So, as much of the internet has already noticed, Google/Gmail “helps” users by automatically archiving users own posts to discussion lists when they are received. This is stupid for a number of reasons, which I won’t go in to, but suffice to say that there is no way to convince Gmail of said stupidity.

While puzzling the situation it occurred to me that, at least for my group of people, the receipt of their own postings didn’t have anything to do with the message itself. The real reason we wanted to get our own posts was that we wanted an acknowledgement of our post letting us know everything went ok and that we could expect others to receive the message. Mailman will happily send you a “Your message entitled … was successfully received…” if you set AcknowledgePosts(ack) in your membership options. Because this email looks nothing like the original message you sent, Gmail won’t try to help you out! Magic!

Naturally the next question, as the title implies, is how to do this from the command line. If you’re like me, using Google Apps for your company, this is a problem for all of your users and one that you’d like to fix for everyone in one go. Mark Sapiro has written many helpful mailman scripts that can accomplish this type of task (http://www.msapiro.net/scripts/), but unfortunately this specific option didn’t get any love.

The following script was shamelessly copied/pasted/find/replaceded from the set_mod.py script found at the above url. Usage is described in the comments.

For one reason or another the –all flag of withlist doesn’t work here. Rather than figure it out, just use xargs…

[[code]]czo4OTpcIg0KbHMgL3Zhci9saWIvbWFpbG1hbi9saXN0cy8gfCB4YXJncyAtSSB7fSAuL3dpdGhsaXN0IC1yIHNldF9hY2sge30gLXN7WyYqJl19IC1kIFwnZXhhbXBsZS5jb21cJw0KXCI7e1smKiZdfQ==[[/code]]

Moving the lists themselves wasn’t terribly bad, I just followed this post on debian-adminstration.org: http://www.debian-administration.org/article/Migrating_mailman_lists

That went fine up until I got down to 4.2, where I realized “Man, I’ve got way too many lists to do this by hand.” No biggie, I just modified the command to use xargs:

<pre>
# cd /usr/lib/mailman/bin
# ls /var/lib/mailman/lists | xargs -I listname ./withlist -l -r fix_url listname -u new-domain.example.org
</pre>

This worked quite well to fix the links on all of Mailmans internal html pages… but it didn’t do a thing for the archives. Many of the archived messages had links to attachments at the old domain name. So yeah:

<pre>
# cd /var/lib/mailman/archives/private
# find -type f -name *.html | xargs -I file perl -pi -w -e 's/olddomain\.example\.org/newdomain\.example\.org/g;' file
</pre>

[[code]]czoxMTpcIg0KOmcvXiQvZA0KXCI7e1smKiZdfQ==[[/code]]

FIrst thing you need to do is install dkms, and kernel-devel.

“No precompiled module for this kernel found — trying to build one. Messages emitted during module compilation will be logged to /var/log/vbox-install.log. Compilation of the kernel module FAILED! VirtualBox will not start until this
problem is fixed. Please consult /var/log/vbox-install.log to find out why the
kernel module does not compile. Most probably the kernel sources are not found.
Install them and execute”

Now we can move on to adding the VirtualBox repo:

[[code]]czo5NDpcIg0Kd2dldCAtcSBodHRwOi8vZG93bmxvYWQudmlydHVhbGJveC5vcmcvdmlydHVhbGJveC9kZWJpYW4vc3VuX3Zib3guYXN7WyYqJl19YyAtTy0gfCBycG0gLS1pbXBvcnQgLQ0KXCI7e1smKiZdfQ==[[/code]]

Now you are good to go…

[[code]]czoxMjA6XCINCnJvb3RAbG9jYWxob3N0JCB5dW0gaW5zdGFsbCBWaXJ0dWFsQm94DQoNCiMgVGhlcmUgaXMgYSBjb21waWxlIHN0ZXB7WyYqJl19IGluIHRoaXMgaW5zdGFsbGVyLCBhbmQgbWFrZSB0YWtlIGEgZmV3IG1pbnV0ZXMuIw0KXCI7e1smKiZdfQ==[[/code]]

If you happened to have tried installing VirtualBox and ran into the above “Compilation of the kernel module FAILED!” message, just install dkms then ‘/etc/init.d/vboxdrv setup’ (Compiling will take some time probably).

KVM issues
Apparently the KVM kernel modules(which you would have if you were using the Fedora Virtual Manager) aren’t compatible with VirtualBox. There is a bug filed but as a workaround (or possibly just the easiest solution) is:

[[code]]czozNzpcInJvb3RAbG9jYWxob3N0JCBtb2Rwcm9iZSAtciBrdm1faW50ZWxcIjt7WyYqJl19[[/code]]

Document Management

What is document management? It seems pretty straight-forward to me, but many people have looked at me cross-eyed when I’ve used the term, and some have come right out and said “What do you mean, document management?”. Well, there’s this wiki, but basically I mean a system that gives me the ability to post my files on the web and have a decent amount of access control(user roles, permissions, etc). Naturally there are other features that are not definitive, but essential: Revisioning, search, and metadata is just a start.

There are many document management systems out there, both proprietary and open-source. Of what’s available, all are lacking. What strengths they have are countered with other hard to justify requirements: cost, complexity, administrative overhead.

I’ll take just a minute to talk about two solutions I’ve tried already: Micro$oft SharePoint 2007, and KnowledgeTree Community Edition.

SharePoint

I’ve been using SharePoint for around 6 months now. It’s certainly not bad at the job of document management – and in fact has the ability to go way beyond – given a full time asp.net developer and a fat budget.

I actually had to take a class to learn how to admin the system. Sure I could have probably picked it up on my own, given enough time…but it just speaks to the complexity issue. So, after 2 days of class – the take away was really this:

SharePoint can do anything. You might have to program it to do that, but it can do it. Community of developers? Oh uh, yeah there’s probably one out there… So what does it do out of the box really well? How about locking you into other M$ products – does that count? The honest truth is that out of the box SharePoint has numerous features that are shiny on the outside, but hollow on the inside.

One of the biggest stumbling blocks for our organization was the inability to use any type of authentication besides Active Directory – quite possibly a limitation imposed by the University to keep the system as secure as possible – but a real deal-breaker when you work with 62 different institutions.

KnowledgeTree Community Edition

Since this section of the article is really getting away from me…I’d better take a shortcut and include a part of some email correspondence with a KT sales rep:

My evaluation of KT is going well – although we have hit a few stumbling blocks. The mechanisms that KT uses to do the indexing seem to be overly done, requiring OpenOffice to be running as a background service, which is quite a bit just to allow for the indexing of word documents.

The documentation seems sparse – which is to be expected…but in addition the style of the commands seems to be an old and very specific style (i.e. chown -R nobody.nogroup as opposed to the more universal nobody:nogroup). We’ve had some difficulty getting the scheduler to run due to permission errors as well. From my point of view – not having expert knowledge of the history and reasoning behind the decisions about how KT is built – it seems to have a somewhat polished look on the outside, while being cobbled together on the inside…in need of a total rewrite.

Lastly, the integration with Drupal that I so badly needed didn’t work out very well. After applying the changes, the KT system became inaccessible, and a restore of the files and db from backups was necessary. This isn’t the fault of the KT software I’m sure, but nonetheless affects my ability to use it for my needs.

I’m afraid my KT project is on hold at this point due to time constraints.

I expect I will take it on again when time allows.

Ultimately, the system is fairly good and has a flexible authentication system, but the Drupal integration even if it did work was pathetic at best: can you say “iframe”?

Coming soon: How To: Setting Up Drupal File Framework On Ubuntu 8.10

Reminds me of this XKCD.