On Thu, May 14, 2026 at 7:40 PM Ivan Zhakov <[email protected]> wrote:
> On Thu, 14 May 2026 at 14:13, Timofei Zhakov <[email protected]> wrote:
>
>> This function counts real printable UTF characters in a string. It
>> currently contains a table of all patterns that is manually checked. I
>> believe it was stolen from elsewhere a long time ago. Before we had
>> utf8proc as a required dependency.
>>
>> I have a few reasons to rewrite it to use the library instead;
>>
>> 1. I'm pretty sure nobody would ever care to update the dataset. On
>> the other hand, utf8proc bundles all available information about the
>> latest Unicode version that is supported on the current platform.
>>
>> 2. There is also a property that defines *display* width, that
>> basically makes symbols like emojis wider than normal characters even
>> on monospace fonts.
>>
>> (For context I want to fix indentation in places throughout our
>> cmdline like the authors in 'svn list -v' that mess up the tables.
>> This is where a function like that will be useful.)
>>
>> 3. Cleanup redundant code.
>>
>> 4. It might be slightly faster to use their dataset because utf8proc
>> only accesses a table in static memory twice (for address and then
>> retrieves properties) instead of binary searching and checking all
>> ranges. Maybe it's slower though idk.
>>
>> Thoughts?
>>
>> Sounds good to me.
>
> Regarding potential performance regression: is it something we can
> measure? As far as I understand svn_utf_cstring_utf8_width() is not used
> for performance critical code, but it would be nice to know if there is
> significant performance regression anyway.
>
>
Well, I guess it was a bad argument because the utf8proc version looks well
optimised anyway and we already use in multiple places.
I'm done with the first version of a patch to make this switch. Please find
it in the attachments. I also added a unit test to confirm.
I think since this function depends on utf8proc it should be implemented in
utf8proc.c file. Hence utf_width.c can be removed entirely. I guess the
main reason it was in a separate file was to keep the dataset away from
other code.
By the way, I also found that before traversing the string, this function
calls svn_utf__cstring_is_valid() to check validity. I
think utf8proc_iterate() does the same check, but I really am not
sure. svn_utf__cstring_is_valid() skips all first octets with values >0x80.
Then it basically does the same work utf8proc_iterate() would.
--
Timofei Zhakov
Index: subversion/libsvn_subr/utf8proc.c
===================================================================
--- subversion/libsvn_subr/utf8proc.c (revision 1934258)
+++ subversion/libsvn_subr/utf8proc.c (working copy)
@@ -59,7 +59,6 @@ svn_utf__utf8proc_runtime_version(void)
#ifdef UTF8PROC_STATIC
/* Unused static function warning removal hack. */
SVN_UNUSED(utf8proc_category_string);
- SVN_UNUSED(utf8proc_charwidth);
SVN_UNUSED(utf8proc_charwidth_ambiguous);
SVN_UNUSED(utf8proc_grapheme_break);
SVN_UNUSED(utf8proc_islower);
@@ -606,3 +605,40 @@ svn_utf__fuzzy_escape(const char *src, apr_size_t
return result->data;
}
+
+int
+svn_utf_cstring_utf8_width(const char *cstr)
+{
+ int width = 0;
+
+ if (*cstr == '\0')
+ return 0;
+
+ /* Ensure the conversion below doesn't fail because of encoding errors. */
+ if (!svn_utf__cstring_is_valid(cstr))
+ return -1;
+
+ /* Convert the UTF-8 string to UTF-32 (UCS4) which is the format
+ * utf8proc_charwidth() expects, and get the width of each character.
+ * We don't need much error checking since the input is valid UTF-8. */
+ while (*cstr)
+ {
+ apr_int32_t ucs;
+ int w;
+
+ int nbytes = utf8proc_iterate((apr_byte_t*)cstr, -1, &ucs);
+
+ if (nbytes < 0)
+ return -1;
+
+ cstr += nbytes;
+
+ /* Determine the width of this character and add it to the total. */
+ w = utf8proc_charwidth(ucs);
+ if (w == -1)
+ return -1;
+ width += w;
+ }
+
+ return width;
+}
Index: subversion/libsvn_subr/utf_width.c
===================================================================
--- subversion/libsvn_subr/utf_width.c (revision 1934258)
+++ subversion/libsvn_subr/utf_width.c (nonexistent)
@@ -1,283 +0,0 @@
-/*
- * This is an implementation of wcwidth() and wcswidth() (defined in
- * IEEE Std 1002.1-2001) for Unicode.
- *
- * http://www.opengroup.org/onlinepubs/007904975/functions/wcwidth.html
- * http://www.opengroup.org/onlinepubs/007904975/functions/wcswidth.html
- *
- * In fixed-width output devices, Latin characters all occupy a single
- * "cell" position of equal width, whereas ideographic CJK characters
- * occupy two such cells. Interoperability between terminal-line
- * applications and (teletype-style) character terminals using the
- * UTF-8 encoding requires agreement on which character should advance
- * the cursor by how many cell positions. No established formal
- * standards exist at present on which Unicode character shall occupy
- * how many cell positions on character terminals. These routines are
- * a first attempt of defining such behavior based on simple rules
- * applied to data provided by the Unicode Consortium.
- *
- * For some graphical characters, the Unicode standard explicitly
- * defines a character-cell width via the definition of the East Asian
- * FullWidth (F), Wide (W), Half-width (H), and Narrow (Na) classes.
- * In all these cases, there is no ambiguity about which width a
- * terminal shall use. For characters in the East Asian Ambiguous (A)
- * class, the width choice depends purely on a preference of backward
- * compatibility with either historic CJK or Western practice.
- * Choosing single-width for these characters is easy to justify as
- * the appropriate long-term solution, as the CJK practice of
- * displaying these characters as double-width comes from historic
- * implementation simplicity (8-bit encoded characters were displayed
- * single-width and 16-bit ones double-width, even for Greek,
- * Cyrillic, etc.) and not any typographic considerations.
- *
- * Much less clear is the choice of width for the Not East Asian
- * (Neutral) class. Existing practice does not dictate a width for any
- * of these characters. It would nevertheless make sense
- * typographically to allocate two character cells to characters such
- * as for instance EM SPACE or VOLUME INTEGRAL, which cannot be
- * represented adequately with a single-width glyph. The following
- * routines at present merely assign a single-cell width to all
- * neutral characters, in the interest of simplicity. This is not
- * entirely satisfactory and should be reconsidered before
- * establishing a formal standard in this area. At the moment, the
- * decision which Not East Asian (Neutral) characters should be
- * represented by double-width glyphs cannot yet be answered by
- * applying a simple rule from the Unicode database content. Setting
- * up a proper standard for the behavior of UTF-8 character terminals
- * will require a careful analysis not only of each Unicode character,
- * but also of each presentation form, something the author of these
- * routines has avoided to do so far.
- *
- * http://www.unicode.org/unicode/reports/tr11/
- *
- * Markus Kuhn -- 2007-05-26 (Unicode 5.0)
- *
- * Permission to use, copy, modify, and distribute this software
- * for any purpose and without fee is hereby granted. The author
- * disclaims all warranties with regard to this software.
- *
- * Latest version: http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c
- */
-
-#include <apr_lib.h>
-
-#include "svn_utf.h"
-#include "private/svn_utf_private.h"
-
-#include "svn_private_config.h"
-
-struct interval {
- apr_uint32_t first;
- apr_uint32_t last;
-};
-
-/* auxiliary function for binary search in interval table */
-static int
-bisearch(apr_uint32_t ucs, const struct interval *table, apr_uint32_t max)
-{
- apr_uint32_t min = 0;
- apr_uint32_t mid;
-
- if (ucs < table[0].first || ucs > table[max].last)
- return 0;
- while (max >= min) {
- mid = (min + max) / 2;
- if (ucs > table[mid].last)
- min = mid + 1;
- else if (ucs < table[mid].first)
- max = mid - 1; /* this is safe because ucs >= table[0].first */
- else
- return 1;
- }
-
- return 0;
-}
-
-
-/* The following two functions define the column width of an ISO 10646
- * character as follows:
- *
- * - The null character (U+0000) has a column width of 0.
- *
- * - Other C0/C1 control characters and DEL will lead to a return
- * value of -1.
- *
- * - Non-spacing and enclosing combining characters (general
- * category code Mn or Me in the Unicode database) have a
- * column width of 0.
- *
- * - SOFT HYPHEN (U+00AD) has a column width of 1.
- *
- * - Other format characters (general category code Cf in the Unicode
- * database) and ZERO WIDTH SPACE (U+200B) have a column width of 0.
- *
- * - Hangul Jamo medial vowels and final consonants (U+1160-U+11FF)
- * have a column width of 0.
- *
- * - Spacing characters in the East Asian Wide (W) or East Asian
- * Full-width (F) category as defined in Unicode Technical
- * Report #11 have a column width of 2.
- *
- * - All remaining characters (including all printable
- * ISO 8859-1 and WGL4 characters, Unicode control characters,
- * etc.) have a column width of 1.
- *
- * This implementation assumes that wchar_t characters are encoded
- * in ISO 10646.
- */
-
-static int
-mk_wcwidth(apr_uint32_t ucs)
-{
- /* sorted list of non-overlapping intervals of non-spacing characters */
- /* generated by "uniset +cat=Me +cat=Mn +cat=Cf -00AD +1160-11FF +200B c" */
- static const struct interval combining[] = {
- { 0x0300, 0x036F }, { 0x0483, 0x0486 }, { 0x0488, 0x0489 },
- { 0x0591, 0x05BD }, { 0x05BF, 0x05BF }, { 0x05C1, 0x05C2 },
- { 0x05C4, 0x05C5 }, { 0x05C7, 0x05C7 }, { 0x0600, 0x0603 },
- { 0x0610, 0x0615 }, { 0x064B, 0x065E }, { 0x0670, 0x0670 },
- { 0x06D6, 0x06E4 }, { 0x06E7, 0x06E8 }, { 0x06EA, 0x06ED },
- { 0x070F, 0x070F }, { 0x0711, 0x0711 }, { 0x0730, 0x074A },
- { 0x07A6, 0x07B0 }, { 0x07EB, 0x07F3 }, { 0x0901, 0x0902 },
- { 0x093C, 0x093C }, { 0x0941, 0x0948 }, { 0x094D, 0x094D },
- { 0x0951, 0x0954 }, { 0x0962, 0x0963 }, { 0x0981, 0x0981 },
- { 0x09BC, 0x09BC }, { 0x09C1, 0x09C4 }, { 0x09CD, 0x09CD },
- { 0x09E2, 0x09E3 }, { 0x0A01, 0x0A02 }, { 0x0A3C, 0x0A3C },
- { 0x0A41, 0x0A42 }, { 0x0A47, 0x0A48 }, { 0x0A4B, 0x0A4D },
- { 0x0A70, 0x0A71 }, { 0x0A81, 0x0A82 }, { 0x0ABC, 0x0ABC },
- { 0x0AC1, 0x0AC5 }, { 0x0AC7, 0x0AC8 }, { 0x0ACD, 0x0ACD },
- { 0x0AE2, 0x0AE3 }, { 0x0B01, 0x0B01 }, { 0x0B3C, 0x0B3C },
- { 0x0B3F, 0x0B3F }, { 0x0B41, 0x0B43 }, { 0x0B4D, 0x0B4D },
- { 0x0B56, 0x0B56 }, { 0x0B82, 0x0B82 }, { 0x0BC0, 0x0BC0 },
- { 0x0BCD, 0x0BCD }, { 0x0C3E, 0x0C40 }, { 0x0C46, 0x0C48 },
- { 0x0C4A, 0x0C4D }, { 0x0C55, 0x0C56 }, { 0x0CBC, 0x0CBC },
- { 0x0CBF, 0x0CBF }, { 0x0CC6, 0x0CC6 }, { 0x0CCC, 0x0CCD },
- { 0x0CE2, 0x0CE3 }, { 0x0D41, 0x0D43 }, { 0x0D4D, 0x0D4D },
- { 0x0DCA, 0x0DCA }, { 0x0DD2, 0x0DD4 }, { 0x0DD6, 0x0DD6 },
- { 0x0E31, 0x0E31 }, { 0x0E34, 0x0E3A }, { 0x0E47, 0x0E4E },
- { 0x0EB1, 0x0EB1 }, { 0x0EB4, 0x0EB9 }, { 0x0EBB, 0x0EBC },
- { 0x0EC8, 0x0ECD }, { 0x0F18, 0x0F19 }, { 0x0F35, 0x0F35 },
- { 0x0F37, 0x0F37 }, { 0x0F39, 0x0F39 }, { 0x0F71, 0x0F7E },
- { 0x0F80, 0x0F84 }, { 0x0F86, 0x0F87 }, { 0x0F90, 0x0F97 },
- { 0x0F99, 0x0FBC }, { 0x0FC6, 0x0FC6 }, { 0x102D, 0x1030 },
- { 0x1032, 0x1032 }, { 0x1036, 0x1037 }, { 0x1039, 0x1039 },
- { 0x1058, 0x1059 }, { 0x1160, 0x11FF }, { 0x135F, 0x135F },
- { 0x1712, 0x1714 }, { 0x1732, 0x1734 }, { 0x1752, 0x1753 },
- { 0x1772, 0x1773 }, { 0x17B4, 0x17B5 }, { 0x17B7, 0x17BD },
- { 0x17C6, 0x17C6 }, { 0x17C9, 0x17D3 }, { 0x17DD, 0x17DD },
- { 0x180B, 0x180D }, { 0x18A9, 0x18A9 }, { 0x1920, 0x1922 },
- { 0x1927, 0x1928 }, { 0x1932, 0x1932 }, { 0x1939, 0x193B },
- { 0x1A17, 0x1A18 }, { 0x1B00, 0x1B03 }, { 0x1B34, 0x1B34 },
- { 0x1B36, 0x1B3A }, { 0x1B3C, 0x1B3C }, { 0x1B42, 0x1B42 },
- { 0x1B6B, 0x1B73 }, { 0x1DC0, 0x1DCA }, { 0x1DFE, 0x1DFF },
- { 0x200B, 0x200F }, { 0x202A, 0x202E }, { 0x2060, 0x2063 },
- { 0x206A, 0x206F }, { 0x20D0, 0x20EF }, { 0x302A, 0x302F },
- { 0x3099, 0x309A }, { 0xA806, 0xA806 }, { 0xA80B, 0xA80B },
- { 0xA825, 0xA826 }, { 0xFB1E, 0xFB1E }, { 0xFE00, 0xFE0F },
- { 0xFE20, 0xFE23 }, { 0xFEFF, 0xFEFF }, { 0xFFF9, 0xFFFB },
- { 0x10A01, 0x10A03 }, { 0x10A05, 0x10A06 }, { 0x10A0C, 0x10A0F },
- { 0x10A38, 0x10A3A }, { 0x10A3F, 0x10A3F }, { 0x1D167, 0x1D169 },
- { 0x1D173, 0x1D182 }, { 0x1D185, 0x1D18B }, { 0x1D1AA, 0x1D1AD },
- { 0x1D242, 0x1D244 }, { 0xE0001, 0xE0001 }, { 0xE0020, 0xE007F },
- { 0xE0100, 0xE01EF }
- };
-
- /* test for 8-bit control characters */
- if (ucs == 0)
- return 0;
- if (ucs < 32 || (ucs >= 0x7f && ucs < 0xa0))
- return -1;
-
- /* binary search in table of non-spacing characters */
- if (bisearch(ucs, combining,
- sizeof(combining) / sizeof(struct interval) - 1))
- return 0;
-
- /* if we arrive here, ucs is not a combining or C0/C1 control character */
-
- return 1 +
- (ucs >= 0x1100 &&
- (ucs <= 0x115f || /* Hangul Jamo init. consonants */
- ucs == 0x2329 || ucs == 0x232a ||
- (ucs >= 0x2e80 && ucs <= 0xa4cf &&
- ucs != 0x303f) || /* CJK ... Yi */
- (ucs >= 0xac00 && ucs <= 0xd7a3) || /* Hangul Syllables */
- (ucs >= 0xf900 && ucs <= 0xfaff) || /* CJK Compatibility Ideographs */
- (ucs >= 0xfe10 && ucs <= 0xfe19) || /* Vertical forms */
- (ucs >= 0xfe30 && ucs <= 0xfe6f) || /* CJK Compatibility Forms */
- (ucs >= 0xff00 && ucs <= 0xff60) || /* Fullwidth Forms */
- (ucs >= 0xffe0 && ucs <= 0xffe6) ||
- (ucs >= 0x20000 && ucs <= 0x2fffd) ||
- (ucs >= 0x30000 && ucs <= 0x3fffd)));
-}
-
-int
-svn_utf_cstring_utf8_width(const char *cstr)
-{
- int width = 0;
-
- if (*cstr == '\0')
- return 0;
-
- /* Ensure the conversion below doesn't fail because of encoding errors. */
- if (!svn_utf__cstring_is_valid(cstr))
- return -1;
-
- /* Convert the UTF-8 string to UTF-32 (UCS4) which is the format
- * mk_wcwidth() expects, and get the width of each character.
- * We don't need much error checking since the input is valid UTF-8. */
- while (*cstr)
- {
- apr_uint32_t ucs;
- int nbytes;
- int lead_mask;
- int w;
- int i;
-
- if ((*cstr & 0x80) == 0)
- {
- nbytes = 1;
- lead_mask = 0x7f;
- }
- else if ((*cstr & 0xe0) == 0xc0)
- {
- nbytes = 2;
- lead_mask = 0x1f;
- }
- else if ((*cstr & 0xf0) == 0xe0)
- {
- nbytes = 3;
- lead_mask = 0x0f;
- }
- else if ((*cstr & 0xf8) == 0xf0)
- {
- nbytes = 4;
- lead_mask = 0x07;
- }
- else
- {
- /* RFC 3629 restricts UTF-8 to max 4 bytes per character. */
- return -1;
- }
-
- /* Parse character data from leading byte. */
- ucs = (apr_uint32_t)(*cstr & lead_mask);
-
- /* Parse character data from continuation bytes. */
- for (i = 1; i < nbytes; i++)
- {
- ucs <<= 6;
- ucs |= (cstr[i] & 0x3f);
- }
-
- cstr += nbytes;
-
- /* Determine the width of this character and add it to the total. */
- w = mk_wcwidth(ucs);
- if (w == -1)
- return -1;
- width += w;
- }
-
- return width;
-}
Property changes on: subversion/libsvn_subr/utf_width.c
___________________________________________________________________
Deleted: svn:eol-style
## -1 +0,0 ##
-native
\ No newline at end of property
Index: subversion/tests/libsvn_subr/utf-test.c
===================================================================
--- subversion/tests/libsvn_subr/utf-test.c (revision 1934258)
+++ subversion/tests/libsvn_subr/utf-test.c (working copy)
@@ -998,6 +998,32 @@ test_utf_xfrm(apr_pool_t *pool)
return SVN_NO_ERROR;
}
+static svn_error_t *
+test_utf8_width(apr_pool_t *pool)
+{
+ /* there are three emojis that each have wcwidth of two */
+ const char *fat_emojis = "\xf0\x9f\xa5\xba\xf0\x9f\x91\x89\xf0\x9f\x91\x88";
+ const char *mixup =
+ "S\xcc\x87\xcc\xa3" /* S with dot above and below */
+ "\xc5\xaf" /* u with ring */
+ "b\xcc\xb1" /* b with macron below */
+ "\xe1\xb9\xbd" /* v with tilde */
+ "e\xcc\xa7\xcc\x86" /* e with breve and cedilla */
+ "\xc8\x91" /* r with double grave */
+ "s\xcc\x8c" /* s with caron */
+ "\xe1\xb8\xaf" /* i with diaeresis and acute */
+ "o\xcc\x80\xcc\x9b" /* o with grave and hook */
+ "\xe1\xb9\x8b"; /* n with circumflex below */
+ const char *invalid = "a" "\xe6" "bc";
+
+ SVN_TEST_INT_ASSERT(svn_utf_cstring_utf8_width("abc123"), 6);
+ SVN_TEST_INT_ASSERT(svn_utf_cstring_utf8_width(fat_emojis), 6);
+ SVN_TEST_INT_ASSERT(svn_utf_cstring_utf8_width(mixup), 10);
+ SVN_TEST_INT_ASSERT(svn_utf_cstring_utf8_width(invalid), -1);
+
+ return SVN_NO_ERROR;
+}
+
/* The test table. */
@@ -1028,6 +1054,8 @@ static struct svn_test_descriptor_t test_funcs[] =
"test svn_utf__normalize"),
SVN_TEST_PASS2(test_utf_xfrm,
"test svn_utf__xfrm"),
+ SVN_TEST_PASS2(test_utf8_width,
+ "test svn_utf_cstring_utf8_width"),
SVN_TEST_NULL
};
