On 16. Oct 2006, at 14:20, Erwan David wrote:
It seems that OE is in this case... (why can it not send in 8bit encoding when answering to a f=f message, if it is not broken ?)
So if you send a f=f letter to an OE user, and he replies, the reply is encoded with QP, but had you not sent it as f=f, it would be 8 bit? That is a rather peculiar pattern, as whether OE can send the reply as 7 bit or 8 bit would depend on which SMTP it goes through.
That said, sending a 7 bit QP reply is not in anyway broken, it is in fact what most mailers do, since you can’t blindly assume that SMTP’s will support 8 bit.
And MUAs not f=f aware will show 1 line per paragraph, which is ugly and may be inconvenient (wrapping in the middle of a word or not showing the whole line).
No they will not. Take my second reply in this thread, if you view it in Mail it appears first as a quoted paragraph and then a non-quoted paragraph. Both flow to the width of the window.
This is how it looks in a UA not supporting f=f: http:// lists.macromates.com/pipermail/textmate/2006-October/013851.html
The reason is that behind the scenes the UA does send it as hard wrapped, but puts a space after each of the lines which the UA wrapped, signaling to the receiver (that does support f=f) that the line break should be removed.
In addition to having the paragraphs reflow to the width of the *receivers* window size, hard wrapped mails have the disadvantage that each time a line is quoted, it grows by at least two bytes, so it will quickly be too long for the current hard wrap setting, meaning the last word is moved to the next line, either producing unquoted lines among the quoted lines, or quoted text with a very ragged right border. f=f solves this problem as well.