RE: [Freetype] Question on MacOS X

[Leonard Rosenthol]

At 07:18 PM 11/30/2001 -0500, Changyuan Hu wrote:
> >          True, but the standard FT_New_Face() on the Mac goes through some
> > special processing to provide transparent support for Mac OS fonts.  I
> > would recommend that you use that call instead, so that FT can deal with
> > all the complexities for you.
>
> Function FT_New_Face() requires font file path name as input.

        Correct.


> That is almost useless for a serious application in Windows or Mac, in 
> which you would allow user to choose font from family name, style and 
> size, but not from font file selection.

        Agreed.  Which is why there is a new routine (not checked in yet!) 
that lets you specify the name and style, and get back the path name and 
FOND index to use for calling FT_New_Fac().


>  It is the same case in Windows, where you probably use Windows standard 
> ChooseFont() to open standard font selection dialogue and upon return, 
> you get the handle to the selected font. Because the Windows ChooseFont() 
> function uses special algorithm to map user requirement to the font 
> Windows sees as the best match, using a file dialog to let user choose a 
> font file directly is not such a nice solution. (That is why I still did 
> not use FreeType in my Windows application.)

        I haven't looked, but I suspect that Windows has a call similar to 
FT_Face_From_FOND that lets you get an FT_Face from an HFONT.


> So, I have to use Mac native way to open a font, i.e. passing a FOND 
> resource handle to FT_New_Face_From_FOND(). Macintosh dev technical note 
> TN2024 has a detailed explaination on how to open the FOND resource hand 
> from both the traditional resource-fork font and the new data-fork font 
> (.dfont).

        Correct.  BUT it also has a problem where it doesn't take into 
account the face offset (as reported to me by someone else on this list, 
and patches WERE added to FT2 to deal with it).


> By the way, the patch code in FT2 version ftmac.c to allow parse_font find 
> a face which is not the first one in FOND resource seems not complete.

        Oh?


> It still does not support style (normal, italic, bold and bold-italic). 
> Some FOND contains sfnt resource of all the four possible styles (such as 
> Times). So, we should loop into the AsscEntry array to find the best fit. 
> I did this patch for the FT1.2.2 code locally and of course I added the 
> style argument in the FT_New_Face_From_FOND() function. Following is my 
> local modification.

        I'll test it out (though I am sure it's fine) and check it in!


LDR