Re: no iso 10646 fonts are used

Stefan> Hi, with the sjt-xft branch and a default font that is a Stefan> truetype font available in 10646 encoding, I don't get the Stefan> correct character if I try evaluate (unicode-to-char 0945) Stefan> This works fine in 21.5-b19 without the xft unfortunately But not via ISO 10646---I'm fairly sure that it's being presented via a CMap as an ISO 8859-7 font, else XEmacs couldn't use it. Stefan> Under xft I get the message: (font/notice) Unable to Stefan> instantiate font for charset greek-iso8859-7, face default Fix the entry for Greek in "struct charset_reporter charset_table" at line 1143 of objects-x.c. (Of course this table should be Lisp-visible, but it hasn't evolved to that point yet.) If I read ISO 639-2 correctly (I can't seem to access ISO 639-1), that ISO 639 tag should be "el", not "gr". Check that your ISO 10646 font supports "el" (it should be in the long string of x|y|z in the lang property, these are part of the full font descriptor spewed by the debug code to stderr when the font is initialized---if your XEmacs doesn't do this, update to current sjt-xft, which does by default). Thanks for the report, and my apologies for the inconvenience. A patch will be in CVS later. Do you mind if I cite you in the ChangeLog? Like this: 2005-03-10 Stephen J. Turnbull <stephen(a)xemacs.org&gt; * objects-x.c (charset_table): Fix ISO 639 code for Greek. Reported by Stefan Holst <holst(a)mathematik.uni-mainz.de&gt;. Call for volunteers: I would *really* appreciate a patch to make that table Lisp-visible if somebody has the skills and time---that would really make this branch much more useful, I think. I don't, right now. Suggestions: Basic approach: make the table a hash-table with Mule charsets as keys, values would be Lisp vectors. Stuffing the table needs to go into complex_vars_of_objects_x or maybe reinit_objects_x. Advanced: (1) allow aliasing by letting the first element of the value be a charset to be rehashed, otherwise the language description string. (2) Cache the FcCharSet objects for the collections (eg, Latin-1 is a collection that covers many Western European languages). Blue sky: Provide design documentation and (if appropriate) implementation for whether and how to unify FcCharSets with Mule char-tables. -- Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Ask not how you can "do" free software business; ask what your business can "do for" free software.