Hans Hagen wrote:

Barry Schwartz wrote:

...

I do not know whether this is a luatex bug or a ConTeXt bug. Mark IV seems to be incorrectly scaling and spacing PostScript fonts that have unconventional em-unit-size settings, such as my Sorts Mill Goudy fonts available here: http://www.theleagueofmoveabletype.com/

can you make a test file? also, such matters can best be posted to the context list as more people can test then

looks like a backend issue to me Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------

Hans Hagen skribis:

this bug is not in luatex but in acrobat (in order to catch it luatex will have to do some additional manipulations)

btw, this was already in the tracker:

http://tracker.luatex.org/view.php?id=200

(for this purpose it can be handy to have multiple pdf viewers on your system as xpdf and okular an d... show the font ok)

As do kpdf, evince, and epdfview. This is an odd bug, because for instance my own software (http://code.google.com/p/kompostilo/source/browse/trunk/python/kt/pdf.py) embeds the font in similar format, but Reader displays my fonts correctly (as do okular et al.). This probably is why I didn't think to check with other viewers. Perhaps it is because my software uses the FontForge Python extension to convert the font to CID CFF, and FF is doing something magical during the conversion.

I wrote:

FYI when I try to use FontForge to extract the embedded fonts from luatex-created documents, including the ConTeXt docs, FF doesn't recognize the documents as PDF.

I think it may be accidental that okular et al. display the font correctly. I ran a ConTeXt-generated pdf through pdftk to make it recognizable to fontforge, and when I extracted the font it had 2048001 units per em instead of 2048. (Also, the font extracted from a pdf made by my software was extracted by fontforge according to its name without the subset prefix, but the font extracted from ConTeXt output had the subset prefix. I suppose that's probably a harmless difference due to the vagueness of the PDF spec.)

Barry Schwartz wrote:

Barry Schwartz skribis:

...

FYI when I try to use FontForge to extract the embedded fonts from luatex-created documents, including the ConTeXt docs, FF doesn't recognize the documents as PDF.

Probably fontforge can't read the xref from a stream, which is how luatex is writing the xref. I'll report it as a fontforge bug.

then try \pdfobjcompreslevel=0 or \nopdfcompression ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------