Vasile Gaburici wrote:
Before I get anymore strawman arguments, here's the definitive experiment:
$ ls -U1 /usr/share/fonts/vista/ | grep erda Verdanai.ttf Verdana.ttf Verdanaz.ttf Verdanab.ttf
$ ls -U1 /xp/wtf/ Verdana.ttf Verdanab.ttf Verdanai.ttf Verdanaz.ttf
did you test the latest beta (this afternoon version?)
XeTeX solves this problem by using *only* the ID 16 and ID 17 OpenType values to create a unique name for system fonts. I actually filed a bug about lack of support for the common ID 1/2 names, but Johnathan WONTFIXED it because he really wants the font names to be unambiguous, and using ID 1 and ID 2 values in addition to ID 16/17 doesn't guarantee that. Your situation is worse, because you want to support all sort of legacy fonts in that single hashtable.
part of the problem was that the mtx-font lookup (shown by list) differed from the one used in practice (in tex); this is now harmonized (and seems to be ok on taco's and my machine)
Anyway, I've wasted enough of my time on this matter. If you sill don't grok it, forget about it.
you don't want to know how much time we spent on fonts so far -)
BTW, the Lua 5.1 "reference" manual is a joke. How do you printf a boolean? What's the syntax for their definitely non-standard regexps.
they are not regexps but just expressions; being a lightweight language lua does not have regexp (as the manual states: it would add more code than lua itself; however, there is lpeg (also in luatex) which is not only more poweeful than regexp but also is efficient wrt the implementation 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 -----------------------------------------------------------------