Re: [ft-devel] Type 1C font that is not rendered correctly

[Werner LEMBERG]

> > Sigh.  I really wonder whether I shall break the specification
> > just to support this particular font...
> 
> Well, i'd say it all boils down to how much the breakage implies in
> terms of work to fix it and if it can break correct fonts and how
> difficult it'll be to maintain in the future.

This is `fixed' now.  Here the ChangeLog entry:

        * src/cff/cffgload.c (cff_decoder_parse_charstrings): No
        longer assume that the first argument on the stack is the
        bottom-most element.  Two reasons:

          o According to people from Adobe it is missing in the Type 2
            specification that pushing of additional, superfluous
            arguments on the stack is prohibited.

          o Acroread in general handles fonts differently, namely by
            popping the number of arguments needed for a particular
            operand (as a PS interpreter would do).  In case of buggy
            fonts this causes a different interpretation which of the
            elements on the stack are superfluous and which not.

        Since there are CFF subfonts (embedded in PDFs) which rely on
        Acroread's behaviour, FreeType now does the same.

This is a rather big change; my test fonts all work fine, and I ask
you to run your test suites too.


    Werner