mirror of
https://github.com/mirkobrombin/raven.git
synced 2026-11-01 21:45:06 +00:00
mirkobrombin
f5eda87880
Assisted-by: Claude Code:claude-opus-5-5 AI-Scope: Wrote NormalizeString and IsNormalizedString on UAX #15, the table generator and the test.
70 lines
2.7 KiB
C
70 lines
2.7 KiB
C
#define _WIN32_WINNT 0x0600
|
|
#define WIN32_LEAN_AND_MEAN
|
|
#include <windows.h>
|
|
#include <winnls.h>
|
|
|
|
static int same(const WCHAR *left, const WCHAR *right, int length)
|
|
{
|
|
for (int index = 0; index < length; index++)
|
|
if (left[index] != right[index])
|
|
return 0;
|
|
return 1;
|
|
}
|
|
|
|
static int check(NORM_FORM form, const WCHAR *in, const WCHAR *expected)
|
|
{
|
|
WCHAR out[32];
|
|
int wanted = lstrlenW(expected);
|
|
int length = NormalizeString(form, in, -1, out, 32);
|
|
return length == wanted + 1 && same(out, expected, length);
|
|
}
|
|
|
|
int raven_entry(void)
|
|
{
|
|
// e + combining acute, in each direction.
|
|
if (!check(NormalizationC, L"e\x0301", L"\x00e9") || !check(NormalizationD, L"\x00e9", L"e\x0301"))
|
|
return 1;
|
|
// Marks reorder by class: dot below (220) goes ahead of dot above (230), then the first composes.
|
|
if (!check(NormalizationC, L"s\x0307\x0323", L"\x1e69"))
|
|
return 2;
|
|
if (!check(NormalizationD, L"\x1e69", L"s\x0323\x0307"))
|
|
return 3;
|
|
// A blocked mark does not join: the second acute stays apart.
|
|
if (!check(NormalizationC, L"a\x0301\x0301", L"\x00e1\x0301"))
|
|
return 4;
|
|
// Hangul by arithmetic, both ways.
|
|
if (!check(NormalizationD, L"\xd55c", L"\x1112\x1161\x11ab") || !check(NormalizationC, L"\x1112\x1161\x11ab", L"\xd55c"))
|
|
return 5;
|
|
// Compatibility forms only change under KC and KD.
|
|
if (!check(NormalizationKC, L"\xfb01", L"fi") || !check(NormalizationC, L"\xfb01", L"\xfb01") || !check(NormalizationKD, L"\x2460", L"1"))
|
|
return 6;
|
|
// A composition exclusion stays decomposed, and a singleton does not come back.
|
|
if (!check(NormalizationC, L"\x0958", L"\x0915\x093c") || !check(NormalizationC, L"\x212b", L"\x00c5"))
|
|
return 7;
|
|
// Beyond the basic plane: U+1D15E has an excluded canonical decomposition.
|
|
if (!check(NormalizationC, L"\xd834\xdd5e", L"\xd834\xdd57\xd834\xdd65"))
|
|
return 8;
|
|
|
|
// Asking for the size, a counted source without terminator, a short buffer, broken text, a bad form.
|
|
WCHAR out[8];
|
|
if (NormalizeString(NormalizationC, L"e\x0301", -1, NULL, 0) < 2)
|
|
return 9;
|
|
if (NormalizeString(NormalizationC, L"e\x0301z", 2, out, 8) != 1 || out[0] != 0xe9)
|
|
return 10;
|
|
SetLastError(0);
|
|
if (NormalizeString(NormalizationD, L"\x00e9\x00e9", 2, out, 1) || GetLastError() != ERROR_INSUFFICIENT_BUFFER)
|
|
return 11;
|
|
SetLastError(0);
|
|
if (NormalizeString(NormalizationC, L"a\xd800z", -1, out, 8) > 0 || GetLastError() != ERROR_NO_UNICODE_TRANSLATION)
|
|
return 12;
|
|
SetLastError(0);
|
|
if (NormalizeString((NORM_FORM)3, L"a", -1, out, 8) || GetLastError() != ERROR_INVALID_PARAMETER)
|
|
return 13;
|
|
|
|
if (!IsNormalizedString(NormalizationC, L"\x00e9", -1) || IsNormalizedString(NormalizationC, L"e\x0301", -1))
|
|
return 14;
|
|
if (!IsNormalizedString(NormalizationD, L"e\x0301", -1) || IsNormalizedString(NormalizationD, L"\x00e9", -1))
|
|
return 15;
|
|
return 0;
|
|
}
|