Re: [widec] end-of-line strange behaviour

[TAKAHASHI Tamotsu]

On Sat, 26 Mar 2005, Thomas Dickey wrote:
> On Sat, 26 Mar 2005, Thomas Dickey wrote:
> 
> >On Sun, 27 Mar 2005, TAKAHASHI Tamotsu wrote:
> >
> >>I have no idea why this problem occurs. But I have
> >>never seen this problem with S-Lang, so I think it's
> >>a bug in ncurses.
> >
> >I'm not sure yet.  Using your example, I see it is for a terminal 70 
> >columns wide (which makes the URL wrap at the "right" place.  I don't
> >see any misplaced characters except that the wrapping of the dashed
> >line looks suspicious - I'll investigate that and see why/what it is
> >doing.

Ah, I forgot to say that my muttrc had "unset smart_wrap".


> The wrapping of the dashed line was incorrect: ncurses wrapped one cell 
> too early, and mutt tried to guess what ncurses did rather than ask it
> nicely.  So mutt became confused about what was on the screen.  I'm
> adding this fix to ncurses (but mutt should ask, rather than guess):

Thank you, your patch fixed my problem!
And yes, applications which use full-screen width - like mutt -
should fully cooperate with (n)curses. Although I don't know how
to implement it, I'd like to investigate mutt's code some day.

Thanks,
-- 
tamo