[XeTeX] Segmentation fault with xdvipdfmx

[Jonathan Kew]

On 10 Jul 2006, at 12:02 pm, Akira Kakuto wrote:

>> Thanks for the report; confirmed as a bug, and should be fixed in
>> rev. 56, just checked in. Triggered by the font having a format 3
>> 'post' table (no glyph names).
>>
>> (Apparently Windows is more tolerant of dereferencing a NULL pointer,
>> at least sometimes...)
>
> My previous test was on windows 2000.
> On windows XP, access violation occurred that was fixed in rev. 56.
> I don't know whether fonts are different or not.

On re-reading your previous message, I realized that the font must  
have been different, as it did include glyph names in the 'post'  
table. So that's why you didn't see an access violation. I guess XP  
ships a version of the font with a format 3 'post' table, while W2K  
had format 2.

While xdvipdfmx obviously needs to handle this without crashing, I  
think it's unfortunate that MS has decided to ship such a font. The  
OpenType spec has this to say about 'post' version 3:

<quote src="http://www.microsoft.com/typography/otspec/post.htm">
This version specifies that no PostScript name information is  
provided for the glyphs in this font file. The printing behavior of  
this version on PostScript printers is unspecified, except that it  
should not result in a fatal or unrecoverable error. Some drivers may  
print nothing, other drivers may attempt to print using a default  
naming scheme.
</quote>

I suppose the assumption is that the MS driver will handle it OK, and  
it doesn't matter if the font happens to fail on Postscript printers  
using other driver software. So much for interoperability....

JK