Exim4 spitting out errors from cron

November 21, 2014

This is a quick post for anyone else who might get exim4 bugs from cron, particularly on Debian. To be specific, this bug; https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=501892 is biting me. I went through all the mail to that bug and tested some of the software used and I think I have a handle on what happened in my case.

I tested exim_tidydb against all the databases in the /var/spool/exim4/db directory, and they all were tidied without error by exim_tidydb. I had another db however in there, called greylist.db. That database was created by hand according to instructions found here; https://github.com/Exim/exim/wiki/SimpleGreylisting

However, looking the messages from exim_tidydb when you call it directly, it seems to state that you can *only* call it on databases it knows about;

# /usr/sbin/exim_tidydb /var/spool/exim/ greylist.db   
Usage: exim_tidydb [-t 

As you can see from the message above, there are a number of databases named on the last line and it appears from the behavior of the program as well as from the error message that if you use a name that isn’t listed, you’ll get an error, error 123, which gets passed along to cron.

I don’t think its a bug, except perhaps in the documentation URL which I’ve added above and filed an issue on.

Comments Off on Exim4 spitting out errors from cron

More on the “Doing Data Science” book

December 18, 2013

We’re off to a good start. In the forward I’ve read this;

“Building models and working with data is not value-neutral. You
choose the problems you will work on, you make assumptions in those
models, you choose metrics, and you design the algorithms.”

I think this is really important, we need to address things like confirmation bias right away so that we don’t put too much faith into what we see in our numbers. I’m glad I’m going to be reading “Thinking Fast and Slow” by Daniel Kahneman to balance all the magic numbers floating in my head.

Comments Off on More on the “Doing Data Science” book

O’Reilly reader review program

December 16, 2013

I review for the O'Reilly Reader Review Program O’Reilly publishing has an interesting Reader Review program, or at least a new one. It allows ordinary O’Reilly readers to get electronic versions of books and review them, pretty straight forward right? I’m planning on reviewing “Doing Data Science” because I feel that I don’t know how to do that and I hope this book will tell me.

Link to book at O'Reilly There’s the cover and here’s the description;

Now that people are aware that data can make the difference in an election or a business model, data science as an occupation is gaining ground. But how can you get started working in a wide-ranging, interdisciplinary field that’s so clouded in hype? This insightful book, based on Columbia University’s Introduction to Data Science class, tells you what you need to know.

Comments Off on O’Reilly reader review program

Planetizing foss-gbg

November 28, 2013


We’re creating a planet of FOSS hackers in Göteborg. If you blog (Swedish or English), feel free to send an email to the mailing list; foss-gbg@lists.pelagicore.com. You’ll need a Hackergotchi!

Mailing list info: http://lists.pelagicore.com/cgi-bin/mailman/listinfo/foss-gbg

Comments Off on Planetizing foss-gbg

Installing Dancer, a perl web framework, on Debian Wheezy

February 19, 2011

This post is going to describe what I did to get Dancer up and running on Debian Wheezy. Dancer is a web development framework and as such aims to help you develop web applications more quickly. Dancer also supports psgi, or “Plack”, which is a tool to glue together your web application with your web server, making the two more interchangeable. Contrast this with mod_perl for example which is tightly coupled to the Apache web server.

I use Debian Testing (codename Wheezy) on my workstation at home, but Dancer is not yet available for Debian Testing although it should be showing up in testing in a day or two. I’m going to drop into a sid chroot to install Dancer since it isn’t in testing. A sid chroot is a chrooted version of Debian which is very useful for testing and writing software. A chroot essentially creates a separate root file system and encloses that root file system in a separate directory. If something goes wrong in this separate root file system it often has no impact on the other parts of your file system, at least if you are careful. This makes it ideal for setting up new tools or testing software. Setting this up is a little beyond what I had intended to talk about so I’ll leave it to the other very good sources of info about Debian chroots on the web to explain.

Like all good Debian perl packages you can install Dancer on Debian with this simple call;

sudo apt-get install libdancer-perl

Now what?

But what happens next? Well, if all else fails, read the documentation. In this case, there is plenty. One of the great things about Perl is that there is so much good documentation that comes with it and often the same can be said for Debian. Usually when I want to learn about a particular perl module on my Debian system I do;

perldoc Foo

And in this case I did perldoc Dancer and up came the perl documentation. There are also Debian specific documents and examples in the usual Debian location; /usr/share/doc/libdancer-perl/. I didn’t even bother looking in the examples directory which is included amongst the documentation, but rather I cut and pasted the code right from ‘perldoc’. I mean, if it is as easy as the author claims, I should be up and running without issue right?

And sure enough, I was. Dancer is remarkably easy to set up. I’ll go through the remaining steps I took to test a web app on my machine. Firstly, I cut and pasted the example code in the perl documentation which is a simply hello world program;

    # webapp.pl

    use Dancer;

    get '/' => sub {
        "Hello There!"

    get '/hello/:name' => sub {
        "Hey ".params->{name}.", how are you?";

    post '/new' => sub {
        "creating new entry: ".params->{name};


Once I created the perl file called mywebapp, which contained the code above, I ran

dancer -a mywebapp

Then all you have to do is to run the resulting web app like this;

perl ./mywebapp

which will handle all of the web serving for you. This truly saves a lot of time and headache.
You can change the last “Dancer->dance” bit with the more simple “start;” if you think Dancer->dance is silly. There is also a micro blog called dancr which may be of interest. Anything more sophisticated that what I’ve show here is most likely going to be better explained by the official documentation. Point your browser to the pod on CPAN and there ought to be something helpful there.

Dancer is simple to set up and get running.


Nokia’s white flag of surrender

February 5, 2011

While we wait for the new Nokia CEO to come out and explain his plans for the struggling phone giant, we get to wildly speculate. This speculation hasn’t been bad for Nokia’s share price but it will be a bit embarrassing for Nokia if some of the speculation proves true.

The speculation, even reported in the NY Times, is that Nokia is going to abandon their software platform(s) for a new one, at least in North America. This speculation has some basis, after all the new CEO, Mr. Elop, has a background at Microsoft and surely the Finns on Nokia’s board were looking at that when they hired him. The departure of Ari Jaaksi to HP and the WebOS is another sign that homegrown platforms like MeeGo won’t be getting as much attention inside Nokia.

Nokia is a business, and they are wise to revamp their operations that have won them only a two percent market share in North America. But if they move to Windows Phone 7 platform, or even to Android, it is an admission that they cannot compete in software in the world’s largest and most lucrative market. This would be the first time that Nokia would run someone else’s software on their phones. Furthermore, it calls into question their other strategies: what happens to MeeGo, what happens to Ovi, what happens to Symbian?

I can’t imagine a move like this will be successful for Nokia. Windows phone software has a tiny market share and combined with Nokia’s tiny market share you get . . . a still tiny market share. While the move may please shareholders, it is bound to disappoint a much more important constituency: Nokia’s customers. Nokia does make great hardware, and an Android phone from Nokia would most likely be pretty cool, but how are they going to compete? On price? Yikes. That won’t be good for the profit margins. And people are staying away from Windows phones in droves, they can’t get rid of the things despite a marketing budget that would pull some countries out of their deficit. Nokia phone buyers are pretty loyal, will this brand loyalty be in jeopardy if they get some cruddy phone that doesn’t even compete well with Symbian?

I’m not unbiased here, I believe in MeeGo and its future. It has a more open model than Android and that helps development scale. If Nokia abandons their investment in MeeGo it is a white flag of surrender, a way of saying “we can’t compete.” That seems like bad business to me.

Comments Off on Nokia’s white flag of surrender

maemo.org turned over to the community

September 17, 2010

In a bit of very good news, Nokia has released maemo.org governance over to the Maemo community. There already is a maemo community council so they will be the governance body (vote ongoing now).

While Maemo is sadly an officially “dead” OS, it may be given new life if the community can take over and run it. There are still lots of excellent and quite dedicated hackers and members in the community so I see no reason why this shouldn’t be a vibrant project. It can also share source code with other projects like MeeGo and Linaro and can potentially help Maemo keep pace with MeeGo. I do have some concerns with the OBS but there are other ways to build packages.

This is good news and I look forward to being more active again in a more open community.


Adobe becoming more open?

August 1, 2010

I noticed in a recent SourceForge update email that Adobe has set up a site on SourceForge for their open collaboration initiatives: http://sourceforge.net/adobe/

This is a great idea. Adobe should leverage the power of the Free Software community to make Flash perform better on Linux, these improvements will surely be able to make Flash overall a better netizen plus an overall better performer – even on non-Free platforms like the iPhone, er iOS, thingy. I think Adobe would be wise to brag about this site a little bit and market their openness a bit more, it binds them more tightly to the community and positions them as an alternative to closed, proprietary pseudo standards like Silverlight.

Comments Off on Adobe becoming more open?

DebConf in NYC

July 15, 2010

I'm going to DebConf10


Comments Off on DebConf in NYC

Incompetence and falsehood from American industry

February 27, 2010

It is perhaps comforting to know that some believe American Empire is not doomed to irrelevance. But much of American Industry seems bent on trying. Aside from their usual shabby treatment of their customers with lawsuits and wholesale purchase of politicians through lobby organizations, they love to write law that suits their complacency, hoping to snuff out competition through litigation.

Now that American Industry has seen that other nations can catch up, many companies are trying to change the rules to keep their advantage. One way they do this is through trying to force the US Trade Representative to ban trade with other countries because they have no use for crappy software from Microsoft. Now, they don’t mind if the country doesn’t pay for the software as long as they use it because they know at some point these foreign governments and companies will get caught out and be forced to pay Microsoft for some license or other. What these trade organizations, like the International Intellectual Property Alliance (IIPA), do not want to see is a choice in the marketplace. Choosing something other than a product from one of these organizations threatens the high paid lawyers and ex-CEOs who can’t find work from earning a handsome living making fools of themselves in the press.

Recently the IIPA, whose members include the BSA, RIAA, and MPAA, (litigious hobgoblins all), issued a report trying to strong arm the USTR into banning countries that use “Open Source” software. The report is conveniently available in pdf (a proprietary format of course) on their web site. A quick gander at that didactic screed brings us some useful nuggets, also highlighted by the Gaurdian. Of course the Guardian, like any good news organization, got it’s news from a blogger, who deserves the real credit.

I point out the Guardian article because they have a useful quote from the report so you don’t have to go through the whole thing:

“The Indonesian government’s policy… simply weakens the software industry and undermines its long-term competitiveness by creating an artificial preference for companies offering open source software and related services, even as it denies many legitimate companies access to the government market.

Rather than fostering a system that will allow users to benefit from the best solution available in the market, irrespective of the development model, it encourages a mindset that does not give due consideration to the value to intellectual creations.

As such, it fails to build respect for intellectual property rights and also limits the ability of government or public-sector customers (e.g., State-owned enterprise) to choose the best solutions.

There is a distinct and glaring lack in these paragraphs from the report. In fact I would say there is a complete absence of truth. Let’s help the author(s) with some corrections shall we?

The first paragraph goes off the rails early with the claim that recommending open source software weakens the software industry. Far from it – it strengthens the software industry, particularly the software industry in Indonesia. Because you are allowed to study and modify Free Software, you can learn about programming and become a better programmer. You can create innovative products and sell them. Perhaps even to the United States? It follows that the long-term competitiveness of the industry would be furthered by the use of Free Software in Indonesia, at least the Indonesian software industry. And call me crazy but I believe sovereign nations have a right to their own software industry.

What the heck is an artificial preference they refer to in the first paragraph? Is that the opposite of the “natural preference” of Microsoft software? Somebody better run and tell the busiest web sites in the world, along with two-thirds of the rest of the web, to stop using the apache web server – it just ain’t natural. They follow that odd statement up with claiming legitimate companies will lose access to the government market. Here is where the arrogance that borders on pathological rears its pointy head. Closing a market by allowing only proprietary software denies access and competition – in short free trade – not the other way around. Their statement is transparent falsehood so contrived you’d have to be dense as a lead cupcake not to see through it.

Now in the second paragraph we begin to see the real incompetence that American jurisprudence has so sweetly bestowed upon the world. Rather than say, research their topic, they inform us of what it isn’t, setting up a straw man cum sock puppet. Fortunately this straw puppet is highly flammable – let’s set it alight; Free Software is built upon copyright.

Boy that was easy. Easy because it is so damn simple. Free Software relies on copyright law for its existence. You assign a copyright when you write software, often that copyright goes to a company if you are working for them, but sometimes it goes to individuals and organizations. They can license their software as they see fit if they possess the copyright. Many companies see a strategic advantage in using Free Software licenses like the GPL because it allows them to re-use high-quality software from other large companies, like Intel, IBM, Red Hat, Sun, Oracle, (stop me if this is boring you) Novell, Nokia, Microsoft, oh yes – even Microsoft has released Free Software. Presumably because it allows them to make money. But I’m just guessing here.

Clearly the IIPA lawyers either do not understand the law or they lie to protect their business. Free Software is directly built upon copyright law and to say “it fails to build respect for intellectual property rights” is to babble illogically. How can copyrighted works undermine copyright? How can you claim that following law does not respect the law? You can’t of course, these are logical impossibilities. You would think it would also be a bit illogical to rail against “open source” software and then go ahead and run your web site on that self-same open source, but this is precisely what the IIPA does; their ranting nonsense web site is hosted on a Solaris machine. Did I mention they were incompetent?

Free Software is a material product which lends itself to capital markets. It is an efficient use of capital and brings with it huge return on investment. Companies large and small use Free Software daily to help them grow and prosper, that is a fact that IIPA fails to mention and one cannot help but wonder why.

Comments Off on Incompetence and falsehood from American industry