Skip to content

Commit

Permalink
Fix glk_buffer_to_title_case_uni() issue
Browse files Browse the repository at this point in the history 
Fixes a long-standing bug with glk_buffer_to_title_case_uni() where the
results would be incorrect for characters that expand into more than one
character in titlecase. We hardcode a list of titlecase exceptions from
Unicode, and otherwise continue to use the GLib 1-to-1 mapping from
g_unichar_totitle().

Closes: #28
  • Loading branch information
@ptomato
ptomato committed Sep 24, 2023
1 parent 78d24e0 commit a68fe96
Showing 1 changed file with 248 additions and 14 deletions.
262 changes: 248 additions & 14 deletions libchimara/case.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -183,20 +183,254 @@ glk_buffer_to_title_case_uni(glui32 *buf, glui32 len, glui32 numchars, glui32 lo
{
g_return_val_if_fail(buf != NULL && (len > 0 || numchars > 0), 0);
g_return_val_if_fail(numchars <= len, 0);

/* FIXME: This is wrong. g_unichar_totitle() which returns the titlecase of
one Unicode code point, but that only returns the correct result if the
titlecase character is also one code point.
For example, the one-character 'ffi' ligature should be title-cased to the
three-character string 'Ffi'. This code leaves it as the 'ffi' ligature,
which is incorrect.
However, nothing much can be done about it unless GLib gets a
g_utf8_strtitle() function, or we roll our own. */
*buf = g_unichar_totitle(*buf);
/* Call lowercase on the rest of the string */
if(lowerrest)
return glk_buffer_to_lower_case_uni(buf + 1, len - 1, numchars - 1) + 1;
return numchars;

/* Special-cases not handled by g_unichar_totitle() because they expand to
* more than one character in title case.
* Source: https://www.unicode.org/Public/UCD/latest/ucd/SpecialCasing.txt
* Does not include any context- or locale-sensitive case mappings. */
unsigned tlen = 0;
glui32 titled[3];
switch (buf[0]) {
case 0x00df: /* LATIN SMALL LETTER SHARP S */
titled[tlen++] = 0x0053;
titled[tlen++] = 0x0073;
break;
case 0xfb00: /* LATIN SMALL LIGATURE FF */
titled[tlen++] = 0x0046;
titled[tlen++] = 0x0066;
break;
case 0xfb01: /* LATIN SMALL LIGATURE FI */
titled[tlen++] = 0x0046;
titled[tlen++] = 0x0069;
break;
case 0xfb02: /* LATIN SMALL LIGATURE FL */
titled[tlen++] = 0x0046;
titled[tlen++] = 0x006c;
break;
case 0xfb03: /* LATIN SMALL LIGATURE FFI */
titled[tlen++] = 0x0046;
titled[tlen++] = 0x0066;
titled[tlen++] = 0x0069;
break;
case 0xfb04: /* LATIN SMALL LIGATURE FFL */
titled[tlen++] = 0x0046;
titled[tlen++] = 0x0066;
titled[tlen++] = 0x006c;
break;
case 0xfb05: /* LATIN SMALL LIGATURE LONG S T */
titled[tlen++] = 0x0053;
titled[tlen++] = 0x0074;
break;
case 0xfb06: /* LATIN SMALL LIGATURE ST */
titled[tlen++] = 0x0053;
titled[tlen++] = 0x0074;
break;
case 0x0587: /* ARMENIAN SMALL LIGATURE ECH YIWN */
titled[tlen++] = 0x0535;
titled[tlen++] = 0x0582;
break;
case 0xfb13: /* ARMENIAN SMALL LIGATURE MEN NOW */
titled[tlen++] = 0x0544;
titled[tlen++] = 0x0576;
break;
case 0xfb14: /* ARMENIAN SMALL LIGATURE MEN ECH */
titled[tlen++] = 0x0544;
titled[tlen++] = 0x0565;
break;
case 0xfb15: /* ARMENIAN SMALL LIGATURE MEN INI */
titled[tlen++] = 0x0544;
titled[tlen++] = 0x056b;
break;
case 0xfb16: /* ARMENIAN SMALL LIGATURE VEW NOW */
titled[tlen++] = 0x054e;
titled[tlen++] = 0x0576;
break;
case 0xfb17: /* ARMENIAN SMALL LIGATURE MEN XEH */
titled[tlen++] = 0x0544;
titled[tlen++] = 0x056d;
break;
case 0x0149: /* LATIN SMALL LETTER N PRECEDED BY APOSTROPHE */
titled[tlen++] = 0x02bc;
titled[tlen++] = 0x004e;
break;
case 0x0390: /* GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TONOS */
titled[tlen++] = 0x0399;
titled[tlen++] = 0x0308;
titled[tlen++] = 0x0301;
break;
case 0x03b0: /* GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND TONOS */
titled[tlen++] = 0x03a5;
titled[tlen++] = 0x0308;
titled[tlen++] = 0x0301;
break;
case 0x01f0: /* LATIN SMALL LETTER J WITH CARON */
titled[tlen++] = 0x004a;
titled[tlen++] = 0x030c;
break;
case 0x1e96: /* LATIN SMALL LETTER H WITH LINE BELOW */
titled[tlen++] = 0x0048;
titled[tlen++] = 0x0331;
break;
case 0x1e97: /* LATIN SMALL LETTER T WITH DIAERESIS */
titled[tlen++] = 0x0054;
titled[tlen++] = 0x0308;
break;
case 0x1e98: /* LATIN SMALL LETTER W WITH RING ABOVE */
titled[tlen++] = 0x0057;
titled[tlen++] = 0x030a;
break;
case 0x1e99: /* LATIN SMALL LETTER Y WITH RING ABOVE */
titled[tlen++] = 0x0059;
titled[tlen++] = 0x030a;
break;
case 0x1e9a: /* LATIN SMALL LETTER A WITH RIGHT HALF RING */
titled[tlen++] = 0x0041;
titled[tlen++] = 0x02be;
break;
case 0x1f50: /* GREEK SMALL LETTER UPSILON WITH PSILI */
titled[tlen++] = 0x03a5;
titled[tlen++] = 0x0313;
break;
case 0x1f52: /* GREEK SMALL LETTER UPSILON WITH PSILI AND VARIA */
titled[tlen++] = 0x03a5;
titled[tlen++] = 0x0313;
titled[tlen++] = 0x0300;
break;
case 0x1f54: /* GREEK SMALL LETTER UPSILON WITH PSILI AND OXIA */
titled[tlen++] = 0x03a5;
titled[tlen++] = 0x0313;
titled[tlen++] = 0x0301;
break;
case 0x1f56: /* GREEK SMALL LETTER UPSILON WITH PSILI AND PERISPOMENI */
titled[tlen++] = 0x03a5;
titled[tlen++] = 0x0313;
titled[tlen++] = 0x0342;
break;
case 0x1fb6: /* GREEK SMALL LETTER ALPHA WITH PERISPOMENI */
titled[tlen++] = 0x0391;
titled[tlen++] = 0x0342;
break;
case 0x1fc6: /* GREEK SMALL LETTER ETA WITH PERISPOMENI */
titled[tlen++] = 0x0397;
titled[tlen++] = 0x0342;
break;
case 0x1fd2: /* GREEK SMALL LETTER IOTA WITH DIALYTIKA AND VARIA */
titled[tlen++] = 0x0399;
titled[tlen++] = 0x0308;
titled[tlen++] = 0x0300;
break;
case 0x1fd3: /* GREEK SMALL LETTER IOTA WITH DIALYTIKA AND OXIA */
titled[tlen++] = 0x0399;
titled[tlen++] = 0x0308;
titled[tlen++] = 0x0301;
break;
case 0x1fd6: /* GREEK SMALL LETTER IOTA WITH PERISPOMENI */
titled[tlen++] = 0x0399;
titled[tlen++] = 0x0342;
break;
case 0x1fd7: /* GREEK SMALL LETTER IOTA WITH DIALYTIKA AND PERISPOMENI */
titled[tlen++] = 0x0399;
titled[tlen++] = 0x0308;
titled[tlen++] = 0x0342;
break;
case 0x1fe2: /* GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND VARIA */
titled[tlen++] = 0x03a5;
titled[tlen++] = 0x0308;
titled[tlen++] = 0x0300;
break;
case 0x1fe3: /* GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND OXIA */
titled[tlen++] = 0x03a5;
titled[tlen++] = 0x0308;
titled[tlen++] = 0x0301;
break;
case 0x1fe4: /* GREEK SMALL LETTER RHO WITH PSILI */
titled[tlen++] = 0x03a1;
titled[tlen++] = 0x0313;
break;
case 0x1fe6: /* GREEK SMALL LETTER UPSILON WITH PERISPOMENI */
titled[tlen++] = 0x03a5;
titled[tlen++] = 0x0342;
break;
case 0x1fe7: /* GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND PERISPOMENI */
titled[tlen++] = 0x03a5;
titled[tlen++] = 0x0308;
titled[tlen++] = 0x0342;
break;
case 0x1ff6: /* GREEK SMALL LETTER OMEGA WITH PERISPOMENI */
titled[tlen++] = 0x03a9;
titled[tlen++] = 0x0342;
break;
case 0x1fb2: /* GREEK SMALL LETTER ALPHA WITH VARIA AND YPOGEGRAMMENI */
titled[tlen++] = 0x1fba;
titled[tlen++] = 0x0345;
break;
case 0x1fb4: /* GREEK SMALL LETTER ALPHA WITH OXIA AND YPOGEGRAMMENI */
titled[tlen++] = 0x0386;
titled[tlen++] = 0x0345;
break;
case 0x1fc2: /* GREEK SMALL LETTER ETA WITH VARIA AND YPOGEGRAMMENI */
titled[tlen++] = 0x1fca;
titled[tlen++] = 0x0345;
break;
case 0x1fc4: /* GREEK SMALL LETTER ETA WITH OXIA AND YPOGEGRAMMENI */
titled[tlen++] = 0x0389;
titled[tlen++] = 0x0345;
break;
case 0x1ff2: /* GREEK SMALL LETTER OMEGA WITH VARIA AND YPOGEGRAMMENI */
titled[tlen++] = 0x1ffa;
titled[tlen++] = 0x0345;
break;
case 0x1ff4: /* GREEK SMALL LETTER OMEGA WITH OXIA AND YPOGEGRAMMENI */
titled[tlen++] = 0x038f;
titled[tlen++] = 0x0345;
break;
case 0x1fb7: /* GREEK SMALL LETTER ALPHA WITH PERISPOMENI AND YPOGEGRAMMENI */
titled[tlen++] = 0x0391;
titled[tlen++] = 0x0342;
titled[tlen++] = 0x0345;
break;
case 0x1fc7: /* GREEK SMALL LETTER ETA WITH PERISPOMENI AND YPOGEGRAMMENI */
titled[tlen++] = 0x0397;
titled[tlen++] = 0x0342;
titled[tlen++] = 0x0345;
break;
case 0x1ff7: /* GREEK SMALL LETTER OMEGA WITH PERISPOMENI AND YPOGEGRAMMENI */
titled[tlen++] = 0x03a9;
titled[tlen++] = 0x0342;
titled[tlen++] = 0x0345;
break;
}

if (tlen == 0) {
/* Easy path. g_unichar_totitle() returns the titlecase of one Unicode
* code point, but that is only correct if the titlecase character is
* also one code point. */
*buf = g_unichar_totitle(*buf);
/* Call lowercase on the rest of the string */
if (lowerrest)
return glk_buffer_to_lower_case_uni(buf + 1, len - 1, numchars - 1) + 1;
return numchars;
}

/* This code handles the special cases from the above switch statement. For
* example, the one-character 'ffi' ligature should be title-cased to the
* three-character string 'Ffi'. */

/* Allocate a new buffer, because the number of characters has already
* changed. */
g_autofree glui32 *newbuf = g_new0(glui32, len + tlen);

memcpy(newbuf, titled, tlen * sizeof(glui32));
memcpy(newbuf + tlen, buf + 1, (numchars - 1) * sizeof(glui32));

if (!lowerrest) {
memcpy(buf, newbuf, len * sizeof(glui32));
return numchars + tlen - 1;
}

glui32 lowercount = glk_buffer_to_lower_case_uni(newbuf + tlen, len - tlen, numchars - 1);
memcpy(buf, newbuf, len * sizeof(glui32));
return lowercount + tlen;
}

/**
Expand Down

0 comments on commit a68fe96

Please sign in to comment.