Re: Found problem with Subrs parsing...

[Just van Rossum]

At 2:59 PM -0400 24-07-2000, Tom Kacvinsky wrote:
>And it is:
>
>Some fonts (including an older version of ITC Garamond Light I had
>laying around) do not define the subrs in increasing order; but the
>t1z driver expects them to be in increasing order.
>
>The "offending" code is in z1load.c (function parse_subrs).  I have
>not yet thought of an elegant patch for this.  David?

Removing these lines fixes the problem for this font:

      /* make sure we get subr index and loop count in sync --  */
      /* in some cases, the dictionary count is simply the last */
      /* subr index + 1, with possible holes in the table       */
      if ( index > n )
        n = index;

But... this is not the correct patch: it'll screw up when there are holes.
So unless the current spec specifically disallows this funny subr ordering
(and it doesn't seem to do that), there needs to be a more solid way to
detect the end of the subrs array. My knowledge of the t2z parser is still
too limited to immediately see how to do that cleanly...

Just