Author: philippn

Status Update

The good news first: The documentation task is finally being adressed. All the infrastructure work is done since the past weekend so I can focus on hacking away the contents. The highlight of the chosen implementation is definitely that the documentation HTML is generated on-the-fly during build from a DocBook 5 XML file. Plus, the infrastructure for context-sensitive-help is also up and running, so things are shaping up nicely.

Now, the bad news is that I’m again behind schedule. Continue reading “Status Update”

Ohloh, I want my half hour back!

After having started with the documentation stuff, I discovered that the docbook stylesheets used for generating the documentation on-the-fly during build process were inflating the project stats on Ohloh.

While this might not be the biggest problem in the world, I still found it annoying as hell. So, now I have implemented a work-around that also unzips the docbook stuff on-the-fly so that it doesn’t need to sit in subversion.

That was probably the most superfluous half-hour that I have spent for a while…

1.0 and Beyond…

Some time has passed since Milestone 3 and it appears to have been quite a success. We still don’t have any reported bugs and during my quite frequent use of it at work I didn’t experience any oddities either. Good stuff 🙂

That’s not ment to say that where isn’t alot of potential for improvement though, which brings us back on topic… Continue reading “1.0 and Beyond…”

Milestone 3 to arrive early

I have decided to change my earlier plans a bit, and release M3 this weekend.

The reasons for this are manifold. First I have fixed alot of rather nasty parsing bugs that I encountered during testing, which alone would justify a release. Also I have finished a really huge reengineering, that should definitely improve overall application robustness. And finally, the WebSphere support is rocking now 🙂

Though, I will have to fix another bug and add a few more JUnit tests, before M3 will be rolling out. Stay tuned.

1.0.0M2 out now!

As promised, Milestone 2 is being released today.

It adds a bit of polish here and there when compared to M1. In particular:

  • When re-opening a log resource, the last active filter is re-applied
  • You can save and load arbitrary filters in the Filter Settings dialog. I hope I have made this feature intuitive enough, atleast until documentation is done…
  • You can drop all log entries from a log resource using the new Truncate action. This is useful i.e. for rolling log files, where you don’t want to keep everything that has happened weeks ago. In the past you had to delete the log resource entirely, and then re-import. Now you can simply use Truncate occasionally 🙂

Apart from that, a bunch of bugfixes also made it in this release. Plus, I could remove the dependency on org.eclipse.core.resources which makes the download a bit smaller.

As usual you can get it from SourceForge.net.

Note: Log resources imported with M1 will not be picked up by M2. You will need to re-import them!

Release Schedule

Here is the tentative release schedule up to 1.0.0 final.

  • M2: April 19th – Usability improvements
  • M3: May 19th – WebSphere dialect, overall documentation, <insert your feature request here>
  • RC: Early June – P2 update function
  • Final: mid-June

After release of 1.0.0, work will continue immediately on 1.1 which will be based on Eclipse Platform 3.6. If you have any super cool feature requests that you would like to see in 1.0, feel free to post them in the comments or open a ticket in our SourceForge.net tracker.

About the Build process — and the last few check-ins

You might wonder where the builds for LogSaw actually come from. Since February 2010, LogSaw is built by Hudson in corporation with Buckminster.

It was a rather tedious act to get it to work, but since then it has been working rather nicely and it definitely saves time. Plus you can hook some great stuff into Hudson, for example Sonar. Sonar integrates a bunch of static code analysis utilities into a simple to use web-interface, which can help to improve overall code quality.

Since the release of M1, I have setup a Sonar instance and started working on the issues identified by it. So this is basically what the last few check-ins were all about…

M1a out now

This application requires Java 6, a tiny fact that I had forgotten somewhere on the way to M1 :-/

So the only change in M1a is that the app will show a proper error message if launched with Java 5 or lower (M1 would just refuse to start with no specific error given).

Some advice for people who already tried M1:

  • If you are happily running M1, you can safely skip this update!
  • If LogSaw M1 refused to start for you, make sure that you have Java 6 installed and then extract it again; otherwise it will continue not to work eventhough Java 6 is installed