<<< Date Index >>>     <<< Thread Index >>>

Re: For 1.5.9: soft fill [patch-1.5.8.dgc.softfill.3]



First of all, I'd like to include this patch.

Ideally without the segfault, and with some meaningful documentation
for the muttrc(5) manual page.

Any takers?

-- 
Thomas Roessler · Personal soap box at <http://log.does-not-exist.org/>.





On 2005-04-02 14:13:39 +0200, Alain Bench wrote:
> From: Alain Bench <veronatif@xxxxxxx>
> To: Mutt dev ml <mutt-dev@xxxxxxxx>
> Date: Sat, 2 Apr 2005 14:13:39 +0200 (CEST)
> Subject: Re: For 1.5.9: soft fill [patch-1.5.8.dgc.softfill.3]
> Mail-Followup-To: Mutt dev ml <mutt-dev@xxxxxxxx>
> X-Spam-Level: 
> 
> Hi Dave!
> 
>  On Sunday, March 6, 2005 at 7:11:35 PM -0600, David Champion wrote:
> 
> > The softfill patch provides a "%*X" expando, similar to the existing
> > "%>X" and %|X". Like "%>X", it inserts padding to push a line of text
> > all the way to the right. Unlike "%>X", it gives fill precedence to
> > the right-hand side, not to the left.
> 
>     Clear, simple, and usefull. Thanks! I'm happy to have retriggered
> interest about it! About usefullness, see on mutt-users msgids
> <20050222162513.GA25614@xxxxxxxxx> and especially
> <20050302220609.GA14092@xxxxxxxxx> (note they talk about old v2 %=XY
> syntax).
> 
> 
> > Now the entire LHS is held in reserve. It's debatable whether this is
> > better, but it's definitely simpler.
> 
>     I agree, but let's pose the debate: Old v2 %=XY gives right-hand
> priority over *one* given Y left field, while new v3 %*X gives right
> priority over everything at left. Important left-hand infos may be
> hidden on narrow screen, which imposes carefull choice of fields and
> safe right width limitations. Compared v2 and v3 behaviours have each
> pros and cons. Ok: Simpler v3 wins.
> 
> 
> > the patch also modifies the basic expando modifier code
> 
>     BTW how can you truncate beginning of a fixed width field? Like a
> long /home/a/alain/long/path/filename.ext, displayed on 20 chars as
> "ng/path/filename.ext"?
> 
> 
>     Bug: With v3 and $index_format="%s%* [%F]" on a 50 columns term I
> get correct:
> 
> | Repeated crashes on x86-64 [Mads Martin Jørgensen]
> | |->                             [Thomas Glanzmann]
> | | `->                      [Mads Martin Jørgensen]
> | `->                                [Brendan Cully]
> |   `->                      [Mads Martin Jørgensen]
> 
>     But on 15 columns term, normally truncated beginning of name comes
> back overwriting end of line:
> 
> | rtin Jø[Mads Ma
> | omas Glanzma[Th
> | rtin Jø[Mads Ma
> | [Brendan Cully]
> | rtin Jø[Mads Ma
> 
> 
>     And while stress testing that effect, doing many terminal resizes, I
> also had at random in attachment menu black screens (staying black after
> ^L), and even one segfault:
> 
> | #0  menu_redraw (menu=0x80da218) at menu.c:841
> | 841       if (menu->dialog)
> | (gdb) print menu
> | $1 = (MUTTMENU *) 0x616c702f
> | (gdb) print menu->dialog
> | Cannot access memory at address 0x616c7057.
> 
> 
>     If this patch is commited, may I suggest default value of
> $attach_format gets changed, old "%.40d%> " sequence replaced by
> "%d%* ". This both removes the annoying fixed 40 chars limit, and
> ensures right-hand [type infos] are never pushed out of screen.
> 
> | set attach_format="%u%D%I %t%4n %T%d%* [%.7m/%.10M, %.6e%?C?, %C?, %s] "
> 
> 
> Bye!  Alain.
> -- 
> Give your computer's unused idle processor cycles to a scientific goal:
> The Folding@home project at <URL:http://folding.stanford.edu/>.
> 
>