Re: [EXTERNAL] Re: Bug: tt_check_single_notdef function skip spaces

[Werner LEMBERG]

>>  Too bad that Microsoft's PDF driver doesn't strip off tables for
>>embedded bitmaps.  Maybe it makes even sense to write a bug report
>>to MS: There shouldn't be any bitmap strikes in fonts part of a PDF.
> 
> I can imagine a PDF where someone is using these bitmas in the right
> size, but the space itself can be custom scaled .... And that won't
> be a bug...

I don't think so: Bitmap strikes make *only* sense for screen display.
As far as I know, there is normally no functionality to scale embedded
bitmaps.[*] Bitmap fonts in PDFs are always Type 3 PS fonts.

> Hmm, with my corporate email, I had trouble signing up here: 
> https://gitlab.freedesktop.org/groups/freetype/-/issues

What exactly do you mean with 'signing up'?  You need to register
first before you can submit a Merge Request.  If everything fails, an
e-mail with the patch as an attachment will do, too.

> sbits-always-scalable and FT_PARAM_TAG_SBITS_ALWAYS_SCALABLE.  That
> resolve this problem, but in my opinion problem is in
> tt_check_single_notdef (don't see glyph at index 3).  If I
> understand correctly this function checking only glyph offsets
> (empty glyph don't have different begin and end offset)?

Correct.

> I expected some corrections there.  Acrobat and FontForge can see
> that this is a (empty) glyph on index 3.

The function works as expected, as far as I can see: If there are
bitmap strikes but only space glyphs besides '.notdef', do not set
'FT_FACE_FLAG_SCALABLE' and handle the font as bitmap-only.  As
mentioned previously, this is (almost) the exact definition of how
bitmap-only glyphs for Windows are constructed by FontForge
(https://fontforge.org/docs/techref/bitmaponlysfnt.html).


    Werner


[*] Well, there exists the `EBSC` OpenType table, but this is for
    screen display, too, and certainly not for PDF output.