Of TimeZones and other silliness (There really should be a standard...)

My eyes were just way too sore to stare at a computer this afternoon, so I napped for a few hours, then headed out to chat with Simon for a few hours more. I drank green tea the whole time, so I got home somewhat wired for sound.

I found and fixed a bug I'd introduced during the latency fixes. No big deal there, code was assuming a dictionary was coming in but it's now a generator producing two-tuples. Next I started working on the matplotlib-generated graphics, mostly small bits and pieces, fixing font sizes, that kind of thing.

Then I started into what should have been a 2-minute fix; the time information was off by 5 hours in the graph ticks. Oh, thinks I, must be a problem with the timezone name linked from /etc/localtime

Turns out that the machine doesn't use a symlink in /etc/localtime , that though common, that setup isn't a required element for unix machines. After reading quite a few pages it seems there really isn't a single consistent mechanism available for finding the Olsen database name for the current timezone on the computer :( . Oh well; it's now a configuration option that can override the attempt to automatically calculate it.

Anyway, my zest for coding this evening is beginning to wane (caffeine wearing off, eyes and hands beginning to complain), so I think maybe I'll call it a night.


Comments are closed.


Pingbacks are closed.