What a week... trying to get a fix from my H/W guy so I can freakin' TEST the changes that have come down the pike, so I can get a decent update to the rest of the team so THEY can test, so we can be done with this damn project.
Of course, we have three revisions of the device and the newer images don't work right on the older hardware (e.g. Rev C firmware on Rev B device; not my fault, it's hard-wired that way; if my H/W guy had used reverse logic on the sleep line it would have worked, but noooo, probably not enough room for an inverter. Or the '!' sign in the VHDL. Who knows.)
So yesterday I send out an e-mail that includes the note that the modified image I was sending FIXES the Rev-B sleep issue, but MAY BREAK four specific modules due to creeping incompatibilities. This gives us a lot more devices we can use for testing, with the effect of just losing 4 of 36 modules... no big deal, big improvement overall.
I get a ping from one of the guys downstairs saying that the RevB sleep hack doesn't work, they can't use modules! Okay, so my H/W guy messed up (don't we all?) and the hack doesn't work. Bring one up and I'll verify the problem.
Umm. Guys. This works. I see the module, I see data...
Oh! It turns out it only fails on THIS module. And this second one as well.
Both of which were listed as possibly broken in my e-mail.
These are smart guys. Did they not read my e-mail at all? It wasn't even a long e-mail! I kept it short!
::headdesk::
Thank god it's Friday; I _may_ get through this day without killing someone.