X-Git-Url: http://git.tuebingen.mpg.de/?p=paraslash.git;a=blobdiff_plain;f=string.c;h=6033a008dbf154953de673f85048eb78812d1722;hp=d9dcc62d2ea45efc6a3c2899185a2033d5e74339;hb=534a94f441767947874cb15d18211edf758e9277;hpb=5971e3303cf800603622c2475e07fdccc33e4915

diff --git a/string.c b/string.c
index d9dcc62d..6033a008 100644
--- a/string.c
+++ b/string.c
@@ -6,19 +6,16 @@
 
 /** \file string.c Memory allocation and string handling functions. */
 
-#define _GNU_SOURCE
+#include "para.h"
 
 #include <pwd.h>
 #include <sys/utsname.h> /* uname() */
-
 #include <string.h>
 #include <regex.h>
-
 #include <langinfo.h>
 #include <wchar.h>
 #include <wctype.h>
 
-#include "para.h"
 #include "string.h"
 #include "error.h"
 
@@ -806,15 +803,18 @@ err:
  * Split a buffer into words.
  *
  * This parser honors single and double quotes, backslash-escaped characters
- * and special characters like \p \\n. The result contains pointers to copies
- * of the words contained in \a buf and has to be freed by using \ref
- * free_argv().
+ * and special characters like \\n. The result contains pointers to copies of
+ * the words contained in buf and has to be freed by using \ref free_argv().
  *
  * \param buf The buffer to be split.
  * \param delim Each character in this string is treated as a separator.
  * \param result The array of words is returned here.
  *
- * \return Number of words in \a buf, negative on errors.
+ * It's OK to pass NULL as the buffer argument. This is equivalent to passing
+ * the empty string.
+ *
+ * \return Number of words in buf, negative on errors. The array returned
+ * through the result pointer is NULL terminated.
  */
 int create_argv(const char *buf, const char *delim, char ***result)
 {
@@ -1070,3 +1070,66 @@ __must_check int strwidth(const char *s, size_t *result)
 	free(dest);
 	return 1;
 }
+
+/**
+ * Truncate and sanitize a (wide character) string.
+ *
+ * This replaces all non-printable characters by spaces and makes sure that the
+ * modified string does not exceed the given maximal width.
+ *
+ * \param src The source string in multi-byte form.
+ * \param max_width The maximal number of cells the result may occupy.
+ * \param result Sanitized multi-byte string, must be freed by caller.
+ * \param width The width of the sanitized string, always <= max_width.
+ *
+ * The function is wide-character aware but falls back to C strings for
+ * non-UTF-8 locales.
+ *
+ * \return Standard. On success, *result points to a sanitized copy of the
+ * given string. This copy was allocated with malloc() and should hence be
+ * freed when the caller is no longer interested in the result.
+ *
+ * The function fails if the given string contains an invalid multibyte
+ * sequence. In this case, *result is set to NULL, and *width to zero.
+ */
+__must_check int sanitize_str(const char *src, size_t max_width,
+		char **result, size_t *width)
+{
+	mbstate_t state;
+	static wchar_t *wcs;
+	size_t num_wchars, n;
+
+	if (!utf8_mode()) {
+		*result = para_strdup(src);
+		/* replace non-printable characters by spaces */
+		for (n = 0; n < max_width && src[n]; n++) {
+			if (!isprint((unsigned char)src[n]))
+				(*result)[n] = ' ';
+		}
+		(*result)[n] = '\0';
+		*width = n;
+		return 0;
+	}
+	*result = NULL;
+	*width = 0;
+	memset(&state, 0, sizeof(state));
+	num_wchars = mbsrtowcs(NULL, &src, 0, &state);
+	if (num_wchars == (size_t)-1)
+		return -ERRNO_TO_PARA_ERROR(errno);
+	wcs = para_malloc((num_wchars + 1) * sizeof(*wcs));
+	memset(&state, 0, sizeof(state));
+	num_wchars = mbsrtowcs(wcs, &src, num_wchars + 1, &state);
+	assert(num_wchars != (size_t)-1);
+	for (n = 0; n < num_wchars && *width < max_width; n++) {
+		if (!iswprint(wcs[n]))
+			wcs[n] = L' ';
+		*width += xwcwidth(wcs[n], *width);
+	}
+	wcs[n] = L'\0';
+	n = wcstombs(NULL, wcs, 0) + 1;
+	*result = para_malloc(n);
+	num_wchars = wcstombs(*result, wcs, n);
+	assert(num_wchars != (size_t)-1);
+	free(wcs);
+	return 1;
+}