mail headers (was: Re: Z80 CLK DRV)

Yes. CFWS. Commas are not CFWS. Yes. This is where most of the comma-containing References: headers probably come from - software written to a very old definition of the header (or possibly, even, written by someone who didn't bother checking the spec). Even 822 specifies References: without the commas (you can use commas in an 822 References:, but only by putting quotes around them, turning them into quoted-strings, or parens, turning them into comments). 733, intermediate between 724 and 822, has the commas. Other sources of commas in References: are software that copies existing References: blindly, adding message-IDs to one end or the other, thus perpetuating mistakes but at least not making them worse. Indeed, the example quoted above must have come from such a thing, since it is lacking the comma between the third and fourth message-IDs. There's also In-Reply-To:, which I regularly see containing free-form text such as (to quote the first example I find in my mail) In-Reply-To: <4FEFB24E.1080206 at telegraphics.com.au> from Toby Thain at "Jun 30, 12 10:13:34 pm" 822 permits that particular case (since the <special>s are all quoted) but 2822 does not. I really need to tweak my own MUA so that it doesn't perpetuate the problem in existing headers (it currently is an instance of the "copy it textually and add" genre; I fix them when I notice, but I don't always notice). /~\ The ASCII Mouse \ / Ribbon Campaign X Against HTML mouse at rodents-montreal.org / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B