From d6daf37df4ccd7589d2d8f6911bf7270f12d1672 Mon Sep 17 00:00:00 2001
From: nicm <nicm>
Date: Thu, 12 Nov 2015 12:19:57 +0000
Subject: Tidy utf8.c a little: build table on first use, and make utf8_width
 take a u_int rather than splitting and then combining again in utf8_split.

---
 utf8.c | 81 ++++++++++++++++++++++++------------------------------------------
 1 file changed, 29 insertions(+), 52 deletions(-)

(limited to 'utf8.c')

diff --git a/utf8.c b/utf8.c
index cb20ea6d..0926f4bc 100644
--- a/utf8.c
+++ b/utf8.c
@@ -35,7 +35,7 @@ struct utf8_width_entry {
 };
 
 /* Sorted, then repeatedly split in the middle to balance the tree. */
-struct utf8_width_entry utf8_width_table[] = {
+static struct utf8_width_entry utf8_width_table[] = {
 	{ 0x00b41, 0x00b44, 0, NULL, NULL },
 	{ 0x008e4, 0x00902, 0, NULL, NULL },
 	{ 0x006d6, 0x006dd, 0, NULL, NULL },
@@ -344,12 +344,9 @@ struct utf8_width_entry utf8_width_table[] = {
 	{ 0xe0100, 0xe01ef, 0, NULL, NULL },
 	{ 0x100000, 0x10fffd, 0, NULL, NULL },
 };
+static struct utf8_width_entry	*utf8_width_root = NULL;
 
-struct utf8_width_entry	*utf8_width_root = NULL;
-
-int	utf8_overlap(struct utf8_width_entry *, struct utf8_width_entry *);
-u_int	utf8_combine(const struct utf8_data *);
-u_int	utf8_width(const struct utf8_data *);
+static void	utf8_build(void);
 
 /* Set a single character. */
 void
@@ -405,40 +402,20 @@ utf8_append(struct utf8_data *utf8data, u_char ch)
 	if (utf8data->have != utf8data->size)
 		return (1);
 
-	utf8data->width = utf8_width(utf8data);
-	return (0);
-}
-
-/* Check if two width tree entries overlap. */
-int
-utf8_overlap(struct utf8_width_entry *item1, struct utf8_width_entry *item2)
-{
-	if (item1->first >= item2->first && item1->first <= item2->last)
-		return (1);
-	if (item1->last >= item2->first && item1->last <= item2->last)
-		return (1);
-	if (item2->first >= item1->first && item2->first <= item1->last)
-		return (1);
-	if (item2->last >= item1->first && item2->last <= item1->last)
-		return (1);
+	utf8data->width = utf8_width(utf8_combine(utf8data));
 	return (0);
 }
 
 /* Build UTF-8 width tree. */
-void
+static void
 utf8_build(void)
 {
 	struct utf8_width_entry	**ptr, *item, *node;
-	u_int			  i, j;
+	u_int			  i;
 
 	for (i = 0; i < nitems(utf8_width_table); i++) {
 		item = &utf8_width_table[i];
 
-		for (j = 0; j < nitems(utf8_width_table); j++) {
-			if (i != j && utf8_overlap(item, &utf8_width_table[j]))
-				log_fatalx("utf8 overlap: %u %u", i, j);
-		}
-
 		ptr = &utf8_width_root;
 		while (*ptr != NULL) {
 			node = *ptr;
@@ -451,6 +428,27 @@ utf8_build(void)
 	}
 }
 
+/* Lookup width of UTF-8 data in tree. */
+u_int
+utf8_width(u_int uc)
+{
+	struct utf8_width_entry	*item;
+
+	if (utf8_width_root == NULL)
+		utf8_build();
+
+	item = utf8_width_root;
+	while (item != NULL) {
+		if (uc < item->first)
+			item = item->left;
+		else if (uc > item->last)
+			item = item->right;
+		else
+			return (item->width);
+	}
+	return (1);
+}
+
 /* Combine UTF-8 into 32-bit Unicode. */
 u_int
 utf8_combine(const struct utf8_data *utf8data)
@@ -481,7 +479,7 @@ utf8_combine(const struct utf8_data *utf8data)
 	return (value);
 }
 
-/* Split a UTF-8 character. */
+/* Split 32-bit Unicode into UTF-8. */
 int
 utf8_split(u_int uc, struct utf8_data *utf8data)
 {
@@ -505,7 +503,7 @@ utf8_split(u_int uc, struct utf8_data *utf8data)
 		utf8data->data[3] = 0x80 | (uc & 0x3f);
 	} else
 		return (-1);
-	utf8data->width = utf8_width(utf8data);
+	utf8data->width = utf8_width(uc);
 	return (0);
 }
 
@@ -522,27 +520,6 @@ utf8_split2(u_int uc, u_char *ptr)
 	return (1);
 }
 
-/* Lookup width of UTF-8 data in tree. */
-u_int
-utf8_width(const struct utf8_data *utf8data)
-{
-	struct utf8_width_entry	*item;
-	u_int			 value;
-
-	value = utf8_combine(utf8data);
-
-	item = utf8_width_root;
-	while (item != NULL) {
-		if (value < item->first)
-			item = item->left;
-		else if (value > item->last)
-			item = item->right;
-		else
-			return (item->width);
-	}
-	return (1);
-}
-
 /*
  * Encode len characters from src into dst, which is guaranteed to have four
  * bytes available for each character from src (for \abc or UTF-8) plus space
-- 
cgit 


From a209ea3953ba16742f6f6bb19b76ffdb1200960e Mon Sep 17 00:00:00 2001
From: nicm <nicm>
Date: Thu, 12 Nov 2015 12:43:36 +0000
Subject: Add utf8_padcstr and use it to align columns in list-keys.

---
 utf8.c | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

(limited to 'utf8.c')

diff --git a/utf8.c b/utf8.c
index 0926f4bc..db738020 100644
--- a/utf8.c
+++ b/utf8.c
@@ -713,3 +713,24 @@ utf8_trimcstr(const char *s, u_int width)
 	free(tmp);
 	return (out);
 }
+
+/* Pad UTF-8 string to width. Caller frees. */
+char *
+utf8_padcstr(const char *s, u_int width)
+{
+	size_t	 slen;
+	char	*out;
+	u_int	  n, i;
+
+	n = utf8_cstrwidth(s);
+	if (n >= width)
+		return (xstrdup(s));
+
+	slen = strlen(s);
+	out = xmalloc(slen + 1 + (width - n));
+	memcpy(out, s, slen);
+	for (i = n; i < width; i++)
+		out[slen++] = ' ';
+	out[slen] = '\0';
+	return (out);
+}
-- 
cgit