RE: [ft-devel] Implementation error of the FreeType2 function FT_Get_MM_Var()

[Chi Nguyen]

James,

Thank-you very much for your quick reply. 

The reported problem was observed using freetype v2.1.1 and v2.3.7.
It only showed up on our 32-bit x86 systems running OpenSuSE11 and Fedora
core 9.

Is it possible to verify that on your end? Shall we file a bug against these
two versions of freetype?

A cursory check on our test-systems did not exhibit that
        (master->axis[i].def)<<16 == master->axis[i].minimum
when an invalid value is reported by the function FT_Get_MM_Var().


Thanks and best regards,


Nguyen Trung Chi



-----Original Message-----
From: James Cloos [mailto:address@hidden 
Sent: Friday, November 28, 2008 2:13 AM
To: Chi Nguyen
Cc: address@hidden; lsb-dev; address@hidden
Subject: Re: [ft-devel] Implementation error of the FreeType2 function
FT_Get_MM_Var()

I tried that (on a 32 bit, x86 linux box) using ft from cvs as of a week
or so ago.

I couldn't trigger the bug using the versions of ZX and ZY included in
the rar.

I changed the printf() from %ld to %lx; hex helped show one issue:

When it does report a false value, it is always the case that:

     (master->axis[i].def)<<16 == master->axis[i].minimum

So thre seems to be a missing or incorrect shift someplace and every
time it occurs the default value is the minimum value.  (Assuming that
/DesignVector is what master->axis[i].def returns.)

-JimC
-- 
James Cloos <address@hidden>         OpenPGP: 1024D/ED7DAEA6