August, 11, 2005 archives

the los angeles times reports that many of the plaques at the old caltrans building were actually cement, although they appeared metallic.

i find it fairly mind-boggling that there is no picture to accompany the article. the building is literally across the street from the los angeles times building.

i am surprised that newspapers don’t use more photographs on their web sites, and use them more effectively.

here’s another example: this article about how to store book collections doesn’t have a single picture. unbelievably lame.

what do we get instead? ten little pictures to tell the story of the watts riots.

quatermass and the pit is vintage british scifi, with all the good and bad that implies. in the end, i was disappointed by how choppy the story is, with just a few too many leaps of logic on the part of the scientists.

it’s also another one of those films i probably appreciated less than i might have because i fell asleep while i was watching it the first time.

equilibrium is a movie that is seems good, but seems to be lacking some vital ingredient. it’s like a loaf of bread baked without salt. it does have a couple of pretty snazzy action scenes, and the underlying idea is sort of neat, if well-worn.

here’s how i spent ninety minutes of work today: having been asked to look at a timing related bug, i came to the conclusion that writing a good test case really required i implement an an old feature request for a sleep() function. ninety minutes later, the code is written, approved by brian aker, and pushed to the main 5.0 tree.

and it’s the first function i’ve ever added to mysql. oh yeah, and i ate dinner while i was waiting for brian to review my patch.

i’ve also got the test case written for the other bug. but there’s some question as to what the standards-defined behavior really is — from my reading of the sql:2003 spec, it’s implementation-defined. apparently our standards gurus (now on vacation) left other impressions.

we’ll probably work it out so you can get either behavior depending on what function you call to get the current timestamp.

update: thirty minutes later, paul dubois has documented the new sleep() function. i love it when a plan comes together.