a4b143
From 3fde78dba9ad834ab264ad6888672833874b9101 Mon Sep 17 00:00:00 2001
a4b143
From: Dave Reisner <dreisner@archlinux.org>
a4b143
Date: Wed, 18 Sep 2013 11:52:14 -0400
a4b143
Subject: [PATCH] shared/utf8: merge implementations, remove cruft
a4b143
a4b143
This unifies the utf8 handling code which was previously duplicated in
a4b143
udev and systemd.
a4b143
---
a4b143
 TODO                 |   1 -
a4b143
 src/shared/utf8.c    | 279 +++++++++++++--------------------------------------
a4b143
 src/shared/utf8.h    |   3 +-
a4b143
 src/test/test-utf8.c |   9 ++
a4b143
 4 files changed, 81 insertions(+), 211 deletions(-)
a4b143
a4b143
diff --git a/TODO b/TODO
a4b143
index 4c3e14f..8b7325e 100644
a4b143
--- a/TODO
a4b143
+++ b/TODO
a4b143
@@ -599,7 +599,6 @@ Features:
a4b143
 * udev:
a4b143
   - remove src/udev/udev-builtin-firmware.c (CONFIG_FW_LOADER_USER_HELPER=n)
a4b143
   - move to LGPL
a4b143
-  - unify utf8 validator code with shared/
a4b143
   - kill scsi_id
a4b143
   - add trigger --subsystem-match=usb/usb_device device
a4b143
 
a4b143
diff --git a/src/shared/utf8.c b/src/shared/utf8.c
a4b143
index 1a68394..732f0f0 100644
a4b143
--- a/src/shared/utf8.c
a4b143
+++ b/src/shared/utf8.c
a4b143
@@ -51,8 +51,6 @@
a4b143
 #include "utf8.h"
a4b143
 #include "util.h"
a4b143
 
a4b143
-#define FILTER_CHAR '_'
a4b143
-
a4b143
 static inline bool is_unicode_valid(uint32_t ch) {
a4b143
 
a4b143
         if (ch >= 0x110000) /* End of unicode space */
a4b143
@@ -67,17 +65,6 @@ static inline bool is_unicode_valid(uint32_t ch) {
a4b143
         return true;
a4b143
 }
a4b143
 
a4b143
-static inline bool is_continuation_char(uint8_t ch) {
a4b143
-        if ((ch & 0xc0) != 0x80) /* 10xxxxxx */
a4b143
-                return false;
a4b143
-        return true;
a4b143
-}
a4b143
-
a4b143
-static inline void merge_continuation_char(uint32_t *u_ch, uint8_t ch) {
a4b143
-        *u_ch <<= 6;
a4b143
-        *u_ch |= ch & 0x3f;
a4b143
-}
a4b143
-
a4b143
 static bool is_unicode_control(uint32_t ch) {
a4b143
 
a4b143
         /*
a4b143
@@ -90,163 +77,97 @@ static bool is_unicode_control(uint32_t ch) {
a4b143
                 (0x7F <= ch && ch <= 0x9F);
a4b143
 }
a4b143
 
a4b143
-bool utf8_is_printable(const char* str, size_t length) {
a4b143
-        uint32_t val = 0;
a4b143
-        uint32_t min = 0;
a4b143
-        const uint8_t *p;
a4b143
+/* count of characters used to encode one unicode char */
a4b143
+static int utf8_encoded_expected_len(const char *str) {
a4b143
+        unsigned char c = (unsigned char)str[0];
a4b143
 
a4b143
-        assert(str);
a4b143
+        if (c < 0x80)
a4b143
+                return 1;
a4b143
+        if ((c & 0xe0) == 0xc0)
a4b143
+                return 2;
a4b143
+        if ((c & 0xf0) == 0xe0)
a4b143
+                return 3;
a4b143
+        if ((c & 0xf8) == 0xf0)
a4b143
+                return 4;
a4b143
+        if ((c & 0xfc) == 0xf8)
a4b143
+                return 5;
a4b143
+        if ((c & 0xfe) == 0xfc)
a4b143
+                return 6;
a4b143
+        return 0;
a4b143
+}
a4b143
 
a4b143
-        for (p = (const uint8_t*) str; length; p++, length--) {
a4b143
-                if (*p < 128) {
a4b143
-                        val = *p;
a4b143
-                } else {
a4b143
-                        if ((*p & 0xe0) == 0xc0) { /* 110xxxxx two-char seq. */
a4b143
-                                min = 128;
a4b143
-                                val = (uint32_t) (*p & 0x1e);
a4b143
-                                goto ONE_REMAINING;
a4b143
-                        } else if ((*p & 0xf0) == 0xe0) { /* 1110xxxx three-char seq.*/
a4b143
-                                min = (1 << 11);
a4b143
-                                val = (uint32_t) (*p & 0x0f);
a4b143
-                                goto TWO_REMAINING;
a4b143
-                        } else if ((*p & 0xf8) == 0xf0) { /* 11110xxx four-char seq */
a4b143
-                                min = (1 << 16);
a4b143
-                                val = (uint32_t) (*p & 0x07);
a4b143
-                        } else
a4b143
-                                return false;
a4b143
-
a4b143
-                        p++;
a4b143
-                        length--;
a4b143
-                        if (!length || !is_continuation_char(*p))
a4b143
-                                return false;
a4b143
-                        merge_continuation_char(&val, *p);
a4b143
-
a4b143
-                TWO_REMAINING:
a4b143
-                        p++;
a4b143
-                        length--;
a4b143
-                        if (!is_continuation_char(*p))
a4b143
-                                return false;
a4b143
-                        merge_continuation_char(&val, *p);
a4b143
-
a4b143
-                ONE_REMAINING:
a4b143
-                        p++;
a4b143
-                        length--;
a4b143
-                        if (!is_continuation_char(*p))
a4b143
-                                return false;
a4b143
-                        merge_continuation_char(&val, *p);
a4b143
-
a4b143
-                        if (val < min)
a4b143
-                                return false;
a4b143
-                }
a4b143
+/* decode one unicode char */
a4b143
+static int utf8_encoded_to_unichar(const char *str) {
a4b143
+        int unichar;
a4b143
+        int len;
a4b143
+        int i;
a4b143
 
a4b143
-                if (is_unicode_control(val))
a4b143
-                        return false;
a4b143
+        len = utf8_encoded_expected_len(str);
a4b143
+        switch (len) {
a4b143
+        case 1:
a4b143
+                return (int)str[0];
a4b143
+        case 2:
a4b143
+                unichar = str[0] & 0x1f;
a4b143
+                break;
a4b143
+        case 3:
a4b143
+                unichar = (int)str[0] & 0x0f;
a4b143
+                break;
a4b143
+        case 4:
a4b143
+                unichar = (int)str[0] & 0x07;
a4b143
+                break;
a4b143
+        case 5:
a4b143
+                unichar = (int)str[0] & 0x03;
a4b143
+                break;
a4b143
+        case 6:
a4b143
+                unichar = (int)str[0] & 0x01;
a4b143
+                break;
a4b143
+        default:
a4b143
+                return -1;
a4b143
         }
a4b143
 
a4b143
-        return true;
a4b143
+        for (i = 1; i < len; i++) {
a4b143
+                if (((int)str[i] & 0xc0) != 0x80)
a4b143
+                        return -1;
a4b143
+                unichar <<= 6;
a4b143
+                unichar |= (int)str[i] & 0x3f;
a4b143
+        }
a4b143
+
a4b143
+        return unichar;
a4b143
 }
a4b143
 
a4b143
-static char* utf8_validate(const char *str, char *output) {
a4b143
-        uint32_t val = 0;
a4b143
-        uint32_t min = 0;
a4b143
-        const uint8_t *p, *last;
a4b143
-        int size;
a4b143
-        uint8_t *o;
a4b143
+bool utf8_is_printable(const char* str, size_t length) {
a4b143
+        const uint8_t *p;
a4b143
 
a4b143
         assert(str);
a4b143
 
a4b143
-        o = (uint8_t*) output;
a4b143
-        for (p = (const uint8_t*) str; *p; p++) {
a4b143
-                if (*p < 128) {
a4b143
-                        if (o)
a4b143
-                                *o = *p;
a4b143
-                } else {
a4b143
-                        last = p;
a4b143
-
a4b143
-                        if ((*p & 0xe0) == 0xc0) { /* 110xxxxx two-char seq. */
a4b143
-                                size = 2;
a4b143
-                                min = 128;
a4b143
-                                val = (uint32_t) (*p & 0x1e);
a4b143
-                                goto ONE_REMAINING;
a4b143
-                        } else if ((*p & 0xf0) == 0xe0) { /* 1110xxxx three-char seq.*/
a4b143
-                                size = 3;
a4b143
-                                min = (1 << 11);
a4b143
-                                val = (uint32_t) (*p & 0x0f);
a4b143
-                                goto TWO_REMAINING;
a4b143
-                        } else if ((*p & 0xf8) == 0xf0) { /* 11110xxx four-char seq */
a4b143
-                                size = 4;
a4b143
-                                min = (1 << 16);
a4b143
-                                val = (uint32_t) (*p & 0x07);
a4b143
-                        } else
a4b143
-                                goto error;
a4b143
-
a4b143
-                        p++;
a4b143
-                        if (!is_continuation_char(*p))
a4b143
-                                goto error;
a4b143
-                        merge_continuation_char(&val, *p);
a4b143
-
a4b143
-                TWO_REMAINING:
a4b143
-                        p++;
a4b143
-                        if (!is_continuation_char(*p))
a4b143
-                                goto error;
a4b143
-                        merge_continuation_char(&val, *p);
a4b143
-
a4b143
-                ONE_REMAINING:
a4b143
-                        p++;
a4b143
-                        if (!is_continuation_char(*p))
a4b143
-                                goto error;
a4b143
-                        merge_continuation_char(&val, *p);
a4b143
-
a4b143
-                        if (val < min)
a4b143
-                                goto error;
a4b143
-
a4b143
-                        if (!is_unicode_valid(val))
a4b143
-                                goto error;
a4b143
-
a4b143
-                        if (o) {
a4b143
-                                memcpy(o, last, (size_t) size);
a4b143
-                                o += size;
a4b143
-                        }
a4b143
-
a4b143
-                        continue;
a4b143
-
a4b143
-                error:
a4b143
-                        if (o) {
a4b143
-                                *o = FILTER_CHAR;
a4b143
-                                p = last; /* We retry at the next character */
a4b143
-                        } else
a4b143
-                                goto failure;
a4b143
-                }
a4b143
+        for (p = (const uint8_t*) str; length; p++) {
a4b143
+                int encoded_len = utf8_encoded_valid_unichar((const char *)p);
a4b143
+                int32_t val = utf8_encoded_to_unichar((const char*)p);
a4b143
 
a4b143
-                if (o)
a4b143
-                        o++;
a4b143
-        }
a4b143
+                if (encoded_len < 0 || val < 0 || is_unicode_control(val))
a4b143
+                        return false;
a4b143
 
a4b143
-        if (o) {
a4b143
-                *o = '\0';
a4b143
-                return output;
a4b143
+                length -= encoded_len;
a4b143
         }
a4b143
 
a4b143
-        return (char*) str;
a4b143
-
a4b143
-failure:
a4b143
-        return NULL;
a4b143
-}
a4b143
-
a4b143
-char* utf8_is_valid (const char *str) {
a4b143
-        return utf8_validate(str, NULL);
a4b143
+        return true;
a4b143
 }
a4b143
 
a4b143
-char* utf8_filter (const char *str) {
a4b143
-        char *new_str;
a4b143
+const char *utf8_is_valid(const char *str) {
a4b143
+        const uint8_t *p;
a4b143
 
a4b143
         assert(str);
a4b143
 
a4b143
-        new_str = malloc(strlen(str) + 1);
a4b143
-        if (!new_str)
a4b143
-                return NULL;
a4b143
+        for (p = (const uint8_t*) str; *p; ) {
a4b143
+                int len = utf8_encoded_valid_unichar((const char *)p);
a4b143
+
a4b143
+                if (len < 0)
a4b143
+                        return NULL;
a4b143
+
a4b143
+                p += len;
a4b143
+        }
a4b143
 
a4b143
-        return utf8_validate(str, new_str);
a4b143
+        return str;
a4b143
 }
a4b143
 
a4b143
 char *ascii_is_valid(const char *str) {
a4b143
@@ -318,64 +239,6 @@ char *utf16_to_utf8(const void *s, size_t length) {
a4b143
         return r;
a4b143
 }
a4b143
 
a4b143
-/* count of characters used to encode one unicode char */
a4b143
-static int utf8_encoded_expected_len(const char *str) {
a4b143
-        unsigned char c = (unsigned char)str[0];
a4b143
-
a4b143
-        if (c < 0x80)
a4b143
-                return 1;
a4b143
-        if ((c & 0xe0) == 0xc0)
a4b143
-                return 2;
a4b143
-        if ((c & 0xf0) == 0xe0)
a4b143
-                return 3;
a4b143
-        if ((c & 0xf8) == 0xf0)
a4b143
-                return 4;
a4b143
-        if ((c & 0xfc) == 0xf8)
a4b143
-                return 5;
a4b143
-        if ((c & 0xfe) == 0xfc)
a4b143
-                return 6;
a4b143
-        return 0;
a4b143
-}
a4b143
-
a4b143
-/* decode one unicode char */
a4b143
-static int utf8_encoded_to_unichar(const char *str) {
a4b143
-        int unichar;
a4b143
-        int len;
a4b143
-        int i;
a4b143
-
a4b143
-        len = utf8_encoded_expected_len(str);
a4b143
-        switch (len) {
a4b143
-        case 1:
a4b143
-                return (int)str[0];
a4b143
-        case 2:
a4b143
-                unichar = str[0] & 0x1f;
a4b143
-                break;
a4b143
-        case 3:
a4b143
-                unichar = (int)str[0] & 0x0f;
a4b143
-                break;
a4b143
-        case 4:
a4b143
-                unichar = (int)str[0] & 0x07;
a4b143
-                break;
a4b143
-        case 5:
a4b143
-                unichar = (int)str[0] & 0x03;
a4b143
-                break;
a4b143
-        case 6:
a4b143
-                unichar = (int)str[0] & 0x01;
a4b143
-                break;
a4b143
-        default:
a4b143
-                return -1;
a4b143
-        }
a4b143
-
a4b143
-        for (i = 1; i < len; i++) {
a4b143
-                if (((int)str[i] & 0xc0) != 0x80)
a4b143
-                        return -1;
a4b143
-                unichar <<= 6;
a4b143
-                unichar |= (int)str[i] & 0x3f;
a4b143
-        }
a4b143
-
a4b143
-        return unichar;
a4b143
-}
a4b143
-
a4b143
 /* expected size used to encode one unicode char */
a4b143
 static int utf8_unichar_to_encoded_len(int unichar) {
a4b143
         if (unichar < 0x80)
a4b143
diff --git a/src/shared/utf8.h b/src/shared/utf8.h
a4b143
index 7a5608c..22e1346 100644
a4b143
--- a/src/shared/utf8.h
a4b143
+++ b/src/shared/utf8.h
a4b143
@@ -25,12 +25,11 @@
a4b143
 
a4b143
 #include "macro.h"
a4b143
 
a4b143
-char *utf8_is_valid(const char *s) _pure_;
a4b143
+const char *utf8_is_valid(const char *s) _pure_;
a4b143
 char *ascii_is_valid(const char *s) _pure_;
a4b143
 
a4b143
 bool utf8_is_printable(const char* str, size_t length) _pure_;
a4b143
 
a4b143
-char *utf8_filter(const char *s);
a4b143
 char *ascii_filter(const char *s);
a4b143
 
a4b143
 char *utf16_to_utf8(const void *s, size_t length);
a4b143
diff --git a/src/test/test-utf8.c b/src/test/test-utf8.c
a4b143
index d2b9771..26cc37b 100644
a4b143
--- a/src/test/test-utf8.c
a4b143
+++ b/src/test/test-utf8.c
a4b143
@@ -47,6 +47,12 @@ static void test_udev_encode_string(void) {
a4b143
         assert_se(expect_encoded_as("s/ash/ng", "s\\x2fash\\x2fng"));
a4b143
 }
a4b143
 
a4b143
+static void test_utf8_is_printable(void) {
a4b143
+        assert_se(utf8_is_printable("ascii is valid\tunicode", 22));
a4b143
+        assert_se(utf8_is_printable("\342\204\242", 3));
a4b143
+        assert_se(!utf8_is_printable("\341\204", 2));
a4b143
+}
a4b143
+
a4b143
 static void test_utf8_is_valid(void) {
a4b143
         assert_se(utf8_is_valid("ascii is valid unicode"));
a4b143
         assert_se(utf8_is_valid("\341\204\242"));
a4b143
@@ -55,5 +61,8 @@ static void test_utf8_is_valid(void) {
a4b143
 
a4b143
 int main(int argc, char *argv[]) {
a4b143
         test_utf8_is_valid();
a4b143
+        test_utf8_is_printable();
a4b143
         test_udev_encode_string();
a4b143
+
a4b143
+        return 0;
a4b143
 }