Re: Umlaut Bug with Adobe Reader and Arial

I tried some options; turning off ascii readability made the problem go away.

Try

set rl_ttfAsciiReadable=0
python broken_umlauts.py

to see what happens. I did some quick checking with an older (2012) version of 
reportlab and can see the problem there as well so this isn't related to all the 
bytes/str changes.


The difference between set rl_ttfAsciiReadable=0 / 1 is that for debugging 
purposes we want to map the standard ascii characters to the first subset in the 
correct position. That means the default setting makes most ttf strings readable 
in the (and the whole file a bit longer). When we have ascii readablity = 0 then 
we try to subset the used characters plus some messing about to preserve space 
positioning which is required for some reason. At any rate a quick fix for 
modern reportlab might be to create a reportlab_settings.py somewhere on the 
python path and put

ttfAsciiReadable=0

in it.

This isn't an answer to the underlying problem, but it may ameliorate the effects.
-- 
Robin Becker