[Qemacs-commit] qemacs qe.h charset.c tty.c

[Charlie Gordon]

CVSROOT:        /sources/qemacs
Module name:    qemacs
Changes by:     Charlie Gordon <chqrlie>        14/01/02 11:28:30

Modified files:
        .              : qe.h charset.c tty.c 

Log message:
        added generic utility function unicode_glyph_tty_width in charset.c
        
        * moved and improved code from tty_term_glyph_width

CVSWeb URLs:
http://cvs.savannah.gnu.org/viewcvs/qemacs/qe.h?cvsroot=qemacs&r1=1.104&r2=1.105
http://cvs.savannah.gnu.org/viewcvs/qemacs/charset.c?cvsroot=qemacs&r1=1.24&r2=1.25
http://cvs.savannah.gnu.org/viewcvs/qemacs/tty.c?cvsroot=qemacs&r1=1.51&r2=1.52

Patches:
Index: qe.h
===================================================================
RCS file: /sources/qemacs/qemacs/qe.h,v
retrieving revision 1.104
retrieving revision 1.105
diff -u -b -r1.104 -r1.105
--- qe.h        27 Dec 2013 01:54:12 -0000      1.104
+++ qe.h        2 Jan 2014 11:28:30 -0000       1.105
@@ -1,8 +1,8 @@
 /*
  * QEmacs, tiny but powerful multimode editor
  *
- * Copyright (c) 2000,2001 Fabrice Bellard.
- * Copyright (c) 2000-2013 Charlie Gordon.
+ * Copyright (c) 2000-2001 Fabrice Bellard.
+ * Copyright (c) 2000-2014 Charlie Gordon.
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -516,6 +516,7 @@
 int decode_8bit(CharsetDecodeState *s);
 u8 *encode_8bit(QECharset *charset, u8 *q, int c);
 
+int unicode_glyph_tty_width(unsigned int ucs);
 int unicode_to_charset(char *buf, unsigned int c, QECharset *charset);
 
 /* arabic.c */

Index: charset.c
===================================================================
RCS file: /sources/qemacs/qemacs/charset.c,v
retrieving revision 1.24
retrieving revision 1.25
diff -u -b -r1.24 -r1.25
--- charset.c   21 May 2008 06:42:32 -0000      1.24
+++ charset.c   2 Jan 2014 11:28:30 -0000       1.25
@@ -1,8 +1,8 @@
 /*
  * Basic Charset functions for QEmacs
  *
- * Copyright (c) 2000, 2001, 2002 Fabrice Bellard.
- * Copyright (c) 2002-2008 Charlie Gordon.
+ * Copyright (c) 2000-2002 Fabrice Bellard.
+ * Copyright (c) 2002-2014 Charlie Gordon.
  *
  * This library is free software; you can redistribute it and/or
  * modify it under the terms of the GNU Lesser General Public
@@ -23,7 +23,64 @@
 
 QECharset *first_charset;
 
-/* specific tables */
+/* Unicode utilities */
+
+/* Compute tty width of unicode characters.  This is a modified
+ * implementation of wcwidth() from Markus Kuhn. We do not handle non
+ * spacing and enclosing combining characters and control chars.
+ */
+
+static unsigned int const unicode_glyph_ranges[] = {
+    0x10FF, 1, 0x115f, 2,     /*  0: Hangul Jamo */
+    0x2328, 1, 0x232a, 2,     /*  2: wide Angle brackets */
+    0x2E7F, 1, 0x2efd, 2,     /*  4: CJK Radicals */
+    0x2EFF, 1, 0x303e, 2,     /*  6: Kangxi Radicals */
+    0x303F, 1, 0x4dbf, 2,     /*  8: CJK */
+    0x4DFF, 1, 0xa4cf, 2,     /* 10: CJK */
+    0xABFF, 1, 0xd7a3, 2,     /* 12: Hangul Syllables */
+    0xF8FF, 1, 0xfaff, 2,     /* 14: CJK Compatibility Ideographs */
+    0xFDFF, 1, 0xFE1F, 2,     /* 16: */
+    0xFE2F, 1, 0xfe6f, 2,     /* 18: CJK Compatibility Forms */
+    0xFEFF, 1, 0xff5f, 2,     /* 20: Fullwidth Forms */
+    0xFFDF, 1, 0xffe6, 2,     /* 22: */
+    0x1FFFF, 1, 0x3fffd, 2,   /* 24: CJK Compatibility */
+    UINT_MAX, 1,              /* 26: catchall */
+};
+
+static unsigned int const unicode_glyph_range_index[16] = {
+    2 * 0,   /* 0000-0FFF */
+    2 * 0,   /* 1000-1FFF */
+    2 * 2,   /* 2000-2FFF */
+    2 * 7,   /* 3000-3FFF */
+    2 * 9,   /* 4000-4FFF */
+    2 * 11,  /* 5000-5FFF */
+    2 * 11,  /* 6000-6FFF */
+    2 * 11,  /* 7000-7FFF */
+    2 * 11,  /* 8000-8FFF */
+    2 * 11,  /* 9000-9FFF */
+    2 * 11,  /* A000-AFFF */
+    2 * 13,  /* B000-BFFF */
+    2 * 13,  /* C000-CFFF */
+    2 * 13,  /* D000-DFFF */
+    2 * 14,  /* E000-EFFF */
+    2 * 14,  /* F000-FFFF */
+};
+
+int unicode_glyph_tty_width(unsigned int ucs)
+{
+    unsigned int const *ip;
+
+    /* Iterative lookup with fast initial jump, no boundary test needed */
+    ip = unicode_glyph_ranges + unicode_glyph_range_index[(ucs >> 12) & 0xF];
+
+    while (ucs > ip[0]) {
+        ip += 2;
+    }
+    return ip[1];
+}
+
+/* utf-8 specific tables */
+
 static unsigned short table_idem[256];
 static unsigned short table_utf8[256];
 static unsigned short table_none[256];
@@ -286,6 +343,12 @@
         if (c < 0x80 || c >= 0xc0)
             nb_chars++;
     }
+    /* CG: nb_chars is the number of character boundaries, trailing
+     * utf-8 sequence at start of buffer is ignored in count while
+     * incomplete utf-8 sequence at end of buffer is counted.  This may
+     * cause problems when counting characters with eb_get_pos with an
+     * offset falling indside an utf-8 sequence.
+     */
     return nb_chars;
 }
 

Index: tty.c
===================================================================
RCS file: /sources/qemacs/qemacs/tty.c,v
retrieving revision 1.51
retrieving revision 1.52
diff -u -b -r1.51 -r1.52
--- tty.c       17 Dec 2013 16:06:35 -0000      1.51
+++ tty.c       2 Jan 2014 11:28:30 -0000       1.52
@@ -1,8 +1,8 @@
 /*
  * TTY handling for QEmacs
  *
- * Copyright (c) 2000,2001 Fabrice Bellard.
- * Copyright (c) 2002-2013 Charlie Gordon.
+ * Copyright (c) 2000-2001 Fabrice Bellard.
+ * Copyright (c) 2002-2014 Charlie Gordon.
  *
  * This library is free software; you can redistribute it and/or
  * modify it under the terms of the GNU Lesser General Public
@@ -879,65 +879,13 @@
     qe_free(&font);
 }
 
-/*
- * Modified implementation of wcwidth() from Markus Kuhn. We do not
- * handle non spacing and enclosing combining characters and control
- * chars.
- */
-
-static unsigned int const tty_term_glyph_ranges[] = {
-    0x10FF, 1, 0x115f, 2,     /*  0: Hangul Jamo */
-    0x2328, 1, 0x232a, 2,     /*  1: wide Angle brackets */
-    0x2E7F, 1, 0x2efd, 2,     /*  2: CJK Radicals */
-    0x2EFF, 1, 0x303e, 2,     /*  3: Kangxi Radicals */
-    0x303F, 1, 0x4dbf, 2,     /*  4: CJK */
-    0x4DFF, 1, 0xa4cf, 2,     /*  5: CJK */
-    0xABFF, 1, 0xd7a3, 2,     /*  6: Hangul Syllables */
-    0xF8FF, 1, 0xfaff, 2,     /*  7: CJK Compatibility Ideographs */
-    0xFDFF, 1, 0xFE1F, 2,     /*  8: */
-    0xFE2F, 1, 0xfe6f, 2,     /*  9: CJK Compatibility Forms */
-    0xFEFF, 1, 0xff5f, 2,     /* 10: Fullwidth Forms */
-    0xFFDF, 1, 0xffe6, 2,     /* 11: */
-    0x1FFFF, 1, 0x3fffd, 2,   /* 12: CJK Compatibility */
-    UINT_MAX, 1,              /* 13: catchall */
-};
-
-static unsigned int const tty_term_glyph_index[16] = {
-    4 * 0,  /* 0000-0FFF */
-    4 * 0,  /* 1000-1FFF */
-    4 * 1,  /* 2000-2FFF */
-    4 * 3,  /* 3000-3FFF */
-    4 * 4,  /* 4000-4FFF */
-    4 * 5,  /* 5000-5FFF */
-    4 * 5,  /* 6000-6FFF */
-    4 * 5,  /* 7000-7FFF */
-    4 * 5,  /* 8000-8FFF */
-    4 * 5,  /* 9000-9FFF */
-    4 * 5,  /* A000-AFFF */
-    4 * 6,  /* B000-BFFF */
-    4 * 6,  /* C000-CFFF */
-    4 * 6,  /* D000-DFFF */
-    4 * 7,  /* E000-EFFF */
-    4 * 7,  /* F000-FFFF */
-};
-
-static int tty_term_glyph_width(__unused__ QEditScreen *s, unsigned int ucs)
+static inline int tty_term_glyph_width(__unused__ QEditScreen *s, unsigned int 
ucs)
 {
-    unsigned int const *ip;
-
     /* fast test for majority of non-wide scripts */
     if (ucs < 0x1100)
         return 1;
 
-    /* Iterative lookup with fast initial jump, no boundary test needed */
-    ip = tty_term_glyph_ranges +
-         tty_term_glyph_index[(ucs >> 12) & 0xF] - 2;
-
-    for (;;) {
-        ip += 2;
-        if (ucs <= ip[0])
-            return ip[1];
-    }
+    return unicode_glyph_tty_width(ucs);
 }
 
 static void tty_term_text_metrics(QEditScreen *s, __unused__ QEFont *font,