mhbuildsbr.c: Flip logic, moving goto to then-block; no need for else.

[nmh] / test / getcwidth.c

diff --git a/test/getcwidth.c b/test/getcwidth.c
index f24172fefa97ea0f52930ffb668e62ab5b1ead59..b692a9f6459a7e6b45ac70333eadf62faa4f7490 100644 (file)
--- a/test/getcwidth.c
+++ b/test/getcwidth.c
@@ -1,7 +1,7 @@
 /*
 /*

- * getcwidth - Get the OS's idea of the width of a combining character
+ * getcwidth - Get the OS's idea of the width of Unicode codepoints

  *
  *

- * This code is Copyright (c) 2012, by the authors of nmh.  See the
+ * This code is Copyright (c) 2013, by the authors of nmh.  See the

  * COPYRIGHT file in the root directory of the nmh distribution for
  * complete copyright information.
  */
  * COPYRIGHT file in the root directory of the nmh distribution for
  * complete copyright information.
  */


@@ -18,63 +18,227 @@
 #ifdef MULTIBYTE_SUPPORT
 #include <locale.h>
 #include <wchar.h>
 #ifdef MULTIBYTE_SUPPORT
 #include <locale.h>
 #include <wchar.h>

+#include <wctype.h>

 #endif
 
 #endif
 

+#ifdef MULTIBYTE_SUPPORT
+static void usage(char *);
+static void dumpwidth(void);
+static void dumpctype(void);
+static void getwidth(const char *);
+#endif /* MULTIBYTE_SUPPORT */
+

 int
 main(int argc, char *argv[])
 {
 int
 main(int argc, char *argv[])
 {

+#ifndef MULTIBYTE_SUPPORT
+       (void) argc;
+       (void) argv;
+       fprintf(stderr, "Nmh was not configured with multibyte support\n");
+       exit(1);
+#else /* MULTIBYTE_SUPPORT */

        wchar_t c;
        wchar_t c;

-       int charlen;
-       char *p;
+       int i;
+
+       if (! setlocale(LC_ALL, "")) {
+               fprintf(stderr, "setlocale failed, check your LC_ALL, "
+                   "LC_CTYPE, and LANG environment variables\n");
+       }
+
+       if (argc < 2)
+               usage(argv[0]);
+
+       if (strcmp(argv[1], "--dump") == 0) {
+               if (argc == 2) {
+                       dumpwidth();
+                       exit(0);
+               } else {
+                       fprintf(stderr, "--dump cannot be combined with "
+                               "other arguments\n");
+                       exit(1);
+               }
+       }
+
+       if (strcmp(argv[1], "--ctype") == 0) {
+               if (argc != 2) {
+                       fprintf(stderr, "--ctype cannot be combined with other arguments\n");
+                       exit(1);
+               }
+               dumpctype();
+               exit(0);
+       }

 
        /*
 
        /*

-        * This is the UTF-8 for "n" + U+0308 (Combining Diaeresis)
+        * Process each argument.  If it begins with "U+", then try to
+        * convert it to a Unicode codepoint.  Otherwise, take each
+        * string and get the total width

         */
 
         */
 

-       unsigned char string[] = "n\xcc\x88";
+       for (i = 1; i < argc; i++) {
+               if (strncmp(argv[i], "U+", 2) == 0) {
+                       /*
+                        * We're making a big assumption here that
+                        * wchar_t represents a Unicode codepoint.
+                        * That technically isn't valid unless the
+                        * C compiler defines __STDC_ISO_10646__, but
+                        * we're going to assume now that it works.
+                        */
+                       errno = 0;
+                       c = strtoul(argv[i] + 2, NULL, 16);
+                       if (errno) {
+                               fprintf(stderr, "Codepoint %s invalid\n",
+                                       argv[i]);
+                               continue;
+                       }
+                       printf("%d\n", wcwidth(c));
+               } else {
+                       getwidth(argv[i]);
+               }
+       }

 
 

-       setlocale(LC_ALL, "");
+       exit(0);
+}

 
 

-       if (argc != 1) {
-               fprintf(stderr, "Usage: %s\n", argv[0]);
-               fprintf(stderr, "Returns the column width of a UTF-8 "
-                       "multibyte character\n");
-               exit(1);
-       }
+static void
+usage(char *argv0)
+{
+       fprintf(stderr, "Usage: %s [--dump]\n", argv0);
+       fprintf(stderr, "       %s [--ctype]\n", argv0);
+       fprintf(stderr, "       %s U+XXXX [...]\n", argv0);
+       fprintf(stderr, "       %s utf-8-sequence [...]\n", argv0);
+       fprintf(stderr, "Returns the column width of a Unicode codepoint "
+               "or UTF-8 character sequence\n");
+       fprintf(stderr, "\t--dump\tDump complete width table\n");
+       fprintf(stderr, "\t--ctype\tPrint wctype(3) table.\n");

 
 

-#ifndef MULTIBYTE_SUPPORT
-       fprintf(stderr, "Nmh was not configured with multibyte support\n");

        exit(1);
        exit(1);

-#else
+}
+
+static void
+getwidth(const char *string)
+{
+       wchar_t c;
+       int charlen, charleft = strlen(string);
+       int length = 0;
+

        /*
        /*

-        * It's not clear to me that we can just call mbtowc() with a
-        * combining character; just to be safe, feed it in a base
-        * character first.
+        * In theory we should be able to use wcswidth(), but since we're
+        * testing out how the format libraries behave we'll do it a character
+        * at a time.

         */
 
         */
 

-       mbtowc(NULL, NULL, 0);
+       if (mbtowc(NULL, NULL, 0)) {}

 
 

-       charlen = mbtowc(&c, (char *) string, strlen((char *) string));
+       while (charleft > 0) {
+               int clen;

 
 

-       if (charlen != 1) {
-               fprintf(stderr, "We expected a beginning character length "
-                       "of 1, got %d instead\n", charlen);
-               exit(1);
-       }
+               charlen = mbtowc(&c, string, charleft);
+
+               if (charlen == 0)
+                       break;

 
 

-       p = (char *) (string + charlen);
+               if (charlen < 0) {
+                       fprintf(stderr, "Unable to convert string \"%s\"\n",
+                               string);
+                       return;
+               }

 
 

-       charlen = mbtowc(&c, p, strlen(p));
+               if ((clen = wcwidth(c)) < 0) {
+                       fprintf(stderr, "U+%04lX non-printable\n",
+                               (unsigned long int) c);
+                       return;
+               }

 
 

-       if (charlen != 2) {
-               fprintf(stderr, "We expected a multibyte character length "
-                       "of 2, got %d instead\n", charlen);
-               fprintf(stderr, "Are you using a UTF-8 locale?\n");
-               exit(1);
+               length += clen;
+               string += charlen;
+               charleft -= charlen;

        }
 
        }
 

-       printf("%d\n", wcwidth(c));
+       printf("%d\n", length);
+}

 
 

-       exit(0);
+typedef struct {
+       wchar_t min, max;
+} unicode_range;
+
+static unicode_range range[] = {
+       /* https://en.wikipedia.org/wiki/Unicode#Code_point_planes_and_blocks */
+       {  L'\x0000',    L'\xff' },
+#if WCHAR_MAX >= 0xffff
+       {  L'\x0100',  L'\xffff' },
+#if WCHAR_MAX >= 0xfffff
+       { L'\x10000', L'\x14fff' },
+       { L'\x16000', L'\x18fff' },
+       { L'\x1b000', L'\x1bfff' },
+       { L'\x1d000', L'\x1ffff' },
+       { L'\x20000', L'\x2ffff' },
+       { L'\xe0000', L'\xe0fff' },
+#endif
+#endif
+       { L'\0', L'\0' }, /* Terminates list. */
+};
+
+static void
+dumpwidth(void)
+{
+       unicode_range *r;
+       int first;
+       wchar_t wc, start;
+       int width, lastwidth;
+
+       for (r = range; r->max; r++) {
+               first = 1;
+               for (wc = r->min; wc <= r->max; wc++) {
+                       width = wcwidth(wc);
+                       if (first) {
+                               start = wc;
+                               lastwidth = width;
+                               first = 0;
+                               continue;
+                       }
+                       if (width != lastwidth) {
+                               printf("%04lX - %04lX = %d\n", (unsigned long)start,
+                                          (unsigned long int)wc - 1, lastwidth);
+                               start = wc;
+                               lastwidth = width;
+                       }
+                       if (wc == r->max) {
+                               printf("%04lX - %04lX = %d\n", (unsigned long)start,
+                                          (unsigned long int)wc, lastwidth);
+                /* wchar_t can be a 16-bit unsigned short. */
+                break;
+            }
+               }
+       }
+}
+
+static void
+dumpctype(void)
+{
+       unicode_range *r;
+       wchar_t wc;
+
+       for (r = range; r->max; r++) {
+               for (wc = r->min; wc <= r->max; wc++) {
+                       printf("%6x  %2d  %c%c%c%c%c%c%c%c%c%c%c%c\n",
+                               wc, wcwidth(wc),
+                               iswcntrl(wc) ? 'c' : '-',
+                               iswprint(wc) ? 'p' : '-',
+                               iswgraph(wc) ? 'g' : '-',
+                               iswalpha(wc) ? 'a' : '-',
+                               iswupper(wc) ? 'u' : '-',
+                               iswlower(wc) ? 'l' : '-',
+                               iswdigit(wc) ? 'd' : '-',
+                               iswxdigit(wc) ? 'x' : '-',
+                               iswalnum(wc) ? 'N' : '-',
+                               iswpunct(wc) ? '@' : '-',
+                               iswspace(wc) ? 's' : '-',
+                               iswblank(wc) ? 'b' : '-');
+
+            if (wc == r->max)
+                /* wchar_t can be a 16-bit unsigned short. */
+                break;
+               }
+       }

 #endif /* MULTIBYTE_SUPPORT */
 }
 #endif /* MULTIBYTE_SUPPORT */
 }