Re: Compiling the CVS Tree

[Lars Clausen <lrclause cs uiuc edu>]

On Mon, 30 Dec 2002, Hans Breuer wrote:
> At 04:57 29.12.02 -0600, Lars Clausen wrote:
> Nope. The FREETYPE code is using this factor _twice_ which IMO is 
> questionable. The one definition as constant like the comment says
> is ok, but just using it a second time as in
> lib/font.c:dia_font_build_layout 
>
> #ifdef HAVE_FREETYPE
>     height *= global_size_one;
> #elif defined G_OS_WIN32
>     height *= 0.7;
> #endif
>
> appears to be wrong. It would be nice to know if changing only
> the second appearence to 1.0 would solve the 'huge fonts problem'.

With a recent enough Pango, the Unix side works with height *= 1.0, and
global_size_one = 20 (the global zoom factor).  That even makes sense.  The
global_size_one (renamed global_zoom_factor to reduce confusion) is now
only used relative to an actual zoom factor.

Using 1.0 actually does make the fonts a little too large, lines can
overlap each other.  0.7 does look better.  

[...]
>   It maybe be possible to get on the correct factor with the attached
> test program. Or after looking at it's output on my notebook maybe 
> not :-(
>
> D:\devel\my-gtk\simples>fontres
> Screensize is 800x572 pixels  211x158 mm
> Resolution 96x91 dpi
> Pixels per cm : 37.9 x 36.2
> Dia's magic font factor should be 0.527500(w) or 0.552448(h)
>
> The best working factor for win32 is between 0.7 .. 0.71 which
> could be :
>       - 72 dpi / 96 dpi
>       - 20 / 28.35
>       - something completely different

It's worrisome to have a strange factor like that.

-Lars

-- 
Lars Clausen (http://shasta.cs.uiuc.edu/~lrclause)| Hårdgrim of Numenor
"I do not agree with a word that you say, but I   |----------------------------
will defend to the death your right to say it."   | Where are we going, and
    --Evelyn Beatrice Hall paraphrasing Voltaire  | what's with the handbasket?