset so I could start over fresh.. all my previous day’s research and gave me some ideas on how I wanted to proceed.

Starting from unboxing the X201:

Note: the goal here is to have a working ThinkVantage button to access the Lenovo Rescue and Recovery partition, as well as a fully functional Windows 7 and Debian Squeeze dual-boot. The key to this is to not touch the Windows/Lenovo MBR. Do not install Grub to the MBR!

Boot Windows 7 and run through the quick user setup stuff. 1) Create the Rescue and Recovery Factory Restore Media – I used an 8GB USB flash disk. (make sure the thing actually boots..) 2) Shrink the Windows7_OS volume from within Windows (Start – right-click on Computer – Manage, then something like storage or disks.. select the Windows partition and shrink it – the space used was about 22GB (sad as there is nothing but the OS, Lenovo tools, and an unlicensed MS Office installed..), so I gave it 40GB. (BTW, ntfsresize within d-i worked fine on my first install, but using the volume shrink within Windows to free up space was simple and certainly assured to be pleasing to windows..)

Boot the Debian Squeeze d-i installer and install as normally in the free space, but do not install Grub to the MBR. Install Grub to the /dev/sdaX device that you set up for the “/” filesystem (or “/boot” if you set up a separate boot partition). I usually use a 256MB USB flash disk with the netinst.iso and non-free firmware for installing on anything that I don’t PXE boot. This is not a d-i lesson – google how to do this.

It is probably possible to re-set the bootable bit on the Windows partition during install – d-i attempted to set the bootable bit on my “/” partition (it’s extended and actually won’t set it), removed the bootable bit from the Windows partition, and upon reboot after install, I got no bootable partitions found. A quick run back into Rescue Mode in d-i, and a toggle bootable on the Windows partition with fdisk is all that took – Windows booted back up fine as normal, the ThinkVantage button works fine, and all is well.. except no default Debian boot, yet.

So as not to screw up the killer ThinkPad feature of the blue ThinkVantage button, this means leaving the Lenovo-customized Windows 7 Master Boot Record intact. I decided to try to use the Windows bootloader to boot to grub on my Debian “/” partition. In my digging around on my Lenovo R&R access issues, I found a reference to EasyBCD by NeoSmart to manage dual-booting Windows versions – this little Windows utility manages the cryptic BCD (Boot Configuration Data database) to easily set up booting a multitude of operating systems. Bravo – trying to do this all by hand was miserable..

After installing EasyBCD, I added a Debian boot menu entry, set it as the default, set a timeout of 3 seconds under “Change Settings”, saved my changes, and rebooted.. Windows boot menu comes up, boots grub from my “/” partition, and everything looks great.. oh yeah, no video

Edit /etc/default/grub and set:

GRUB_CMDLINE_LINUX="i915.modeset=0"

run (as root):

update-grub

Edit /etc/modprobe.d/i915-kms.conf and set:

options i915 modeset=0

run (as root):

update-initramfs -u

Create /etc/X11/xorg.conf (as root) with:

# Minimal xorg.conf for the device driver
Section "Device"
	Identifier	"Default screen"
	Driver		"vesa"
EndSection

Reboot. Perfection. (until Intel gets their video driver support sorted out..)

By the way, this X201 is the type-model 3249-CTO – here’s the lspci output:

00:00.0 Host bridge: Intel Corporation Core Processor DRAM Controller (rev 02)
00:02.0 VGA compatible controller: Intel Corporation Core Processor Integrated Graphics Controller (rev 02)
00:16.0 Communication controller: Intel Corporation 5 Series/3400 Series Chipset HECI Controller (rev 06)
00:16.3 Serial controller: Intel Corporation 5 Series/3400 Series Chipset KT Controller (rev 06)
00:19.0 Ethernet controller: Intel Corporation 82577LM Gigabit Network Connection (rev 06)
00:1a.0 USB Controller: Intel Corporation 5 Series/3400 Series Chipset USB2 Enhanced Host Controller (rev 06)
00:1b.0 Audio device: Intel Corporation 5 Series/3400 Series Chipset High Definition Audio (rev 06)
00:1c.0 PCI bridge: Intel Corporation 5 Series/3400 Series Chipset PCI Express Root Port 1 (rev 06)
00:1c.3 PCI bridge: Intel Corporation 5 Series/3400 Series Chipset PCI Express Root Port 4 (rev 06)
00:1c.4 PCI bridge: Intel Corporation 5 Series/3400 Series Chipset PCI Express Root Port 5 (rev 06)
00:1d.0 USB Controller: Intel Corporation 5 Series/3400 Series Chipset USB2 Enhanced Host Controller (rev 06)
00:1e.0 PCI bridge: Intel Corporation 82801 Mobile PCI Bridge (rev a6)
00:1f.0 ISA bridge: Intel Corporation Mobile 5 Series Chipset LPC Interface Controller (rev 06)
00:1f.2 SATA controller: Intel Corporation 5 Series/3400 Series Chipset 6 port SATA AHCI Controller (rev 06)
00:1f.3 SMBus: Intel Corporation 5 Series/3400 Series Chipset SMBus Controller (rev 06)
00:1f.6 Signal processing controller: Intel Corporation 5 Series/3400 Series Chipset Thermal Subsystem (rev 06)
02:00.0 Network controller: Intel Corporation Centrino Ultimate-N 6300 (rev 35)
ff:00.0 Host bridge: Intel Corporation Core Processor QuickPath Architecture Generic Non-core Registers (rev 02)
ff:00.1 Host bridge: Intel Corporation Core Processor QuickPath Architecture System Address Decoder (rev 02)
ff:02.0 Host bridge: Intel Corporation Core Processor QPI Link 0 (rev 02)
ff:02.1 Host bridge: Intel Corporation Core Processor QPI Physical 0 (rev 02)
ff:02.2 Host bridge: Intel Corporation Core Processor Reserved (rev 02)
ff:02.3 Host bridge: Intel Corporation Core Processor Reserved (rev 02)

If you have 4+GB RAM and are using the i386 architecture, install linux-image-2.6-686-bigmem (I thought about installing x86_64, but I’m currently having to hack around day-to-day issues like using flash and enigmail on my 64-bit desktop install..). If you didn’t set up non-free firmware during install, then go do that. I have the Intel Centrino Ultimate-N 6300 wireless card, so d-i installed the firmware-iwlwifi package for me.

While I’m sure it’s great for most people and cases, if you have the slightest bit of complexity in how you might connect to your mail server(s), the auto-account-creation wizard blows.. With one personal email account (on my own mail server), the account setup wizard worked out most of the settings pretty closely, however, my work account and a couple others went haywire and set up everything completely wrong. If you let the wizard run for any amount of time, it attempts to find servers by DNS and writes the account profile to disk in the wrong place just because some DNS name exists. The trick is to enter your Name and Email account in the very first window, then get ready to mash Stop immediately after continuing on to the next window. Then you can hit Manual Setup and add all the settings correctly and things are written to disk in the correct locations, etc.. There really needs to be a Manual Setup option on that first account creation window that bypasses the auto-account discovery, or a config setting to disable the discovery altogether. I like automagic things, but there always needs to be a sane manual option. Period.

Next up is the thread pane headers (thread, starred, sender, date, etc.).. In TB2, you set the headers in one view, and all folder views are the same. I can understand people wanting to have different views, but there is no option to set all views the same, so every single folder has to be modified.. This sucks when you have lots of folders and I have tons of them. There is a workaround – do not log into the account on the first prompt – set the thread pane headers you want on the Inbox, and most (not all) will inherit the Inbox view. Then log in later and fix the few incorrect views that remain (Drafts, Trash, etc.).

While you have the account created and have not logged in the first time, yet, you might want to dig through the account settings and decide if you want a few things changed – particularly under Synchronization & Storage.. no, I don’t want to keep all messages locally – I use IMAP – uncheck that box. I really don’t want to store gigs of messages locally by default – the whole point of IMAP is to keep the messages on the server. I also read some notes from people that use POP3 and (used to) “keep messages on server” – the default TB3 POP3 account creation pops all the messages off the server and deletes them from the server – be aware.. (and you should really be using IMAP, if you want this behavior anyway)

As of writing, the GPG/OpenPGP plugin, Enigmail, does not have a functional x86_64 Linux port for TB3 in the Add-Ons widget.. sigh.. There is a contributed port built on Arch Linux on the Enigmail download page that works for me. This should be fixed soon in the official Add-Ons widget, I suppose.

Some of my about:config modifications for sane behavior (Edit – Preferences – Advanced – Config Editor):

• mail.check_all_imap_folders_for_new; true
• mail.operate_on_msgs_in_collapsed_threads; false
• mail.tabs.autoHide; true
• mailnews.display.prefer_plaintext; true
• mailnews.nav_crosses_folders; 0

OK, it’s just about time to actually log in and fetch mail! A bit of IMAP header fetching and on to reading some mail, changing the toolbar buttons to small, and then I found the message header collapse is gone.. Ah, there is a nicely working add-on for that: CompactHeader by jmozmoz Thank you, thank you!

And on restart of TB3, I end up in whatever folder I was last viewing on shutdown.. and very odd behavior of password prompting, if it was a nested folder (duplicate login prompts..). And no where is there an option to configure “Start if my f’ing Inbox”. This is retarded, in my opinion – who asked for this “feature” anyway? Once again, there looks to be a sort-of-working plugin for this (and once again, this is a manual download/install, at the moment): Select Inbox by Bogdan Rechi This works ok in starting up in the default account Inbox, but the weird password prompts still remain – looking at the plugin code, this is run at start time – my suggestion for TB3 is for the plugin to switch to the default Inbox at shutdown, and let TB3 just start normally. When I start up TB3 after hitting my Inbox prior to shutdown, I don’t get the strange duplicate login prompts at startup.

The default “New Messages in Folder” text color (they dumped the little star..) and the basic highlighting on folder pane, thread pane, message pane selections and focus are crap – almost unreadable blue text on blue background for “New Messages” on the default Debian Gnome theme – blue selected folders look nearly identical to the blue selected+focus folders. The colors are essentially just pulled from the OS color theme, which is fine, but it is very difficult to differentiate where the focus is – I use this visual queue a lot when scrolling up/down with the keyboard while drilling through folders or message threads. The only way to configure this is by setting up custom CSS rules via userChrome.css. My ideas came from the Mozilla Folder Pane Color for New Mail notes and from this killer resource for TB mods – here is my custom userChrome.css:

/*  Tree Text Highlighting (folders, thread pane, address pane, etc.)
    Highlight focused items, but use black on lightgrey when not in focus */
    treechildren::-moz-tree-cell-text(selected) {
        background-color: lightgrey !important;
        color: black !important;
    }
    treechildren::-moz-tree-cell-text(selected, focus) {
        background-color: Highlight !important;
        color: white !important;
    }
    tree[selstyle="primary"] > treechildren::-moz-tree-row {
        border: none !important;
        background-color: transparent !important;
    }
    treechildren::-moz-tree-row(selected) {
        background-color: lightgrey !important;
    }
    treechildren::-moz-tree-row(selected, focus) {
        background-color: Highlight !important;
    }
/*  Folder Pane Color for New Mail */
    treechildren::-moz-tree-cell-text(folderNameCol, newMessages-true) {
        color: black !important;
    }
    treechildren::-moz-tree-cell-text(folderNameCol, newMessages-true, selected) {
        color: black !important;
    }
    treechildren::-moz-tree-cell-text(folderNameCol, newMessages-true, selected, focus) {
        color: white !important;
    }

That about covers my fixes for my Thunderbird 3 sanity – if I need to set it up again and remember any other changes, I’ll add them in.

Update: I added in the config option to disable the new brain-dead collapsed thread display – nice try, Mozilla, but you can’t actually expand it and navigate with the keyboard, so FAIL.. mail.operate_on_msgs_in_collapsed_threads; false

mshuler@gaea:~$ file Life.Of.Brian.1979.avi
Life.Of.Brian.1979.avi: RIFF (little-endian) data, AVI, 640 x 336, 25.00 fps, video: XviD, audio: MPEG-1 Layer 3 (stereo, 48000 Hz)

or not..:

mshuler@gaea:~$ file Three.Days.of.the.Condor.1975.mkv
Three.Days.of.the.Condor.1975.mkv: Matroska data

A quick look at the ‘mplayer’ man page gave me the options I needed to get all the properties without playback of the file:

mshuler@gaea:~$ mplayer -identify -frames 0 Three.Days.of.the.Condor.1975.mkv
MPlayer SVN-r30656 (C) 2000-2010 MPlayer Team
Playing Three.Days.of.the.Condor.1975.mkv.
ID_CHAPTER_ID=0
ID_CHAPTER_0_START=7000
ID_CHAPTER_0_NAME=The American Literary Historical Society
ID_CHAPTER_ID=1
ID_CHAPTER_1_START=557633
ID_CHAPTER_1_NAME=Out to Lunch
ID_CHAPTER_ID=2
ID_CHAPTER_2_START=1125799
ID_CHAPTER_2_NAME=Condor
ID_CHAPTER_ID=3
ID_CHAPTER_3_START=1649965
ID_CHAPTER_3_NAME=The Ansonia Hotel
ID_CHAPTER_ID=4
ID_CHAPTER_4_START=1990799
ID_CHAPTER_4_NAME=Still in Flight
ID_CHAPTER_ID=5
ID_CHAPTER_5_START=2430633
ID_CHAPTER_5_NAME=Some Safe Quiet Time
ID_CHAPTER_ID=6
ID_CHAPTER_6_START=3135166
ID_CHAPTER_6_NAME=Sam's Apartment
ID_CHAPTER_ID=7
ID_CHAPTER_7_START=3594665
ID_CHAPTER_7_NAME=Lonely Pictures
ID_CHAPTER_ID=8
ID_CHAPTER_8_START=4166199
ID_CHAPTER_8_NAME=The Mailman
ID_CHAPTER_ID=9
ID_CHAPTER_9_START=4772166
ID_CHAPTER_9_NAME=Sparrow Hawk
ID_CHAPTER_ID=10
ID_CHAPTER_10_START=5111499
ID_CHAPTER_10_NAME=The Suspicion Business
ID_CHAPTER_ID=11
ID_CHAPTER_11_START=5316665
ID_CHAPTER_11_NAME=Leonard Atwood
ID_CHAPTER_ID=12
ID_CHAPTER_12_START=5778032
ID_CHAPTER_12_NAME=Fine Qualities
ID_CHAPTER_ID=13
ID_CHAPTER_13_START=6068332
ID_CHAPTER_13_NAME=Oil Fields
ID_CHAPTER_ID=14
ID_CHAPTER_14_START=6338665
ID_CHAPTER_14_NAME=For that Day
ID_CHAPTER_ID=15
ID_CHAPTER_15_START=6653499
ID_CHAPTER_15_NAME=Telling Stories
ID_VIDEO_ID=0
[mkv] Track ID 1: video (V_MPEG4/ISO/AVC), -vid 0
ID_AUDIO_ID=0
ID_AID_0_LANG=eng
[mkv] Track ID 2: audio (A_AC3), -aid 0, -alang eng
ID_SUBTITLE_ID=0
ID_SID_0_LANG=eng
[mkv] Track ID 3: subtitles (S_VOBSUB), -sid 0, -slang eng
[mkv] Will play video track 1.
Matroska file format detected.
VIDEO:  [avc1]  704x352  24bpp  23.976 fps    0.0 kbps ( 0.0 kbyte/s)
ID_FILENAME=Three.Days.of.the.Condor.1975.mkv
ID_DEMUXER=mkv
ID_VIDEO_FORMAT=avc1
ID_VIDEO_BITRATE=0
ID_VIDEO_WIDTH=704
ID_VIDEO_HEIGHT=352
ID_VIDEO_FPS=23.976
ID_VIDEO_ASPECT=2.3693
ID_AUDIO_FORMAT=8192
ID_AUDIO_BITRATE=0
ID_AUDIO_RATE=48000
ID_AUDIO_NCH=2
ID_LENGTH=7032.90
ID_SEEKABLE=1
ID_CHAPTERS=16
==========================================================================
Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
Selected video codec: [ffh264] vfm: ffmpeg (FFmpeg H.264)
==========================================================================
ID_VIDEO_CODEC=ffh264
==========================================================================
Opening audio decoder: [ffmpeg] FFmpeg/libavcodec audio decoders
AUDIO: 48000 Hz, 2 ch, s16le, 448.0 kbit/29.17% (ratio: 56000->192000)
ID_AUDIO_BITRATE=448000
ID_AUDIO_RATE=48000
ID_AUDIO_NCH=2
Selected audio codec: [ffac3] afm: ffmpeg (FFmpeg AC-3)
==========================================================================
AO: [alsa] 48000Hz 2ch s16le (2 bytes per sample)
ID_AUDIO_CODEC=ffac3
Starting playback...
Exiting... (End of file)
ID_EXIT=EOF

For one, it’s difficult for me to provide a completely objective review – this device is intended to evoke emotion and that’s exactly what Steve Jobs wants – you are *supposed* to fall in love with the iPad. Secondly, I do not own an iPhone nor an iPad – I’ve played with other people’s iPhones quite a bit and have not seen an iPad in person, yet. So, if you absolutely adore your iPhone, compare every other handheld device to it, and within 30 seconds declare that all other handhelds are bred of lesser mobile genes, sleep with it not just on the nightstand, but perhaps under the pillow, then you may wish to stop reading.. I have an opinion or two.

First and foremost, Apple is crazy brilliant at marketing. Everyone is iPad-caffeinated right now because deliveries have started and people now have these devices in their hands and flooding the twitterverse with their excitement – same has happened with each iPhone hardware and software release. There is love in the air, and Apple loves you for loving them. People do not do the same with ThinkPad hardware releases and Windows updates (unless MS breaks shit, which happens..)

There is a lot to love with the iPad! I don’t want to be a party-pooper, far from it – I want an iPad, too! The form factor is killer – this type of device will definitely change the way people interact with the web, social media, medical patient rounds, watch movies, etc. The physical hardware engineering – the beautiful touchscreen, the weight, the *feel* (yeah.. when I actually get to feel one), the applications – these type of devices *will* change the Internet world, so I would like one, too, please.

Will I buy one? Should you? That depends.

I probably won’t – not unless Apple comes out with a compelling Gen2 iPad.. Look around at some of the (non-iFanBoy) reviews in the last couple days: no camera (the chassis has the identical insert as iPhone’s camera CCD, so it will eventually get one), no multi-tasking, difficult data transfer, etc. For me to be compelled, at the very least, Apple needs to allow me to run more than one application at a time, if I would consider it anything more than a browsing toy. I think the #1 design failure of iPad was basing the OS off the iPhone OS, as opposed to a mobile port of OS-X, but I understand the business decision. The iPad is a fantastically gorgeous toy and if that’s your thing, it is something you should absolutely have! Go get one! Just don’t expect to do any actual work on it, unless your work is directly related to the social media arena – even then, I would almost guarantee that you will still do most of your *real* work on a laptop or desktop. After all, Apple doesn’t want to kill their laptop business, do they?

http://www.businessinsider.com/i-really-hate-what-apple-is-trying-to-do-with-the-ipad-2010-4

“If it can’t print what’s the point..?” – Can your iPhone print? Yeah, that’s what I thought. With the Gen1 iPad, you get essentially a neutered iPhone (it can’t make calls and no camera) with a big screen and a few extras like a very, very cool eReader and the potential to run iWork. You can absolutely work on documents, presentations, etc., but you can do that efficiently only if you do it the Apple way. If you have existing documents you wish to load up on the iPad, I understand from second-hand knowledge that it is rather difficult. Things will get better, and people will figure out how to do some work on the iPad (do you *work* on your iPhone?) – there will be some cool apps that make interesting use of the screen real estate, for sure, but..

The apps.. you really only get one choice of where to shop for your applications. And the iTunes App Store has full control over the extent of your freedoms. If you keep your entire music and video media collection in iTunes, and if you are fine with paying for almost every song and every piece of software you use (some are free), and if you either you don’t know anything about, nor understand, nor really care about software freedom, then by all means keep doing what you are doing – it is a system that really does work well. Apple does this very well – the Apple ecosystem is very functional, very pretty – even sexy, but you are bound to it by design. This ecosystem (and that’s really what it is – a technological island environment with a wonderful array of cohabitating hardwares and softwares that compliment one another – if you have the ability to build your own boat and can brave the waters, you can do some cool things outside of Apple Island) does not go well with my philosophy and beliefs in Open Source Software (google it, non-techies..) and the extreme idiocy of a great deal of software patents granted in the last 10-ish years. (sidenote: New Zealand is considering a patent reform bill that bans some forms of software patents)

http://computerworld.co.nz/news.nsf/news/thumbs-down-for-software-patents-in-nz

So where I am going with this.. buy an iPad, but don’t expect it to be something that it is not. You might want to wait until they come out with the next version – the iPhone 3G kicks the pants off the original.

I truly believe that this type of device is the perfect (although not yet perfected) device for non-technical users – I’m fairly convinced that my mom would learn to love it. You really don’t have to think much, once you figure it out; the UI and the UX are wonderful; the hardware/software island leads to little (I hope) in the way of trouble from trojans and other exploits.

As a good schizophrenic, yeah, I would love to have an iPad. I love computing devices and I could certainly find ways to use an iPad that suit it’s design very well and it would be fun. Most likely I would hack it – I’ll build my boat (using boat plans of more talented boat designers than I), see how far I can travel off the island, and wave back at the curious looks from the sheep on the beach.

I’d guess that more likely, if I decide to spend money on a keyboardless tablet, I will get something in the same form factor, but one that allows me to have the ability to use Open Source Software, one that allows me to run an email client *and* browse the web at the same time(!), and most definitely one that has a kick-ass SSH client (is there a good one for iPad?) – much of what I do is on remote servers, so SSH is an absolute must.

Update: I understand there are a couple really good SSH apps – this alone might be the “killer app” for me.

http://eu.techcrunch.com/2010/03/19/apple-ipad-how-about-a-little-german-innovation-instead/

I’m a bit of a nutcase when it comes to free software (free as in speech, not beer) – just ask my mom how fired up I get in conversations about technology.. I breath it. I tattooed it on myself. So take my words with a grain of salt, if you wish – these are only the opinions of a highly opinionated technology advocate.

First, copy the example config.php out of the docs/ directory to the root of the PMWiki site, in my case, I installed to $DOCROOT/pmwiki/:

cd $DOCROOT
cp pmwiki/docs/sample-config.php pmwiki/config.php

Following the Clean URL notes, in pmwiki/config.php I set:

$EnablePathInfo = 1;

then I hit the wiki front page and started to look at the resulting link URLs like http://wiki.example.com/pmwiki/index.php/Main/WikiSandbox – they all 404 at this point, but this gives an idea of what I want to do next. I want to get rid of the /index.php/ bit, first, so in pmwiki/config.php I set:

$ScriptUrl = 'http://wiki.example.com/pmwiki';

This worked – the resulting URLs are like http://wiki.example.com/pmwiki/Main/WikiSandbox – but, let’s make that a little more portable for other vhost names or aliases and set it to:

$ScriptUrl = 'http://'.$_SERVER['HTTP_HOST'].'/pmwiki';

This works perfectly – now to take care of all the 404s with nginx rewrites. From yesterday’s config, I added the “location /pmwiki” and “location @pmwiki” blocks – here’s my complete nginx server configuration:

server {
  listen 80;
  charset utf-8;
  server_name wiki.example.com;
  access_log /var/log/nginx/wiki.access.log;
  root /var/www/wiki;
  index index.html index.php;

  location /pmwiki {
    error_page 404 = @pmwiki;
    log_not_found off;
  }

  location @pmwiki {
    rewrite ^/pmwiki/(.*)$ /pmwiki/?n=$1 last;
  }

  # configs from pmwiki .htaccess files
  location ~ ^/pmwiki/(cookbook|local|scripts|wiki.d|wikilib.d) {
    deny all;
  }
  location ^~ /pmwiki/docs {
    autoindex on;
    types {
      text/plain php;
    }
  }

  # pass PHP scripts to FastCGI server listening on 127.0.0.1:8080
  location ~ \.php$ {
    fastcgi_pass 127.0.0.1:8080;
    fastcgi_index index.php;
    fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
    include /etc/nginx/fastcgi_params;
  }

  # deny access to all .dot-files
  location ~ /\. {
    deny all;
  }
}

If PmWiki is installed in the document root of the site (not under http://…/pmwiki/), then adjust the ScriptUrl in config.php:

$ScriptUrl = 'http://'.$_SERVER['HTTP_HOST']'';

and the nginx config blocks above to:

  location / {...}

  location @pmwiki {
    rewrite ^/(.*)$ /?n=$1 last;
  }

  location ~ ^/(cookbook|local|scripts|wiki.d|wikilib.d) {...}

  location ^~ /docs {...}

The vast majority of web software packages assume that the admin is running Apache. It is always fun and interesting to sort out how to incorporate the various configurations for nginx, when most of the software packages only document mod_rewrite rules or just throw things in .htaccess files.

This is a very simple install of PmWiki focused only on getting the nginx configuration right. This assumes that you already have an nginx site configured to run PHP scripts through fastcgi – I’m not documenting that here.

Grab the source tarball from the Download page on http://pmwiki.org/, or use Subversion to fetch the latest stable release ($DOCROOT is your site document root):

cd $DOCROOT
svn export svn://pmwiki.org/pmwiki/tags/latest pmwiki --force

Now, hitting the URL in a browser ($HOST is your site, of course) – http://$HOST/pmwiki/pmwiki.php – you should see a little note about PmWiki needing write access to the document store directory – a quick permissions change:

chmod 2775 pmwiki/

Then hit the URL again – you should see the Main page of PmWiki after it does it’s configuration but no so fast! Secure that directory now that we are basically done installing:

chmod 0755 pmwiki/

Follow the notes for adding a one-liner wrapper index.php if you like.

OK, so PmWiki should be working, but probably has some hidden configurations:

$ find pmwiki/ -name .ht*
pmwiki/scripts/.htaccess
pmwiki/cookbook/.htaccess
pmwiki/local/.htaccess
pmwiki/docs/.htaccess
pmwiki/wiki.d/.htaccess

Yep.. each of those .htaccess files, except pmwiki/docs/.htaccess, is a “Deny from all” so that files in those directories cannot be accessed publicly. I also added the directory pmwiki/wikilib.d/ which has the default local PmWiki documentation files and other local wiki pages – OK.

pmwiki/docs/.htaccess is a modification of mime type, so that any files in that directory ending in ‘.php’ will be rendered as plain text – ah, the sample config.php is in there – OK.

Here is my resulting nginx vhost configuration for wiki.example.com (replace with your own server_name and site root directory):

server {
  listen 80;
  charset utf-8;
  server_name wiki.example.com;
  access_log /var/log/nginx/wiki.access.log;
  root /var/www/wiki;
  index index.html index.php;

  # configs from pmwiki .htaccess files
  location ~ ^/pmwiki/(cookbook|local|scripts|wiki.d|wikilib.d) {
    deny all;
  }
  location ^~ /pmwiki/docs {
    autoindex on;
    types {
      text/plain php;
    }
  }

  # pass PHP scripts to FastCGI server listening on 127.0.0.1:8080
  location ~ \.php$ {
    fastcgi_pass 127.0.0.1:8080;
    fastcgi_index index.php;
    fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
    include /etc/nginx/fastcgi_params;
  }

  # deny access to all .dot-files
  location ~ /\. {
    deny all;
  }
}

This should give a working default PmWiki site running on nginx with the same functionality and secured directories as if you installed on Apache – make sure this basic setup is working right before moving on.

Next configuration steps for my install will be rewrite rules for shortened clean URLs.

Update: here’s the next post on PmWiki Clean URLs on nginx

One of my favorite apt preference resources is maintained by Roderick Schertler, and I always keep his apt pinning page listed in my /etc/apt/preferences file.

I track Debian stable (currently “Lenny”) on production servers and upgrade only software that is absolutely necessary from backports or testing – the gotcha I mention above is when Debian releases a new “stable” (“Squeeze” is next). Without “oldstable” listed in the preferences file at a higher priority than “stable”, you might be in for a surprise one day with a long list of packages set to be upgraded to the new “stable” This is not particularly desirable in a production environment that will require a good deal of work to make sure a dist-upgrade to the new stable release will function properly.

On my workstations and laptop, I typically roll along with the “testing” (currently “Squeeze) apt repositories and update/upgrade daily for that new software fresh feeling. I use the same apt configurations and simply bump the testing preference to 910. This allows me to do things like ‘apt-cache policy $FOO’ to see what versions of a package are in all the releases, without trolling around packages.d.o.

Here’s a tarball of my Lenny apt configs and the contents:

/etc/apt/apt.conf – set lenny as the default release and bump the cache limit a lot higher

APT::Default-Release "lenny";
APT::Cache-Limit 33554432;

/etc/apt/sources.list – Lenny/testing/sid along with security, backports, debian-multimedia, and alpha kernel builds

# Lenny
deb http://ftp.us.debian.org/debian/ lenny main contrib non-free
deb-src http://ftp.us.debian.org/debian/ lenny main contrib non-free
deb http://security.debian.org/ lenny/updates main contrib non-free
deb-src http://security.debian.org/ lenny/updates main contrib non-free

# Testing
deb http://ftp.us.debian.org/debian/ testing main contrib non-free
deb-src http://ftp.us.debian.org/debian/ testing main contrib non-free
deb http://security.debian.org/ testing/updates main contrib non-free
deb-src http://security.debian.org/ testing/updates main contrib non-free

# Sid
deb http://ftp.us.debian.org/debian/ unstable main contrib non-free
deb-src http://ftp.us.debian.org/debian/ unstable main contrib non-free

# Experimental
deb http://ftp.us.debian.org/debian/ experimental main contrib non-free
deb-src http://ftp.us.debian.org/debian/ experimental main contrib non-free

#####
# Lenny Backports
deb http://www.backports.org/debian lenny-backports main contrib non-free
deb-src http://www.backports.org/debian lenny-backports main contrib non-free

# Debian Multimedia
deb http://www.debian-multimedia.org/ lenny main
deb-src http://www.debian-multimedia.org/ lenny main

# buildserver.net kernel buildd repo
# http://wiki.debian.org/DebianKernel
deb http://kernel-archive.buildserver.net/debian-kernel/ trunk main
deb-src http://kernel-archive.buildserver.net/debian-kernel/ trunk main

/etc/apt/preferences – the magic to not dork up a stable box.. read Roderick’s pinning page for really great explanations on how this all works

Explanation: see http://www.argon.org/~roderick/apt-pinning.html
Package: *
Pin: release o=Debian,a=oldstable
Pin-Priority: 905

Package: *
Pin: release o=Debian,a=stable
Pin-Priority: 900

Package: *
Pin: release o=Debian,a=testing
Pin-Priority: 400

Package: *
Pin: release o=Debian,a=unstable
Pin-Priority: 300

Package: *
Pin: release o=Debian
Pin-Priority: -1

Package: *
Pin: origin www.backports.org
Pin-Priority: 600

Package: *
Pin: origin www.debian-multimedia.org
Pin-Priority: 600

Package: *
Pin: release o=Debian-Kernel,a=kernel-dists-trunk
Pin-Priority: 200

Brilliant!