[libc-commits] [libc] [libc] Implemented CharacterConverter push/pop for utf32->utf8 conversions (PR #143971)
Michael Jones via libc-commits
libc-commits at lists.llvm.org
Thu Jun 12 15:28:05 PDT 2025
================
@@ -22,13 +23,134 @@ bool CharacterConverter::isComplete() {
return state->bytes_processed == state->total_bytes;
}
-int CharacterConverter::push(char8_t utf8_byte) {}
+int CharacterConverter::push(char32_t utf32) {
+ state->partial = utf32;
+ state->bytes_processed = 0;
+ state->total_bytes = 0;
-int CharacterConverter::push(char32_t utf32) {}
+ // determine number of utf-8 bytes needed to represent this utf32 value
+ char32_t ranges[] = {0x7f, 0x7ff, 0xffff, 0x10ffff};
+ const int num_ranges = 4;
+ for (uint8_t i = 0; i < num_ranges; i++) {
+ if (state->partial <= ranges[i]) {
+ state->total_bytes = i + 1;
+ break;
+ }
+ }
+ if (state->total_bytes == 0) {
+ return -1;
+ }
-utf_ret<char8_t> CharacterConverter::pop_utf8() {}
+ return 0;
+}
+
+utf_ret<char8_t> CharacterConverter::pop_utf8_seqlength1() {
+ utf_ret<char8_t> result;
+ result.error = 0;
+
+ // 0xxxxxxx
+ switch (state->bytes_processed) {
+ case 0:
+ result.out = (char8_t)(state->partial);
+ break;
+ default:
+ result.error = -1;
+ return result;
+ }
+
+ state->bytes_processed++;
+ return result;
+}
+
+utf_ret<char8_t> CharacterConverter::pop_utf8_seqlength2() {
+ utf_ret<char8_t> result;
+ result.error = 0;
+
+ // 110xxxxx 10xxxxxx
+ char32_t utf32 = state->partial;
+ switch (state->bytes_processed) {
+ case 0:
+ result.out = (char8_t)(0xC0 | (utf32 >> 6));
+ break;
+ case 1:
+ result.out = (char8_t)(0x80 | (utf32 & 0x3f));
+ break;
+ default:
+ result.error = -1;
+ return result;
+ }
+
+ state->bytes_processed++;
+ return result;
+}
+
+utf_ret<char8_t> CharacterConverter::pop_utf8_seqlength3() {
+ utf_ret<char8_t> result;
+ result.error = 0;
-utf_ret<char32_t> CharacterConverter::pop_utf32() {}
+ // 1110xxxx 10xxxxxx 10xxxxxx
+ char32_t utf32 = state->partial;
+ switch (state->bytes_processed) {
+ case 0:
+ result.out = (char8_t)(0xE0 | (utf32 >> 12));
+ break;
+ case 1:
+ result.out = (char8_t)(0x80 | ((utf32 >> 6) & 0x3f));
+ break;
+ case 2:
+ result.out = (char8_t)(0x80 | (utf32 & 0x3f));
+ break;
+ default:
+ result.error = -1;
+ return result;
+ }
+
+ state->bytes_processed++;
+ return result;
+}
+
+utf_ret<char8_t> CharacterConverter::pop_utf8_seqlength4() {
+ utf_ret<char8_t> result;
+ result.error = 0;
+
+ // 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
+ char32_t utf32 = state->partial;
+ switch (state->bytes_processed) {
+ case 0:
+ result.out = (char8_t)(0xF0 | (utf32 >> 18));
+ break;
+ case 1:
+ result.out = (char8_t)(0x80 | ((utf32 >> 12) & 0x3f));
+ break;
+ case 2:
+ result.out = (char8_t)(0x80 | ((utf32 >> 6) & 0x3f));
+ break;
+ case 3:
+ result.out = (char8_t)(0x80 | (utf32 & 0x3f));
+ break;
+ default:
+ result.error = -1;
+ return result;
+ }
+
+ state->bytes_processed++;
+ return result;
+}
----------------
michaelrj-google wrote:
I'm pretty sure it's feasible to collapse these all down into one function. It seems like for the first char you're always shifting `(total_bytes - 1) * 6` bits, then adding `{0, 2, 3, 4}` leading 1s. For all the others you're shifting `(3 - bytes_processed) * 6` bits then using the same mask.
https://github.com/llvm/llvm-project/pull/143971
More information about the libc-commits
mailing list