Mercurial > hg > octave-nkf > gnulib-hg
view tests/uniwbrk/test-uc-wordbreaks.c @ 17870:04ed097a2048
uniwbrk/u32-wordbreaks-tests: add conformance test
* modules/uniwbrk/u32-wordbreaks-tests (Files): Add
tests/uniwbrk/test-uc-wordbreaks.c,
tests/uniwbrk/test-uc-wordbreaks.sh, and
tests/uniwbrk/WordBreakTest.txt.
(Makefile.am): Add uniwbrk/test-uc-wordbreaks.sh to $(TESTS), add
test-uc-wordbreaks to $(check_PROGRAMS), and define
test_uc_wordbreaks_SOURCES and test_uc_wordbreaks_LDADD.
* tests/uniwbrk/test-uc-wordbreaks.sh: New file.
* tests/uniwbrk/test-uc-wordbreaks.c: New file.
author | Daiki Ueno <ueno@gnu.org> |
---|---|
date | Thu, 15 Jan 2015 12:06:30 +0900 |
parents | |
children | ab67a6ee6dc8 |
line wrap: on
line source
/* Word break function test, using test data from UCD. Copyright (C) 2010-2014 Free Software Foundation, Inc. This program is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 3 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this program. If not, see <http://www.gnu.org/licenses/>. */ /* Written by Daiki Ueno <ueno@gnu.org>, 2014. Largely based on unigbrk/test-uc-is-grapheme-break.c, written by Ben Pfaff <blp@cs.stanford.edu>, 2010. */ #include <config.h> /* Specification. */ #include <uniwbrk.h> #include <stdio.h> #include <stdlib.h> #include <string.h> const char * wordbreakproperty_to_string (int wbp) { switch (wbp) { #define CASE(VALUE) case WBP_##VALUE: return #VALUE; CASE(OTHER) CASE(CR) CASE(LF) CASE(NEWLINE) CASE(EXTEND) CASE(FORMAT) CASE(KATAKANA) CASE(ALETTER) CASE(MIDNUMLET) CASE(MIDLETTER) CASE(MIDNUM) CASE(NUMERIC) CASE(EXTENDNUMLET) } abort (); } int main (int argc, char *argv[]) { const char *filename; char line[4096]; int exit_code; FILE *stream; int lineno; if (argc != 2) { fprintf (stderr, "usage: %s FILENAME\n" "where FILENAME is the location of the WordBreakTest.txt\n" "test file.\n", argv[0]); exit (1); } filename = argv[1]; stream = fopen (filename, "r"); if (stream == NULL) { fprintf (stderr, "error during fopen of '%s'\n", filename); exit (1); } exit_code = 0; lineno = 0; while (fgets (line, sizeof line, stream)) { char *comment; const char *p; uint32_t input[100]; char breaks[101]; char breaks_expected[101]; int i; lineno++; memset (breaks, 0, sizeof (breaks)); memset (breaks_expected, 0, sizeof (breaks_expected)); comment = strchr (line, '#'); if (comment != NULL) *comment = '\0'; if (line[strspn (line, " \t\r\n")] == '\0') continue; i = 0; p = line; do { p += strspn (p, " \t\r\n"); if (!strncmp (p, "\303\267" /* ÷ */, 2)) { breaks_expected[i] = 1; p += 2; } else if (!strncmp (p, "\303\227" /* × */, 2)) { breaks_expected[i] = 0; p += 2; } else { fprintf (stderr, "%s:%d.%d: syntax error expecting '÷' or '×'\n", filename, lineno, (int) (p - line + 1)); exit (1); } p += strspn (p, " \t\r\n"); if (*p != '\0') { unsigned int next_int; int n; if (sscanf (p, "%x%n", &next_int, &n) != 1) { fprintf (stderr, "%s:%d.%d: syntax error at '%s' " "expecting hexadecimal Unicode code point number\n", filename, lineno, (int) (p - line + 1), p); exit (1); } p += n; input[i] = next_int; } p += strspn (p, " \t\r\n"); i++; } while (*p != '\0'); u32_wordbreaks (input, i - 1, breaks); /* u32_wordbreaks always set BREAKS[0] to 0. */ breaks[0] = breaks_expected[0] = 1; if (memcmp (breaks, breaks_expected, i - 1) != 0) { int j; fprintf (stderr, "%s:%d: expected: ", filename, lineno); for (j = 0; j < i - 1; j++) { int input_wbp = uc_wordbreak_property (input[j]); fprintf (stderr, "%s U+%04X (%s) ", breaks_expected[j] == 1 ? "\303\267" : "\303\227", input[j], wordbreakproperty_to_string (input_wbp)); } fprintf (stderr, "\n"); fprintf (stderr, "%s:%d: actual: ", filename, lineno); for (j = 0; j < i - 1; j++) { int input_wbp = uc_wordbreak_property (input[j]); fprintf (stderr, "%s U+%04X (%s) ", breaks[j] == 1 ? "\303\267" : "\303\227", input[j], wordbreakproperty_to_string (input_wbp)); } fprintf (stderr, "\n"); exit_code = 1; } } return exit_code; }