From e6ddd22060e49ddd8382069ef03fcc3d2618af13 Mon Sep 17 00:00:00 2001
From: Thomas Voss <mail@thomasvoss.com>
Date: Sat, 13 Apr 2024 11:42:12 +0200
Subject: Add u8upper()

---
 include/unicode/string.h     |  9 +++++++++
 lib/unicode/string/u8upper.c | 36 ++++++++++++++++++++++++++++++++++++
 2 files changed, 45 insertions(+)
 create mode 100644 lib/unicode/string/u8upper.c

diff --git a/include/unicode/string.h b/include/unicode/string.h
index c2a99c1..10e16ae 100644
--- a/include/unicode/string.h
+++ b/include/unicode/string.h
@@ -6,7 +6,16 @@
 #include "__charN_t.h"
 #include "__u8view.h"
 
+enum [[clang::__flag_enum__]] upper_flags {
+	UF_LANG_AZ = 1 << 0,
+	UF_LANG_TR = 1 << 0,
+	UF_LANG_LT = 1 << 1,
+	UF_ẞ       = 1 << 2,
+};
+
 size_t u8glen(const char8_t *, size_t);
 size_t u8gnext(struct u8view *, const char8_t **, size_t *);
 
+size_t u8upper(char8_t *, size_t, const char8_t *, size_t, enum upper_flags);
+
 #endif /* !MLIB_UNICODE_STRING_H */
diff --git a/lib/unicode/string/u8upper.c b/lib/unicode/string/u8upper.c
new file mode 100644
index 0000000..df4a086
--- /dev/null
+++ b/lib/unicode/string/u8upper.c
@@ -0,0 +1,36 @@
+#include "mbstring.h"
+#include "unicode/prop.h"
+#include "unicode/string.h"
+
+size_t
+u8upper(char8_t *dst, size_t dstn, const char8_t *src, size_t srcn,
+        enum upper_flags flags)
+{
+	struct ucctx ctx = {
+		.az_or_tr = flags & UF_LANG_AZ,
+		.cap_eszett = flags & UF_ẞ,
+	};
+
+	rune ch;
+	size_t n = 0;
+	bool prev_was_i = false;
+
+	while (u8next(&ch, &src, &srcn)) {
+		if (ch == 0x307 && prev_was_i && (flags & UF_LANG_LT))
+			ctx.lt_after_i = true;
+
+		struct rview rv = uprop_get_uc(ch, ctx);
+		for (size_t i = 0; i < rv.len; i++) {
+			if (n >= dstn) {
+				char8_t buf[U8_LEN_MAX];
+				n += rtou8(buf, sizeof(buf), rv.p[i]);
+			} else
+				n += rtou8(dst + n, dstn - n, rv.p[i]);
+		}
+
+		prev_was_i = ch == 'i';
+		ctx.lt_after_i = false;
+	}
+
+	return n;
+}
-- 
cgit v1.2.3