Bill,
I've got some good news and bad news for you. The bad news is that you can't predict the future. Sorry. The good news (or bad news, depending on what perspective you have) is that I screwed up.
When I read your post (#1473) which was a reply to my PM that the "alias" problem had also occurred on Thursday morning, my gut reaction was, "This can't be. We didn't release the code until Thursday evening."
I even verified with the engineers that we hadn't released code until Thursday evening, so the error "could not possibly" have occurred before then. Therefore, I was certain that you must have been mistaken, and I replied as I did.
When I got you second reply, my gut reaction was, "What the hell?" Then, I read your post to me from Thursday morning (for the first time), and my next reaction was, "Oh S h - -!"
I did some further investigation and learned that we did, in fact, do a test rollout of the new code in the wee hours of Thursday morning. That was not a full rollout. We only rolled it on a couple servers, not the entire system, and we rolled it back very quickly. Frankly, with everything else we've been dealing with the last couple days, it never occurred to anyone that the alias problem might also have existed during that brief test rollout period.
And, yes, if we had read your post earlier, or in more detail, we might have put 2 and 2 together...but we didn't. I feel bad enough that we've had these problems, but I feel far worse that I've now actually added to the confusion with misinformation. I hope this post helps to clear it up.
I sincerely, and personally apologize to all.
Bryan
PS: For those of you who care, I prefer to eat my crow with a side of humble pie. |
