RE: [ft-devel] serious bug introduced 2005-10-29

[Turner, David]

OK,

I just commited a fix to the problem. There was a bug in "ft_trigon_prenorm".
the optimization introduced recently

as for why I couldn't find it on Windows or Jam, apparently because I was used
modified versions of ftoption.h. Dohhhhh.

Also included, a fix to the truetype interpreter FT_UNUSED_EXEC unwanted memcpy

Hope this helps,

- David


> -----Message d'origine-----
> De : address@hidden
> [mailto:address@hidden
> la part de Turner, David
> Envoyé : mardi 15 novembre 2005 15:11
> À : Ismail Donmez; address@hidden
> Objet : RE: [ft-devel] serious bug introduced 2005-10-29
>
>
> ok, i've done some investigation:
>
> - i can't reproduce the problem on Windows, with either "make
> setup devel" or "jam"
>   I'm using gcc 3.2.3
>
> - on Linux, I can't reproduce the problem when compiling with
> Jam at the moment.
>   However, I *can* with "make setup devel". Don't know why
> just yet. This is
>   puzzling...
>
> - disabling the latest optimizations in ftcalc.c doesn't
> change anything to the
>   output (still buggy on Linux with "make setup devel")
>
> - running valgrind reveals some bad things in the bytecode
> interpreter.
>   Everytime FT_UNUSED_EXEC is evaluated, a full structure copy is
>   performed  ( *exc = *exc ). this makes Valgrind scream because
>   the contents of a memcpy are overlapping... Ah ah ah, never noticed
>   this one, which must be a performance killer, by the way...
>
> - no other errors are reported by Valgrind (except some memory leaks
>   in the graph library, I should fix these some days, but
> it's so low priority)
>
> I'm continuing my tests to see what's happening here.
>
> Stay tuned,
>
> - David
>
>
> > -----Message d'origine-----
> > De : address@hidden
> > [mailto:address@hidden
> > la part de Ismail Donmez
> > Envoyé : mardi 15 novembre 2005 13:08
> > À : address@hidden
> > Objet : Re: [ft-devel] serious bug introduced 2005-10-29
> >
> >
> > On Tuesday 15 November 2005 12:36, Werner LEMBERG wrote:
> > > David,
> > >
> > >
> > > a comparison between a FreeType checkout with -D 2005-10-28 and -D
> > > 2005-10-29 shows that the latter has introduced a serious bug
> > > w.r.t. TTF handling.  Attached two images from the font
> > Dustismo.ttf.
> > >
> > > Perhaps your arithmetic performance improvents are buggy...
> > >
> > > This might also be the cause for the strange Vera
> rendering issues.
> >
> > I am seeing many bad artifacts with Tahoma.ttf from Windows
> > installation. I
> > was thinking optimization would break Qt but I guess it breaks more.
> >
> > Regards,
> > ismail
> >
> >
> > _______________________________________________
> > Freetype-devel mailing list
> > address@hidden
> > http://lists.nongnu.org/mailman/listinfo/freetype-devel
> >
> **************************************************************
> *********************
> Information contained in this email message is confidential
> and may be privileged, and is intended only for use of the
> individual or entity named above. If the reader of this
> message is not the intended recipient, or the employee or
> agent responsible to deliver it to the intended recipient,
> you are hereby notified that any dissemination, distribution
> or copying of this communication is strictly prohibited. If
> you have received this communication in error, please 
> immediately notify the address@hidden and destroy the
> original message.
> **************************************************************
> *********************
>
>
>
> _______________________________________________
> Freetype-devel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/freetype-devel
>
***********************************************************************************
Information contained in this email message is confidential and may be 
privileged, and is intended only for use of the individual or entity named 
above. If the reader of this message is not the intended recipient, or the 
employee or agent responsible to deliver it to the intended recipient, you are 
hereby notified that any dissemination, distribution or copying of this 
communication is strictly prohibited. If you have received this communication 
in error, please immediately notify the address@hidden and destroy the original 
message.
***********************************************************************************