On 3 jun 2003, Simon Josefsson outgrape: > Jesper Harder writes: > > > Simon Josefsson writes: > > > > > Jesper Harder writes: > > > > > > > Simon Josefsson writes: > > > > > > > > It is valid, so it should work. > > > > > > > > Nope, multiple To fields are not allowed, cf. the table in RFC > > > > 2822, Section 3.6. > > > > > > See RFC 2822 4.5 for the obs-to field definition, which is > > > identical to the to field definition, but can occur several > > > times. > > > > Ah, yes you're right -- it's obsolete, but clients should still be > > able to interpret it. > > Perhaps Gnus could warn if the user types multiple To: lines though? Not a too bad idea. In fact I don't really know why user can enter 2 or more 'To' fields. That doesn't sound natural to me. > OTOH, minibuffer queries are pure evil, so the warning must be > non-obtrusive if it is present at all. Perhaps coloring the second > To red and having a balloon help with information would work; or some > tty-friendlier variation of the same idea. Seems good. > Perhaps other message.el warnings could be recast like this too. The > non-legible character set question would be really nice to get rid of > IMHO. Just highlight the illegible text in the message buffer, and > when the user sends the message, really do send it. That's an idea we can work on and maybe generalize to all the same situation. I vote for. Thanx for all your enlightenments and your quick and clear answers guys. Maybe some day I will read the mail dedicated RFCs. BTW, which RFCs are the most important ? I mean which one should be read and known for somebody wnating to hack a little more into the Gnus core ? zeDek -- http://www.gnusfr.org -- French Gnus user site Anti-war disclaimer: "Bombing for peace is like fucking for virginity"