Re: flags, beams and stem length in forced directions - output improvement

[Janek Warchoł]

W dniu 22 stycznia 2011 19:06 użytkownik Carl Sorensen
<address@hidden> napisał:
> On 1/22/11 10:59 AM, "Janek Warchoł" <address@hidden>
> wrote:
> >>> Now i'm almost done. However i still don't know how to attach
> >>> different versions of flags to the stems.
>
> Look in lily/stem.cc.  Method Stem::calc_flag.
> lines 643-662 are about where it's currently done.

I see! That's exactly what i want.
Now i ran into another problem. I wanted to insert this

  if (flag_style == "") // i'm preparing additional flags only for
ordinary font style
    {
      if (length < 6.83)
        {
          if (length > 6.5) font_char += "1";
                    // flags.u31 would be a 8th flag that is only a
bit shorter than regular 8th flag (flags.u3)
          else if (length > 6.16) font_char += "2";
                    // flags.u32 would be a 8th flag that is a bit
shorter than flags.u31, and so on
          else if (length > 5.83) font_char += "3";
          else if (length > 5.5) font_char += "4";
          else if (length > 5.16) font_char += "5";
          else font_char += "6";
        }
    }

into line 639 of stem.cc (just before Font_metric *fm =
Font_interface::get_default_font (me);)
but unfortunately length variable is local to Stem::calc_length. I've
tried defining it in the top-level, but failed.
May i ask for help?
Sorry for asking such basic questions, i'm still new to all this.

thanks,
Janek

2011/1/22 Werner LEMBERG <address@hidden>:
>> And there is one more thing - is it a bug in font sources? When i change
>> total_depth# = 3.5 staff_space# - blot_diameter# / 2;
>> to
>> total_depth# = 3 staff_space# - blot_diameter# / 2;
>> (in definition of u4 flag, line 163 of feta-flags.mf), i get the
>> output in attachment. I've tried other values, and:
>> 2.95, 3, 3.1 are bugged
>> 2,9, 3.05, 3.15 are not bugged
>> What do you think about it?
>
> Do you see this `spike' in the DVI output also?

umm... DVI output? I've searched the manuals and didn't found how to
achieve it :/

> Alternatively, if you
> call mf2pt1 as
>
>  FONTFORGE=foo mf2pt1 --rounding=0.0001 <mf-font>   ,
>
> do you see the irregularity if you open the font with fontforge?

Yes, i attach a screenshot. The font file is here:
http://www.mediafire.com/?hadr9nt5w5iqulb

> This
> would indicate a problem with the MF code.  Otherwise (and I believe
> this is the cause), it's a bug in fontforge's intersection resolution
> algorithm.  You might try the recent CVS version of fontforge which
> has a lot of fixed issues.  Don't forget to pass `--enable-double' to
> the `configure' script of fontforge.

Umm.. i'm afraid i don't understand :(
But since fontforge shows the spike in pfb file, i suppose it's irrelevant?

cheers,
Janek

flagspike.png
Description: PNG image