Re: adaptive-fill bug?

[Stefan Monnier]

>     2. it rejects fill-prefixes guessed by adaptive-fill-mode if they are
>     longer than a line of text (longer than the difference between
>     fill-column and left-margin).

>     The second part is the crucial one and should be 100% uncontroversial.

> I agree.

>     The first part should be harmless as well: such a check used to be
>     done on both the first and the second line of a paragraph.  I removed
>     the check on the first line several years ago because it introduced
>     a bug, and moved the paragraph-start check further down, just before
>     returning.  This late-check also makes the second check unnecessary.

> I don't follow the logic here; I don't see why it is unnecessary.

AFAIK, the check is there to make sure we do not use a fill-prefix which
matches paragraph-start.  The late-check ensures this without caring where
the guessed prefix comes from, so it covers adaptive-fill-regexp matches on
both the first and second line (and third line for people like me ;-).


        Stefan