[AC sjt-xft] Get a sane estimate of average character width

APPROVE COMMIT sjt-xft Index: src/ChangeLog =================================================================== RCS file: /pack/xemacscvs/XEmacs/xemacs/src/ChangeLog,v retrieving revision 1.758.2.8 diff -u -U0 -r1.758.2.8 ChangeLog --- src/ChangeLog 11 Feb 2005 15:50:40 -0000 1.758.2.8 +++ src/ChangeLog 20 Feb 2005 14:41:59 -0000 ＠＠ -0,0 +1,6 ＠＠ +2005-02-20 Stephen J. Turnbull <stephen(a)xemacs.org&gt; + + * objects-x.c (x_initialize_font_instance): Estimate average cell + width based on dimensions of "representative" string. Suppress + some excessive debug output. Get rid of spurious tests. + Index: src/objects-x.c =================================================================== RCS file: /pack/xemacscvs/XEmacs/xemacs/src/objects-x.c,v retrieving revision 1.26.2.6 diff -u -r1.26.2.6 objects-x.c --- src/objects-x.c 1 Feb 2005 15:46:56 -0000 1.26.2.6 +++ src/objects-x.c 20 Feb 2005 14:42:07 -0000 ＠＠ -261,22 +261,29 ＠＠ /* #### This shit is just plain wrong unless we have a character cell font. It really hoses us on large repertoire Unicode fonts with "double-width" characters. */ - /* #### We just init'ed the memory; these are all zero to start. */ - if (f->ascent < rf->ascent) - f->ascent = rf->ascent; - if (f->descent < rf->descent) - f->descent = rf->descent; - if (f->width < rf->max_advance_width) - f->width = rf->max_advance_width; - if (f->height < rf->height) - f->height = rf->height; + f->ascent = rf->ascent; + f->descent = rf->descent; + { + /* This is an approximation that AFAIK only gets used to compute + cell size for estimating window dimensions. The test_string8 + is an ASCII string whose characters should approximate the + distribution of widths expected in real text. */ + static const char test_string8[] = "Mmneei"; + static const int len = sizeof (test_string8) - 1; + XGlyphInfo glyphinfo; + + XftTextExtents8 (dpy, rf, test_string8, len, &glyphinfo); + /* #### maybe should be glyphinfo.xOff - glyphinfo.x? */ + f->width = glyphinfo.width/len + 1; + } + f->height = rf->height; /* #### Is this right? See comment on proportional_p below. */ f->proportional_p = 1; /* we can't recognize monospaced fonts! */ - if (debug_xft > 1) - stderr_out ("initialized ascent %d descent %d width %d height %d\n", - f->ascent, f->descent, f->width, f->height); if (debug_xft > 0) + stderr_out ("initialized metrics ascent %d descent %d width %d height %d\n", + f->ascent, f->descent, f->width, f->height); + if (debug_xft > 2) /* we also output on initialization of any font below */ stderr_out ("initialized Xft font %s\n", extname); } else ＠＠ -355,10 +362,13 ＠＠ { int n = 3, d = 5; /* #### check for weirdness */ - if (n*f->height < d*f->width) - stderr_out ("width of %s is %d, larger than %d/%d its height %d\n", - extname, f->width, n, d, f->height); - stderr_out ("initialized Xft font %s\n", extname); + if (n * f->height < d * f->width) + stderr_out ("font %s: width:height is %d:%d, larger than %d:%d\n", + extname, f->width, f->height, n, d); + if (f->height <= 0 || f->width <= 0) + stderr_out ("bogus dimensions of font %s: width = %d, height = %d\n", + extname, f->width, f->height); + stderr_out ("initialized font %s\n", extname); } #else #undef rf -- 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.