lib/mbalign: add mbs_safe_width() from tt.c

Signed-off-by: Karel Zak <kzak@redhat.com>

1 files changed, 136 insertions, 1 deletions

diff --git a/lib/mbsalign.c b/lib/mbsalign.c
index 89ef03298..5933f1863 100644
--- a/lib/mbsalign.c
+++ b/lib/mbsalign.c
@@ -23,17 +23,152 @@
 #include <stdio.h>
 #include <stdbool.h>
 #include <limits.h>
+#include <ctype.h>
 
 #include "c.h"
 #include "mbsalign.h"
 #include "widechar.h"
 
-
 #ifdef HAVE_WIDECHAR
 /* Replace non printable chars.
    Note \t and \n etc. are non printable.
    Return 1 if replacement made, 0 otherwise.  */
 
+/*
+ * Counts number of cells in multibyte string. For all control and
+ * non-printable chars is the result width enlarged to store \x?? hex
+ * sequence. See mbs_safe_encode().
+ */
+size_t mbs_safe_width(const char *s)
+{
+	mbstate_t st;
+	const char *p = s;
+	size_t width = 0;
+
+	memset(&st, 0, sizeof(st));
+
+	while (p && *p) {
+		if (iscntrl((unsigned char) *p)) {
+			width += 4;			/* *p encoded to \x?? */
+			p++;
+		}
+#ifdef HAVE_WIDECHAR
+		else {
+			wchar_t wc;
+			size_t len = mbrtowc(&wc, p, MB_CUR_MAX, &st);
+
+			if (len == 0)
+				break;
+
+			if (len == (size_t) -1 || len == (size_t) -2) {
+				len = 1;
+				width += (isprint((unsigned char) *p) ? 1 : 4);
+
+			} if (!iswprint(wc))
+				width += len * 4;	/* hex encode whole sequence */
+			else
+				width += wcwidth(wc);	/* number of cells */
+			p += len;
+		}
+#else
+		else if (!isprint((unsigned char) *p)) {
+			width += 4;			/* *p encoded to \x?? */
+			p++;
+		} else {
+			width++;
+			p++;
+		}
+#endif
+	}
+
+	return width;
+}
+
+/*
+ * Returns allocated string where all control and non-printable chars are
+ * replaced with \x?? hex sequence.
+ */
+char *mbs_safe_encode(const char *s, size_t *width)
+{
+	mbstate_t st;
+	const char *p = s;
+	char *res, *r;
+	size_t sz = s ? strlen(s) : 0;
+
+
+	if (!sz)
+		return NULL;
+
+	memset(&st, 0, sizeof(st));
+
+	res = malloc((sz * 4) + 1);
+	if (!res)
+		return NULL;
+
+	r = res;
+	*width = 0;
+
+	while (p && *p) {
+		if (iscntrl((unsigned char) *p)) {
+			sprintf(r, "\\x%02x", (unsigned char) *p);
+			r += 4;
+			*width += 4;
+			p++;
+		}
+#ifdef HAVE_WIDECHAR
+		else {
+			wchar_t wc;
+			size_t len = mbrtowc(&wc, p, MB_CUR_MAX, &st);
+
+			if (len == 0)
+				break;		/* end of string */
+
+			if (len == (size_t) -1 || len == (size_t) -2) {
+				len = 1;
+				/*
+				 * Not valid multibyte sequence -- maybe it's
+				 * printable char according to the current locales.
+				 */
+				if (!isprint((unsigned char) *p)) {
+					sprintf(r, "\\x%02x", (unsigned char) *p);
+					r += 4;
+					*width += 4;
+				} else {
+					width++;
+					*r++ = *p;
+				}
+			} else if (!iswprint(wc)) {
+				size_t i;
+				for (i = 0; i < len; i++) {
+					sprintf(r, "\\x%02x", (unsigned char) *p);
+					r += 4;
+					*width += 4;
+				}
+			} else {
+				memcpy(r, p, len);
+				r += len;
+				*width += wcwidth(wc);
+			}
+			p += len;
+		}
+#else
+		else if (!isprint((unsigned char) *p)) {
+			sprintf(r, "\\x%02x", (unsigned char) *p);
+			p++;
+			r += 4;
+			*width += 4;
+		} else {
+			*r++ = *p++;
+			*width++;
+		}
+#endif
+	}
+
+	*r = '\0';
+
+	return res;
+}
+
 static bool
 wc_ensure_printable (wchar_t *wchars)
 {