freetype-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Devel] infinite loop details


From: Ian Brown
Subject: RE: [Devel] infinite loop details
Date: Fri, 14 Dec 2001 08:30:18 +0100

Scott, see my post of July 5 2001 about PCL fonts.

In addition to these points, I came across the following issue when generating a TTF from a windows PCL:

The glyph descriptions in the glyf table need to be consequtive. This is because the size of a glyph is found by differencing the offset of the glyph and the offset of the following glyph as found in the loca table. Unfortunately, windows downloads the glyphs as they occur in the document, so you need to do some dynamic re-ordering.

I haven't looked at your TTF, but I remember I had the same freetype crash that you did, but it was because I wasn't putting everything together properly. I suspect you have the same problem, and the Printer Driver is correct.

Given that you develop a somewhat competing product from my company, I won't actually post any code :-) ... but you should be able to figure out what you need to do from the above hint.

Ian Brown
PrintSoft Systems GmbH

-----Original Message-----
From: Scott Long [mailto:address@hidden]
Sent: Friday, December 14, 2001 2:05 AM
To: address@hidden
Subject: [Devel] infinite loop details

I just realized that my previous post didn't include enough information. It
seems that the root of the problem is this:

Either the PCL printer driver, or the Windows GDI, is producing "funky" glyph
representations. The glyph data included in the PCL print stream is seriously
wrong for glyph index 3 (which I assume is supposed to be an empty glyph,
since it represents a space). The glyph information does not even contain a
complete glyph header.

The problem does not seem to be in our translation process. This is likely to
be a bug in the Windows PCL printer driver, or something else in Windows. Our
translation process just takes the glyph data from the stream and memcpy()'s
it into the TTF file image. I really don't see any way that this process
could go wrong. The PCL data itself is bad.

So the root of the problem seems to be a bug in a printer driver :-) But I
still think that FreeType should handle this problem gracefully.

Thanks for helping out,
Scott

_______________________________________________
Devel mailing list
address@hidden
http://www.freetype.org/mailman/listinfo/devel


reply via email to

[Prev in Thread] Current Thread [Next in Thread]