Bit rot explained

It wasn’t bit rot that was breaking my Batik tests.  It was maven.  I emerged to the latest version of maven, and dutifully upgraded the various plugins, and all my Batik tests were failing.  I even tried upping the version of junit, etc., to no avail.  But when I finally figured out how to test in Eclipse (just press the run button, who knew that meant test as well?), the tests passed. I’d file a bug with Maven, but I haven’t a clue how to reproduce the bug in a simple way.

Driving a Chrysler

Seems like the Chrysler part of DCX is finally revving its engine.  Funny how nothing changed except rumors of buyouts and spin-offs, and the stock pops.  The company itself, its products and workers, are all the same.  Why does the market think that getting it out from under Daimler will help the company, especially when Daimler has invested so much time and effort trying to merge the two companies.  Perhaps that is it, the market is saying that the merger isn’t happening, and more investing to make it happen is a bad idea, while *cutting and running* (stupid republicans) is a good idea.

inexplicable bit rot

I am documenting old code, and in the process upgrading small things—new version of Maven2, etc.  While admittedly things have changed due to the “advice” that Eclipse has been giving me about my Java code, suddenly I find old tests that passed are now failing.  Batik is the culprit, but I just can’t make any headway.

Is “Ruby red grapefruit in its own juice” good for you?

So I forgot to take my antibiotics this morning.  Which probably means nothing, but I am sick of strep throat.

Finally getting work done again in the evenings (and at work).  Hooking up Dojo widgets to edit entries, pushing through some false starts.  Turns out that the Inline Edit widget doesn’t really like to be instantiated programmatically (read the code,  there is a little comment in there that says just that).  But that’s fine, I decided it is simpler to just make a regular form, and hook up timers to everything (onChange events, or onChangeValue if it is a dojo widget) to a timeout function, as I did with my ATMSDataFeed plotter.  When something changes, the timer starts.  When something else changes, the timer is reset, so rapid changes will get passed all at once, and slow dilly dally changes or just one change get passed off as singletons to the backend.

I still think the tree widget is broken (right click to get menu is cool, but for some reason it doesn’t un-highlight the previously selected node??), but looking at the latest nightly tests it seems like drag and drop isn’t as broken as it used to be.  I don’t really have a need for drag and drop at the moment, but that is a good canary for the quality of the tree code overall.

I really need to read up on POE too.  I don’t understand it, but I think it might help out quite a bit with setting up servers for our SOA stupidness.