Oracle Linux
Sunday, February 11th, 2007I ought to pay attention to my ranting more often (and maybe publish it more often). Back in March 2004 I offered the glib “It’s amazing Oracle don’t ship their own Linux“. Weird.
I ought to pay attention to my ranting more often (and maybe publish it more often). Back in March 2004 I offered the glib “It’s amazing Oracle don’t ship their own Linux“. Weird.
Yep, I’m a Struts committer now! Fear me.
So today saw a few JIRA issues closed and a couple of commits made to the Struts 1.x trunk. My aim will be to spend a bit of time each week looking at the 1.x issues and working on them. But first….
I need to get the Jakarta Standard Taglib released. We’re getting close to having all the issues tidied away. I’ve got a request in for the TCK, no idea what to expect there. It’s an interesting exercise in making the least amount of change to have the most amount of impact. I hacked in a caching SPI today, but I think it’s probably a bit too much change.
The Lang release failed its RC2. The legal files were missing from the Maven2 -sources/-javadoc jars. Bugger. Drawing board etc.
I’m definitely feeling the urge to do some coding rather than bugfixing/releasing. JIRA customizations are top of the list - just some simple things, like having a drop down selection list of your filters to conserve space, and being able to hardcode a list of filters. I’d also like to reverse engineer some rules from our triage process at work for use in a JBoss Rule engine somehow - seems like fun, time permitting.
A couple of days of doing little bugfixes to libraries. Along with work bits, yesterday I went through a list of Quartz issues that I’d identified as interesting and uploaded patches and tests to their JIRA, then today I’ve worked on a couple of outstanding Commons Lang 2.3 issues and built the 2.3-RC2 build for the community to weigh in on. Hopefully it’ll be good and I can get that released.
It’s 2 weeks ’til the next board meeting, so I’ve created the board agenda file. Now that I grok what goes on etc, I’m trying to help out where I can with the organizing. Dull boring stuff that I’m usually good at remembering to do.
I was remarking in an IM conversation that the last few blogs have been more enjoyable to write than previous attempts over the last N months because for some reason I kept having the feeling that I should be writing articles and not just random witterings about what I was doing. I’m a crap preacher, so thinking I should be trying to be bringing forth wise thoughts is a complete joke.
Articles are one way conversation - it’s dull. Even when people are adding comments, it’s usually to discuss the subject of the preaching and not to engage in discussion with the author. Blogging is about the articles being the elements of a conversation, so it seems fit that when I start doing things properly it feels much better.
A subsequent thought was that the good blog entries that become articles come out of conversations causing enlightenment and not out of sitting waiting for Eureka.
This is a habit of mine recently that I like:
When working on an issue, I attach any code I was working on to the issue (usually a unit test, I do avoid attaching the Poo.java test files that splatter my harddrive). This has the useful value of being available to anyone else working on the issue - and as they so often say - that someone else is usually yourself.
It also proves useful when your hard drive dies and you haven’t backed up for a month.
Incidentally - cool feature for a piece of backup software would be a bold piece of text on the display showing the number of days since the last backup. That way it is always in my face and I’ll think ‘ACK - backupbackupbackup’.
Today I built my first rpm. Pretty minor as it turns out - CruiseControl ships with a spec file and an Ant build.xml to do the work, so it was mostly about learning a bit about Yum and rpmbuild and .rpmmacros. Worked pretty well and then I sat and newbie’d with the installed CruiseControl, following the documentation to setup a build for Commons Lang.
CruiseControl (unsurprisingly) reminds me of Ant. You have to do a lot of work to get things setup, but you also have a lot of control. Continuum is the Maven (wow..shock!) in that it does more for you, but it’s harder to have control. Of course this is from a short period of use, but I like that CruiseControl is configuration file based, somehow that feels better than Continuum’s database approach. It is noisy though, so I have an urge to make a CruiseControl config file generator.
Thanks to Niall nudging me on the previous post, I took a look at the other validwhen issues. A couple are basically the same thing, decimals aren’t supported, so I commented on that and had a quick play to see if I could add it. No luck - but I do like doing lexer bits (it’s very like regexp) now that I have an actual example to use it on. I wonder what I did with my Lexx & Yacc book.
I added Cayenne and ActiveMQ to the front page of the Apache site. If you didn’t know - both projects have left the Incubator and become full on Apache projects in this ever growing enterprise borg that we’ve become. Also a link to ApacheCon as a Related Site and a search bar that uses Google rather than the 4-year old and utterly not maintained search.apache.org (swish-e perl module).
This evening I tried showing Nathan a Spiderman & His Amazing Friends cartoon after he liked The Incredibles the other day. It was pretty lame, but he seemed to find it watchable. Definitely not amazing though.
Time to make cookies… (cocoa oat cookies).
A much lighter day today - mostly because the morning was filled with going to the dentist to have my teeth worshipped (yes, I’m British and the US dentist does not run screaming… huzzah!). Apart from the usual work stuff of checking out the day’s new bugs (and reading work email and the various ASF lists), I spent a couple of hours figuring out Antlr generated code for the Struts validwhen validator (cf: STR-2997). I’m pretty chuffed that I got a fix+test patch in and that I improved my understanding of Antlr in the process.
Looks like Lang 2.3 may be held up a bit. There are some minor issues that’ll probably get improved a touch, but most importantly it looks like refactoring to one of the classes (Entities) undid the optimisation that a previous issue had asked for. So time for a bit of thinking to figure out how to change the refactoring so as not to hurt the speed.
I started listening to the Open Rights Group mailing list the other day. A Slashdot link (yeah I still read that occasionally) referred to them as the UK’s version of the EFF. Interesting stuff to keep track of.