[reportlab-users] Umlaut Bug with Adobe Reader and Arial

[Robin Becker]

On 23/05/2014 14:01, Volker Haas wrote:
> I just realized that the ttfAsciiReadable=0 trick does only work in the
> minimal example script I posted. If I use it in our rendering stack the
> fix does not work (I am confident that I am actually changing the
> configuration since the PDF files are significantly smaller after
> setting ttfAsciiReadable=0).

well there goes one theory :(
>
> Therefore my question is: if you think this is a bug in reportlab, do
> you plan on fixing it?
>
> Best Regards,
> Volker
>
........

I think this must be a bug related to the actual tables we are including in the 
font (or perhaps some miscalculation). I will have to try and see what 
libreoffice does in its output to make their subset work reasonably. If I get 
our code to write out the binary subset ttf file than it is not acceptable to 
windows font display tool, but fontTools & fontforge think it's ok. Using 
fontTools I converted ariabld to ttx changed all the Arial internal names to 
NRial and recreated a ttf. That font is acceptable to windows and even though 
there cannot be any kind of name clash with an adobe builtin font it also shows 
the umlaut issue.

When I see this error then Adobe Preflight is telling me there's a glyph width 
mismatch for one of the umlauted characters SMALL LETTER U WITH DIARESIS.
-- 
Robin Becker