BoA & DST
|
| Author | Content |
|---|---|
| dcparris Mar 15, 2007 5:06 PM EDT |
I've just received word - I'm talking about excellent authority here - that BoA was not thrilled with MS and Sun in the midst of the DST change. Apparently Sun found a bug in their patch at the last minute. Try patching 40k servers and over 100k workstations over the weekend. I don't think he noticed I was smiling when I told him my Debian boxes were all just fine. And to think I'm just a dumb security guard. ;-) |
| jimf Mar 15, 2007 5:20 PM EDT |
> I don't think he noticed I was smiling It's to your credit that you could stifle your outright laughter :D |
| dcparris Mar 15, 2007 5:34 PM EDT |
Yeah, that was the hard part. :-D |
| phsolide Mar 15, 2007 7:29 PM EDT |
I know that the Solaris system admins at my employer were less than thrilled with that extra DST patch, or rather a patch of a non-working patch. Still, lets put blame where its due: squarely on Microsoft. The Sun patches were for Java Virtual Machines. Why did the JVMs need patching, when stuff like /etc/localtime exists? Because the JVMs have to run on Windows. Windows has traditionally hosed up leap years and DST. In fact, one of the (ignored!) objections to the MSOOXML spec is that it institutionalizes a famous Windows bug that has 1900 as a leap year. It was not, despite its divisible-by-4 numerical nature. Therefore, JVMs keep DST info in some kind of compiled-in lookup table. Disgusting and redundant, but necessary for the "write once, run anywhere" thing that makes Java so darn irresistable to big, stupid corporations. |
You cannot post until you login.