Re: ruler-mode is broken

[Lute Kamstra]

address@hidden (Kim F. Storm) writes:

>>>> I seems that set-fringe-style insists on setting the sum of the
>>>> widths of the left and right fringe to a multiple of
>>>> `(frame-char-width)', but does so in a (for me) counterintuitive
>>>> way.
>
> The reasoning is that emacs should never select a width that is less
> than what the user requested.
>
> Currently, this may not make a lot of sense, but I plan to make the
> fringe bitmaps customizable (and add the ability to define
> additional bitmaps in the fringes).  In that case, I could easily
> envision a mode which would use bitmaps of width 4 in the left
> fringe, and a desire to narrow that fringe to exactly (or at least)
> 4 pixels.
>
> Now suppose, you specify the fringes as (4 . 8) with a char width of
> 11, and emacs chooses to give you (3 . 8).  So the current logic was
> chosen to avoid having to modify the behaviour later on.

Ah, now I'm beginning to understand.  So far, I did not see a reason
to round up as opposed to round to the nearest.  The "inconsistent"
behavior of rounding when 0 is used:

  input     set to    distance
  (0 4)     (0 11)    7
  (1 3)     (4 7)     5
  (2 2)     (5 6)     5
  (3 1)     (6 5)     5
  (4 0)     (11 0)    7

as opposed to:

  (0 4)     (3 8)     5
  (1 3)     (4 7)     5
  (2 2)     (5 6)     5
  (3 1)     (6 5)     5
  (4 0)     (8 3)     5

makes sense too: when a user says 0, (s)he definitely doesn't want a
fringe.

Thanks for explaining this,

  Lute.