summaryrefslogtreecommitdiff
path: root/test
diff options
context:
space:
mode:
authorSteven G. Johnson <stevenj@mit.edu>2015-03-08 21:23:43 (GMT)
committer Steven G. Johnson <stevenj@mit.edu>2015-03-12 16:10:19 (GMT)
commita4c84d2063bb6f2218bc20f89e59001319361f6d (patch)
treebf9799783e0f66c67f0303ea13a778fe51eff539 /test
parent08f101a9e8c6a72dfdb1c9b913df880e13a36333 (diff)
downloadlibutf8proc-a4c84d2063bb6f2218bc20f89e59001319361f6d.tar.gz
libutf8proc-a4c84d2063bb6f2218bc20f89e59001319361f6d.tar.bz2
fix #2: add charwidth function
Diffstat (limited to 'test')
-rw-r--r--test/charwidth.c61
-rw-r--r--test/printproperty.c10
2 files changed, 67 insertions, 4 deletions
diff --git a/test/charwidth.c b/test/charwidth.c
new file mode 100644
index 0000000..bdae9fa
--- a/dev/null
+++ b/test/charwidth.c
@@ -0,0 +1,61 @@
+#include "tests.h"
+#include <ctype.h>
+#include <wchar.h>
+
+int my_isprint(int c) {
+ int cat = utf8proc_get_property(c)->category;
+ return (UTF8PROC_CATEGORY_LU <= cat && cat <= UTF8PROC_CATEGORY_ZS) ||
+ (c == 0x0601 || c == 0x0602 || c == 0x0603 || c == 0x06dd);
+}
+
+int main(int argc, char **argv)
+{
+ int prevc, c, error = 0;
+
+ (void) argc; /* unused */
+ (void) argv; /* unused */
+
+ /* some simple sanity tests of the character widths */
+ for (c = 0; c <= 0x110000; ++c) {
+ int cat = utf8proc_get_property(c)->category;
+ int w = utf8proc_charwidth(c);
+ if ((cat == UTF8PROC_CATEGORY_MN || cat == UTF8PROC_CATEGORY_ME) &&
+ w > 0) {
+ fprintf(stderr, "nonzero width %d for combining char %x\n", w, c);
+ error = 1;
+ }
+ if (c <= 127 && ((!isprint(c) && w > 0) ||
+ (isprint(c) && wcwidth(c) != w))) {
+ fprintf(stderr, "wcwidth %d mismatch %d for %s ASCII %x\n",
+ wcwidth(c), w,
+ isprint(c) ? "printable" : "non-printable", c);
+ error = 1;
+ }
+ if (!my_isprint(c) && w > 0) {
+ fprintf(stderr, "non-printing %x had width %d\n", c, w);
+ error = 1;
+ }
+ }
+ check(!error, "utf8proc_charwidth FAILED tests.");
+
+ /* print some other information by compariing with system wcwidth */
+ printf("Mismatches with system wcwidth (not necessarily errors):\n");
+ for (c = 0; c <= 0x110000; ++c) {
+ int w = utf8proc_charwidth(c);
+ int wc = wcwidth(c);
+ if (sizeof(wchar_t) == 2 && c >= (1<<16)) continue;
+#if 0
+ /* lots of these errors for out-of-date system unicode tables */
+ if (wc == -1 && my_isprint(c) && w > 0)
+ printf(" wcwidth(%x) = -1 for printable char\n", c);
+#endif
+ if (wc == -1 && !my_isprint(c) && w > 0)
+ printf(" wcwidth(%x) = -1 for non-printable width-%d char\n", c, w);
+ if (wc >= 0 && wc != w)
+ printf(" wcwidth(%x) = %d != charwidth %d\n", c, wc, w);
+ }
+
+ printf("Character-width tests SUCCEEDED.\n");
+
+ return 0;
+}
diff --git a/test/printproperty.c b/test/printproperty.c
index 37e4fa3..6be8cb1 100644
--- a/test/printproperty.c
+++ b/test/printproperty.c
@@ -11,7 +11,7 @@ int main(int argc, char **argv)
check(sscanf(argv[i],"%x",&c) == 1, "invalid hex input %s", argv[i]);
const utf8proc_property_t *p = utf8proc_get_property(c);
printf("U+%s:\n"
- " category = %d\n"
+ " category = %s\n"
" combining_class = %d\n"
" bidi_class = %d\n"
" decomp_type = %d\n"
@@ -24,9 +24,10 @@ int main(int argc, char **argv)
" comp_exclusion = %d\n"
" ignorable = %d\n"
" control_boundary = %d\n"
- " boundclass = %d\n",
+ " boundclass = %d\n"
+ " charwidth = %d\n",
argv[i],
- p->category,
+ utf8proc_category_string(c),
p->combining_class,
p->bidi_class,
p->decomp_type,
@@ -39,7 +40,8 @@ int main(int argc, char **argv)
p->comp_exclusion,
p->ignorable,
p->control_boundary,
- p->boundclass);
+ p->boundclass,
+ utf8proc_charwidth(c));
}
return 0;
}