The OSBC Conference is coming up next week, May 22-23 in San Francisco.  This is the absolute best business conference on open source, bar none.  Matt Asay, built this conference up from nothing when the idea of open source and business was an oxymoron (like government efficiency or software quality).  Matt is still the conference chair and also a top exec at Alfresco, the leading open source content management system. 

Keynotes this year include Matthew Szulik from Red Hat (nice to see them participate at industry events, a rare phenomena in the past), Marc West from H&R Block, Eben Moglen on the GPL, Rob Curley from the Washington Post, …

I downloaded the MySQL source code over two weeks ago and finally it is built with code coverage statistics. I am going to go through the quick step by step process I used, although it wasn't quick at the time I tried to figure all of this out.

1) Download the source and unpack to a directory called "mysql".
2) cd mysql
3) BUILD/compile-pentium-gcov

This step has a tendency to give errors, especially those requiring aclocal and ncurses/termcap libraries. Make sure you have automake installed, but to have the proper aclocal executable automake must be of version 1.5 or higher. The ncurses libraries can be found, but they need the ncurses development libraries installed as well.

4) When the build is done run the tests with 'make test-force-full'. This will take some time, so walk away, watch tv, go the gym, make dinner, go to bed, etc.
5) lcov -c -d . -o lcov.output
6) mkdir …

$ wc -l toast.c
1219 toast.c

$ git log --pretty=oneline | wc -l
46

fixing bugs, making plans. Most of the time spent on it has been adding packet parsing support. Parsing data packets into structs, using a basic state machine. Rewriting that into a more "complete" state machine right now. Presently it can fully parse and understand the handshake process and COM_QUERY commands.

We'll see what happens.

MySQL's protocol is fairly nasty. Before moving onto buildnig the next part of the scaffolding I'm just getting the basic packets and commands dealt with. Luckily it allows me to set flags for what features work and what don't. So the proxy (will) disable compression, prepared statements, blah blah.

The ugliness I'm dealing with right now is with the client command packets. They're pretty simple:

4 byte header, omnipresent for all mysql network …

I’m now in Chicago, two days prior to PHPTek. I’ve never been to Chicago, so I wanted to take the conference as an excuse for a mini vacation. I’m staying at the Embassy Suites Rosemont, which is within a mile or so of the conference Hyatt. If you’re already here, or want to meet up with me during the conference.. here is my intended schedule..

Tuesday, May 15th 2007:

• Track #1 Zend Cert Crash Course (all day).

Wednesday, May 16th 2007:

• 10:15 - 11:15 : Writing maintainable PHP Code
• 11:30 - 12:30 : …

Oh fab. Microsoft claims that free software like Linux, which runs a big chunk of corporate America, violates 235 of its patents. It wants royalties from distributors and users.See http://money.cnn.com/magazines/fortune/fortune_archive/2007/05/28/100033867/index.htm

(Software) patents that (IMHO) shouldn't exist in the first place, but that's not the point I want to make here. Companies now need to compete on speed of innovation and "getting it out there", not on ability to hide, hoard and protect some perceived advantages. The above will just cost a lot of money, hinder innovation, annoy users, and in the end everybody loses. What a sad waste of energy.

The best performance improvement is the transition from the nonworking state to the working state --John Ousterhout

Aggressive use of pipeline parallelism is required to tame the ETL processing window for large data volumes, and a good ETL tool must provide easy access to this potent optimization technique.

As we start loading data into the staging tables, we invariably come across bad source data. Consider the sales payment file shown in the data …

I finished implementing the missing optional REGEXP function parameters today, all regular expression functions should now work as similar to their Oracle counterparts as possible with the following restrictions:

* The MySQL UDF API predates the extended character set support added with MySQL 4.1 and so UDF functions have no idea about charsets and collations at all. As a consequence the functions for now are always case sensitve by default and are always assuming their input to be Latin1 encoded
* Only the 'c' and 'i' pattern modifiers for case sensitive and insensitive matching are implemented yet, the 'm' and 'n' modifiers for multi line input are not yet supported
* I haven't tested back references on REGEXP_REPLACE() yet. They may or may not work, as i borrowed the actual implementation from the PHP source i'm not really sure about whether it takes care of this or not

Anyway, the REGEXP UDFs are hosted on …

Hi. I am doing a select like this in MySQL 5:

select * from foo where bar rlike '(.*),(.*)';

The specific example here is made up. Anyway, I'd like to be able to get to the matched text from bar, like I can with various languages regexp libraries. Is this functionality exposed at all in MySQL? I've looked at the docs and can't see any indication that it is, so this might just be wishful thinking.

Tags for this post: mysql regexp rlike select
Related posts: Managing MySQL …

A few people have asked me how fast MySQL Table Checksum is. As with so many other things, it depends. This article shows how long it takes to checksum real data on a production server I help manage, which might give you a rough idea of how long it'll take on your servers.

OSDC is a grass-roots conference providing Open Source developers with an opportunity to meet, share, learn, and of course show-off. OSDC focuses on Open Source developers building solutions directly for customers and other end users, anything goes as long as the code or the development platform is Open Source. Last year's conference attracted over 180 people, 60 talks, and 6 tutorials. Entry for delegates is kept easy by maintaining a low registration fee (approx $300), which always includes the conference dinner.

This year OSDC will be held in Brisbane from the 26th to the 29th of November, with an extra dedicated stream for presentations on Open Source business development, case studies, software process, and project management. The theme for this year's conference is "Success in Development & Business". If you are an Open Source maintainer, developer or user we would encourage you to submit a talk proposal on the open-source tools, …