2004-09-02

Standards error

You see, folks, this is why you always check your work. Bad things happen when you let your guard down and assume that something is just going to work.

Take yesterday's reinjector script that I wrote. It worked, but if I had had the luxury of time to really develop it, I'd have included a section that hashed the Message-ID field into SHA1 and wrote the value to disk. This way, duplicate messages could have been identified and rejected. As it was, people got x many messages, where x was the number of local recipients in the To and Cc fields. I realized this shortly after running the script, but by then it was too late. People got all the mail they'd lost, which was of primary importance. What mattered was that folks who were waiting for their mail got it. The only ones affected were those who were on large lists where the sender decided to CC half the building or such. I don't think there were many of these, but there were probably some. Always check your work.

I will give points to the woman who, in the middle of the whole mail-not-going-where-it-should problem, sent me an e-mail telling me that her mail wasn't working, and asking if I had any suggestions as to why. Points for being an idiot, maybe. I mean, c'mon. If the phones aren't working, are you going to try to dial 911?

No comments: