[XeTeX] Missing ligatures (0.997)

Jonathan Kew jonathan_kew at sil.org
Tue Jul 31 19:51:43 CEST 2007

On 31 Jul 2007, at 6:19 pm, Stephen Moye wrote:

>  On Tuesday, July 31, 2007, at 12:26PM, "Stephen Moye"  
> <stephenmoye at mac.com> wrote:
>> Thank you, Bruno, for confirming my discovery that LY1 was indeed  
>> the answer.
>> Sorry I did not point this out: I am on a Mac and using TeXShop
>> Now that I have LB working, I am getting one more error. This is  
>> from the .xdv file:
>> =====
>> pdf_font>> Simple font "lsd" (enc_id=3) found at id=7.
>>> (Image:./emstm-2.pdf[./emstm-2.pdf][PDF])][7(Image:./emstm-1.pdf 
>>> [./emstm-1.pdf][PDF])<hlcrv at 7.97pt(TFM:hlcrv[/usr/local/texlive/ 
>>> 2007/../texmf-local/fonts/tfm/bh/lucida/hlcrv.tfm])
>> fontmap: hlcrv -> lbme
>> pdf_font>> Simple font "lbme" enc_id=<builtin,-1> opened at  
>> font_id=<hlcrv,10>.
>>> (Image:./emsmono-3.pdf[./emsmono-3.pdf][PDF])][8(Image:./ams- 
>>> short-course-08.pdf[./ams-short-course-08.pdf][PDF])(Image:./ 
>>> ast-310.pdf[./ast-310.pdf][PDF])](lbr[LucidaBright] 
>>> [TeXnANSIEncoding][Type1]
>> ** WARNING ** Obsolete four arguments of "endchar" will be used  
>> for Type 1 "seac" operator.
>> ** ERROR ** Invalid SID
>> Output file removed.
>> =====
>> What does this mean? Everything else that I have tried has worked  
>> perfectly...
> Once again, I seem, after a fashion, to have answered my own  
> question. It appears that any accented character (\`e, \"u or è, ü)  
> brings about this error using the xdvipdfmx causes this error. Does  
> the LY1 encoding have anything to do with this?

I suspect that any 8-bit encoding that provides precomposed accented  
letters would cause it with this font. You could try using OT1 (which  
I don't think has such letters) and see if the problem goes away.  
(Note that this will disrupt hyphenation of such words.)

What I think is going on is that the font uses an obsolete Type 1  
construction to build the accented letters, and xdvipdfmx is not  
handling this. It's generating a warning, which is reasonable, but  
probably proceeds to do something silly with the font data, and then  
it falls over completely.


More information about the XeTeX mailing list