Fold LIBC_UNICODE into LIBC_STR

libc/fmt/stoa.c

@@ -27,7 +27,7 @@
 #include "libc/str/thompike.h"
 #include "libc/str/tpenc.h"
 #include "libc/str/utf16.h"
-#include "libc/unicode/unicode.h"
+#include "libc/str/unicode.h"
 
 typedef int (*out_f)(const char *, void *, size_t);
 typedef int (*emit_f)(out_f, void *, uint64_t);

libc/fmt/strtoll_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/fmt/conv.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 long long strtoll_l(const char *nptr, char **endptr, int base, locale_t l) {
   return strtoll(nptr, endptr, base);

libc/fmt/strtoull_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/fmt/conv.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 unsigned long long strtoull_l(const char *nptr, char **endptr, int base,
                               locale_t l) {

libc/fmt/wcstoll_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/fmt/conv.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 long long wcstoll_l(const wchar_t *nptr, wchar_t **endptr, int base,
                     locale_t l) {

libc/fmt/wcstoull_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/fmt/conv.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 unsigned long long wcstoull_l(const wchar_t *nptr, wchar_t **endptr, int base,
                               locale_t l) {

libc/isystem/langinfo.h

@@ -1,5 +1,5 @@
 #ifndef COSMOPOLITAN_LIBC_ISYSTEM_LANGINFO_H_
 #define COSMOPOLITAN_LIBC_ISYSTEM_LANGINFO_H_
-#include "libc/unicode/langinfo.h"
+#include "libc/str/langinfo.h"
 
 #endif /* COSMOPOLITAN_LIBC_ISYSTEM_LANGINFO_H_ */

libc/isystem/locale.h

@@ -1,4 +1,4 @@
 #ifndef COSMOPOLITAN_LIBC_ISYSTEM_LOCALE_H_
 #define COSMOPOLITAN_LIBC_ISYSTEM_LOCALE_H_
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 #endif /* COSMOPOLITAN_LIBC_ISYSTEM_LOCALE_H_ */

libc/isystem/nl_types.h

@@ -1,4 +1,4 @@
 #ifndef COSMOPOLITAN_LIBC_ISYSTEM_NL_TYPES_H_
 #define COSMOPOLITAN_LIBC_ISYSTEM_NL_TYPES_H_
-#include "libc/unicode/nltypes.h"
+#include "libc/str/nltypes.h"
 #endif /* COSMOPOLITAN_LIBC_ISYSTEM_NL_TYPES_H_ */

libc/libc.mk

@@ -30,7 +30,6 @@ o/$(MODE)/libc:	o/$(MODE)/libc/calls		\
 		o/$(MODE)/libc/thread		\
 		o/$(MODE)/libc/time		\
 		o/$(MODE)/libc/tinymath		\
-		o/$(MODE)/libc/unicode		\
 		o/$(MODE)/libc/x		\
 		o/$(MODE)/libc/zipos		\
 		$(LIBC_CHECKS)

libc/log/log.mk

@@ -42,7 +42,6 @@ LIBC_LOG_A_DIRECTDEPS =					\
 	LIBC_SYSV_CALLS					\
 	LIBC_TIME					\
 	LIBC_TINYMATH					\
-	LIBC_UNICODE					\
 	THIRD_PARTY_DLMALLOC				\
 	THIRD_PARTY_GDTOA
 

libc/sock/sock.mk

@@ -44,7 +44,6 @@ LIBC_SOCK_A_DIRECTDEPS =			\
 	LIBC_STUBS				\
 	LIBC_SYSV_CALLS				\
 	LIBC_TIME				\
-	LIBC_UNICODE				\
 	LIBC_SYSV
 
 LIBC_SOCK_A_DEPS :=				\

libc/stdio/iconv.c

@@ -31,7 +31,7 @@
 #include "libc/stdio/iconv.h"
 #include "libc/str/str.h"
 #include "libc/thread/thread.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 // clang-format off
 
 asm(".ident\t\"\\n\\n\

libc/str/freelocale.c

@@ -16,7 +16,7 @@
 │ TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR             │
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 void freelocale(locale_t l) {
   // TODO: implement me

libc/str/isdigit_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int isdigit_l(int c, locale_t l) {
   return iswdigit(c);

libc/str/islower_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int islower_l(int c, locale_t l) {
   return islower(c);

libc/str/isupper_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int isupper_l(int c, locale_t l) {
   return isupper(c);

libc/str/iswalpha_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int iswalpha_l(wint_t c, locale_t l) {
   return iswalpha(c);

libc/str/iswblank_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int iswblank_l(wint_t c, locale_t l) {
   return iswblank(c);

libc/str/iswcntrl_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int iswcntrl_l(wint_t c, locale_t l) {
   return iswcntrl(c);

libc/str/iswdigit_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int iswdigit_l(wint_t c, locale_t l) {
   return iswdigit(c);

libc/str/iswlower_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int iswlower_l(wint_t c, locale_t l) {
   return iswlower(c);

libc/str/iswprint_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int iswprint_l(wint_t c, locale_t l) {
   return iswprint(c);

libc/str/iswpunct_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int iswpunct_l(wint_t c, locale_t l) {
   return iswpunct(c);

libc/str/iswspace_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int iswspace_l(wint_t c, locale_t l) {
   return iswspace(c);

libc/str/iswupper_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int iswupper_l(wint_t c, locale_t l) {
   return iswupper(c);

libc/str/iswxdigit_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int iswxdigit_l(wint_t c, locale_t l) {
   return iswxdigit(c);

libc/str/isxdigit_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int isxdigit_l(int c, locale_t l) {
   return iswxdigit(c);

libc/str/langinfo.c

@@ -27,9 +27,9 @@
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/nexgen32e/gettls.h"
 #include "libc/thread/thread.h"
-#include "libc/unicode/langinfo.h"
-#include "libc/unicode/locale.h"
-#include "libc/unicode/nltypes.h"
+#include "libc/str/langinfo.h"
+#include "libc/str/locale.h"
+#include "libc/str/nltypes.h"
 
 asm(".ident\t\"\\n\\n\
 Musl libc (MIT License)\\n\

libc/str/langinfo.h

@@ -1,5 +1,5 @@
-#ifndef COSMOPOLITAN_LIBC_UNICODE_LANGINFO_H_
-#define COSMOPOLITAN_LIBC_UNICODE_LANGINFO_H_
+#ifndef COSMOPOLITAN_LIBC_STR_LANGINFO_H_
+#define COSMOPOLITAN_LIBC_STR_LANGINFO_H_
 #if !(__ASSEMBLER__ + __LINKER__ + 0)
 COSMOPOLITAN_C_START_
 
@@ -83,4 +83,4 @@ char *nl_langinfo(int);
 
 COSMOPOLITAN_C_END_
 #endif /* !(__ASSEMBLER__ + __LINKER__ + 0) */
-#endif /* COSMOPOLITAN_LIBC_UNICODE_LANGINFO_H_ */
+#endif /* COSMOPOLITAN_LIBC_STR_LANGINFO_H_ */

libc/str/locale.h

@@ -1,5 +1,5 @@
-#ifndef COSMOPOLITAN_LIBC_UNICODE_LOCALE_H_
-#define COSMOPOLITAN_LIBC_UNICODE_LOCALE_H_
+#ifndef COSMOPOLITAN_LIBC_STR_LOCALE_H_
+#define COSMOPOLITAN_LIBC_STR_LOCALE_H_
 #include "libc/fmt/conv.h"
 #include "libc/time/struct/tm.h"
 
@@ -79,4 +79,4 @@ wint_t towupper_l(wint_t, locale_t);
 
 COSMOPOLITAN_C_END_
 #endif /* !(__ASSEMBLER__ + __LINKER__ + 0) */
-#endif /* COSMOPOLITAN_LIBC_UNICODE_LOCALE_H_ */
+#endif /* COSMOPOLITAN_LIBC_STR_LOCALE_H_ */

libc/str/localeconv.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/limits.h"
-#include "libc/unicode/unicode.h"
+#include "libc/str/unicode.h"
 
 static const struct lconv kLocaleConv = {
     .decimal_point = ".",

libc/str/newlocale.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/sysv/errfuns.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 locale_t newlocale(int catmask, const char *locale, locale_t base) {
   // TODO: implement me

libc/str/nltypes.h

@@ -1,5 +1,5 @@
-#ifndef COSMOPOLITAN_LIBC_UNICODE_NLTYPES_H_
-#define COSMOPOLITAN_LIBC_UNICODE_NLTYPES_H_
+#ifndef COSMOPOLITAN_LIBC_STR_NLTYPES_H_
+#define COSMOPOLITAN_LIBC_STR_NLTYPES_H_
 
 #define NL_SETD       1
 #define NL_CAT_LOCALE 1
@@ -16,4 +16,4 @@ int catclose(nl_catd);
 
 COSMOPOLITAN_C_END_
 #endif /* !(__ASSEMBLER__ + __LINKER__ + 0) */
-#endif /* COSMOPOLITAN_LIBC_UNICODE_NLTYPES_H_ */
+#endif /* COSMOPOLITAN_LIBC_STR_NLTYPES_H_ */

libc/str/setlocale.c

@@ -18,7 +18,7 @@
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/intrin/safemacros.internal.h"
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 /**
  * Sets program locale.

libc/str/str.mk

@@ -94,6 +94,32 @@ o/$(MODE)/libc/fmt/strsignal.greg.o: private			\
 			-ffreestanding				\
 			$(NO_MAGIC)
 
+o/$(MODE)/libc/str/eastasianwidth.bin:				\
+		libc/str/eastasianwidth.txt			\
+		o/$(MODE)/tool/decode/mkwides.com
+	@$(COMPILE) -AMKWIDES -T$@ o/$(MODE)/tool/decode/mkwides.com -o $@ $<
+o/$(MODE)/libc/str/eastasianwidth.bin.lz4:			\
+		o/$(MODE)/libc/str/eastasianwidth.bin		\
+		o/$(MODE)/third_party/lz4cli/lz4cli.com
+	@$(COMPILE) -ALZ4 -T$@ o/$(MODE)/third_party/lz4cli/lz4cli.com -q -f -9 --content-size $< $@
+o/$(MODE)/libc/str/eastasianwidth.s:				\
+		o/$(MODE)/libc/str/eastasianwidth.bin.lz4	\
+		o/$(MODE)/tool/build/lz4toasm.com
+	@$(COMPILE) -ABIN2ASM -T$@ o/$(MODE)/tool/build/lz4toasm.com -s kEastAsianWidth -o $@ $<
+
+o/$(MODE)/libc/str/combiningchars.bin:				\
+		libc/str/strdata.txt				\
+		o/$(MODE)/tool/decode/mkcombos.com
+	@$(COMPILE) -AMKCOMBOS -T$@ o/$(MODE)/tool/decode/mkcombos.com -o $@ $<
+o/$(MODE)/libc/str/combiningchars.bin.lz4:			\
+		o/$(MODE)/libc/str/combiningchars.bin		\
+		o/$(MODE)/third_party/lz4cli/lz4cli.com
+	@$(COMPILE) -ALZ4 -T$@ o/$(MODE)/third_party/lz4cli/lz4cli.com -q -f -9 --content-size $< $@
+o/$(MODE)/libc/str/combiningchars.s:				\
+		o/$(MODE)/libc/str/combiningchars.bin.lz4	\
+		o/$(MODE)/tool/build/lz4toasm.com
+	@$(COMPILE) -ABIN2ASM -T$@ o/$(MODE)/tool/build/lz4toasm.com -s kCombiningChars -o $@ $<
+
 LIBC_STR_LIBS = $(foreach x,$(LIBC_STR_ARTIFACTS),$($(x)))
 LIBC_STR_SRCS = $(foreach x,$(LIBC_STR_ARTIFACTS),$($(x)_SRCS))
 LIBC_STR_HDRS = $(foreach x,$(LIBC_STR_ARTIFACTS),$($(x)_HDRS))

libc/str/strcoll_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int strcoll_l(const char *p, const char *q, locale_t l) {
   return strcmp(p, q);

libc/str/strnwidth.c

@@ -23,7 +23,7 @@
 #include "libc/nexgen32e/bsf.h"
 #include "libc/str/str.h"
 #include "libc/str/thompike.h"
-#include "libc/unicode/unicode.h"
+#include "libc/str/unicode.h"
 
 /**
  * Returns monospace display width of UTF-8 string.

libc/str/strnwidth16.c

@@ -19,7 +19,7 @@
 #include "libc/intrin/safemacros.internal.h"
 #include "libc/str/oldutf16.internal.h"
 #include "libc/str/str.h"
-#include "libc/unicode/unicode.h"
+#include "libc/str/unicode.h"
 
 /**
  * Returns monospace display width of UTF-16 or UCS-2 string.

libc/str/strwidth.c

@@ -16,7 +16,7 @@
 │ TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR             │
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
-#include "libc/unicode/unicode.h"
+#include "libc/str/unicode.h"
 
 /**
  * Returns monospace display width of UTF-8 string.

libc/str/strwidth16.c

@@ -18,7 +18,7 @@
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/intrin/safemacros.internal.h"
 #include "libc/limits.h"
-#include "libc/unicode/unicode.h"
+#include "libc/str/unicode.h"
 
 /**
  * Returns monospace display width of UTF-16 or UCS-2 string.

libc/str/strxfrm_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 size_t strxfrm_l(char *dest, const char *src, size_t count, locale_t l) {
   return strxfrm(dest, src, count);

libc/str/tolower_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int tolower_l(int c, locale_t l) {
   return tolower(c);

libc/str/toupper_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int toupper_l(int c, locale_t l) {
   return toupper(c);

libc/str/towlower_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 wint_t towlower_l(wint_t c, locale_t l) {
   return towlower(c);

libc/str/towupper_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 wint_t towupper_l(wint_t c, locale_t l) {
   return towupper(c);

libc/str/unicode.h

@@ -1,5 +1,5 @@
-#ifndef COSMOPOLITAN_LIBC_UNICODE_UNICODE_H_
-#define COSMOPOLITAN_LIBC_UNICODE_UNICODE_H_
+#ifndef COSMOPOLITAN_LIBC_STR_UNICODE_H_
+#define COSMOPOLITAN_LIBC_STR_UNICODE_H_
 #if !(__ASSEMBLER__ + __LINKER__ + 0)
 COSMOPOLITAN_C_START_
 
@@ -41,4 +41,4 @@ struct lconv *localeconv(void);
 
 COSMOPOLITAN_C_END_
 #endif /* !(__ASSEMBLER__ + __LINKER__ + 0) */
-#endif /* COSMOPOLITAN_LIBC_UNICODE_UNICODE_H_ */
+#endif /* COSMOPOLITAN_LIBC_STR_UNICODE_H_ */

libc/str/update.sh

@@ -0,0 +1,15 @@
+#!/bin/sh
+[ -d libc/unicode ] || exit
+[ -x o//examples/curl.com ] || make -j8 o//examples/curl.com || exit
+mkdir -p o/tmp/ || exit
+
+shineget() {
+  echo $2
+  o//examples/curl.com $2 >o/tmp/$$ || exit
+  mv o/tmp/$$ $1 || exit
+}
+
+shineget libc/str/blocks.txt         https://www.unicode.org/Public/UCD/latest/ucd/Blocks.txt
+shineget libc/str/unicodedata.txt    https://www.unicode.org/Public/UCD/latest/ucd/UnicodeData.txt
+shineget libc/str/eastasianwidth.txt https://www.unicode.org/Public/UCD/latest/ucd/EastAsianWidth.txt
+shineget libc/str/SpecialCasing.txt  https://www.unicode.org/Public/UCD/latest/ucd/SpecialCasing.txt

libc/str/uselocale.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/sysv/errfuns.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 locale_t uselocale(locale_t l) {
   // TODO: implement me!

libc/str/wcscoll_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 int wcscoll_l(const wchar_t *p, const wchar_t *q, locale_t l) {
   return wcscmp(p, q);

libc/str/wcsnwidth.c

@@ -16,7 +16,7 @@
 │ TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR             │
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
-#include "libc/unicode/unicode.h"
+#include "libc/str/unicode.h"
 
 /**
  * Returns monospace display width of wide character string.

libc/str/wcswidth.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/limits.h"
-#include "libc/unicode/unicode.h"
+#include "libc/str/unicode.h"
 
 /**
  * Returns monospace display width of wide character string.

libc/str/wcsxfrm_l.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 size_t wcsxfrm_l(wchar_t *dest, const wchar_t *src, size_t count, locale_t l) {
   return wcsxfrm(dest, src, count);

libc/str/wcwidth.c

@@ -17,7 +17,7 @@
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/intrin/likely.h"
-#include "libc/unicode/unicode.h"
+#include "libc/str/unicode.h"
 
 extern const uint8_t kEastAsianWidth[];
 extern const uint32_t kEastAsianWidthBits;

libc/testlib/testlib.mk

@@ -108,7 +108,6 @@ LIBC_TESTLIB_A_DIRECTDEPS =					\
 	LIBC_SYSV_CALLS						\
 	LIBC_TIME						\
 	LIBC_TINYMATH						\
-	LIBC_UNICODE						\
 	LIBC_X							\
 	LIBC_ZIPOS						\
 	THIRD_PARTY_DLMALLOC					\

libc/thread/thread.h

@@ -3,7 +3,7 @@
 #include "libc/calls/struct/timespec.h"
 #include "libc/intrin/pthread.h"
 #include "libc/runtime/runtime.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 
 #define CTHREAD_CREATE_DETACHED 1
 #define CTHREAD_CREATE_JOINABLE 0

libc/time/strftime.c

@@ -21,7 +21,7 @@
 #include "libc/stdio/stdio.h"
 #include "libc/time/time.h"
 #include "libc/time/tz.internal.h"
-#include "libc/unicode/locale.h"
+#include "libc/str/locale.h"
 // clang-format off
 // converts broken-down timestamp to string
 

libc/unicode/unicode.mk

@@ -1,82 +0,0 @@
-#-*-mode:makefile-gmake;indent-tabs-mode:t;tab-width:8;coding:utf-8-*-┐
-#───vi: set et ft=make ts=8 tw=8 fenc=utf-8 :vi───────────────────────┘
-
-PKGS += LIBC_UNICODE
-
-LIBC_UNICODE_ARTIFACTS += LIBC_UNICODE_A
-LIBC_UNICODE = $(LIBC_UNICODE_A_DEPS) $(LIBC_UNICODE_A)
-LIBC_UNICODE_A = o/$(MODE)/libc/unicode/unicode.a
-LIBC_UNICODE_A_FILES := $(wildcard libc/unicode/*)
-LIBC_UNICODE_A_HDRS = $(filter %.h,$(LIBC_UNICODE_A_FILES))
-LIBC_UNICODE_A_SRCS_A = $(filter %.s,$(LIBC_UNICODE_A_FILES))
-LIBC_UNICODE_A_SRCS_S = $(filter %.S,$(LIBC_UNICODE_A_FILES))
-LIBC_UNICODE_A_SRCS_C = $(filter %.c,$(LIBC_UNICODE_A_FILES))
-
-LIBC_UNICODE_A_SRCS =						\
-	$(LIBC_UNICODE_A_SRCS_A)				\
-	$(LIBC_UNICODE_A_SRCS_S)				\
-	$(LIBC_UNICODE_A_SRCS_C)
-
-LIBC_UNICODE_A_OBJS =						\
-	$(LIBC_UNICODE_A_SRCS_A:%.s=o/$(MODE)/%.o)		\
-	$(LIBC_UNICODE_A_SRCS_S:%.S=o/$(MODE)/%.o)		\
-	$(LIBC_UNICODE_A_SRCS_C:%.c=o/$(MODE)/%.o)
-
-LIBC_UNICODE_A_CHECKS =						\
-	$(LIBC_UNICODE_A).pkg					\
-	$(LIBC_UNICODE_A_HDRS:%=o/$(MODE)/%.ok)
-
-LIBC_UNICODE_A_DIRECTDEPS =					\
-	LIBC_INTRIN						\
-	LIBC_NEXGEN32E						\
-	LIBC_STR						\
-	LIBC_STUBS
-
-LIBC_UNICODE_A_DEPS :=						\
-	$(call uniq,$(foreach x,$(LIBC_UNICODE_A_DIRECTDEPS),$($(x))))
-
-$(LIBC_UNICODE_A):						\
-		libc/unicode/					\
-		$(LIBC_UNICODE_A).pkg				\
-		$(LIBC_UNICODE_A_OBJS)
-
-$(LIBC_UNICODE_A).pkg:						\
-		$(LIBC_UNICODE_A_OBJS)				\
-		$(foreach x,$(LIBC_UNICODE_A_DIRECTDEPS),$($(x)_A).pkg)
-
-LIBC_UNICODE_LIBS = $(foreach x,$(LIBC_UNICODE_ARTIFACTS),$($(x)))
-LIBC_UNICODE_SRCS = $(foreach x,$(LIBC_UNICODE_ARTIFACTS),$($(x)_SRCS))
-LIBC_UNICODE_HDRS = $(foreach x,$(LIBC_UNICODE_ARTIFACTS),$($(x)_HDRS))
-LIBC_UNICODE_BINS = $(foreach x,$(LIBC_UNICODE_ARTIFACTS),$($(x)_BINS))
-LIBC_UNICODE_CHECKS = $(foreach x,$(LIBC_UNICODE_ARTIFACTS),$($(x)_CHECKS))
-LIBC_UNICODE_OBJS = $(foreach x,$(LIBC_UNICODE_ARTIFACTS),$($(x)_OBJS))
-LIBC_UNICODE_TESTS = $(foreach x,$(LIBC_UNICODE_ARTIFACTS),$($(x)_TESTS))
-
-.PHONY: o/$(MODE)/libc/unicode
-o/$(MODE)/libc/unicode: $(LIBC_UNICODE) $(LIBC_UNICODE_CHECKS)
-
-o/$(MODE)/libc/unicode/eastasianwidth.bin:			\
-		libc/unicode/eastasianwidth.txt			\
-		o/$(MODE)/tool/decode/mkwides.com
-	@$(COMPILE) -AMKWIDES -T$@ o/$(MODE)/tool/decode/mkwides.com -o $@ $<
-o/$(MODE)/libc/unicode/eastasianwidth.bin.lz4:			\
-		o/$(MODE)/libc/unicode/eastasianwidth.bin	\
-		o/$(MODE)/third_party/lz4cli/lz4cli.com
-	@$(COMPILE) -ALZ4 -T$@ o/$(MODE)/third_party/lz4cli/lz4cli.com -q -f -9 --content-size $< $@
-o/$(MODE)/libc/unicode/eastasianwidth.s:			\
-		o/$(MODE)/libc/unicode/eastasianwidth.bin.lz4	\
-		o/$(MODE)/tool/build/lz4toasm.com
-	@$(COMPILE) -ABIN2ASM -T$@ o/$(MODE)/tool/build/lz4toasm.com -s kEastAsianWidth -o $@ $<
-
-o/$(MODE)/libc/unicode/combiningchars.bin:			\
-		libc/unicode/unicodedata.txt			\
-		o/$(MODE)/tool/decode/mkcombos.com
-	@$(COMPILE) -AMKCOMBOS -T$@ o/$(MODE)/tool/decode/mkcombos.com -o $@ $<
-o/$(MODE)/libc/unicode/combiningchars.bin.lz4:			\
-		o/$(MODE)/libc/unicode/combiningchars.bin	\
-		o/$(MODE)/third_party/lz4cli/lz4cli.com
-	@$(COMPILE) -ALZ4 -T$@ o/$(MODE)/third_party/lz4cli/lz4cli.com -q -f -9 --content-size $< $@
-o/$(MODE)/libc/unicode/combiningchars.s:			\
-		o/$(MODE)/libc/unicode/combiningchars.bin.lz4	\
-		o/$(MODE)/tool/build/lz4toasm.com
-	@$(COMPILE) -ABIN2ASM -T$@ o/$(MODE)/tool/build/lz4toasm.com -s kCombiningChars -o $@ $<

libc/unicode/update.sh

@@ -1,15 +0,0 @@
-#!/bin/sh
-[ -d libc/unicode ] || exit
-[ -x o//examples/curl.com ] || make -j8 o//examples/curl.com || exit
-mkdir -p o/tmp/ || exit
-
-shineget() {
-  echo $2
-  o//examples/curl.com $2 >o/tmp/$$ || exit
-  mv o/tmp/$$ $1 || exit
-}
-
-shineget libc/unicode/blocks.txt         https://www.unicode.org/Public/UCD/latest/ucd/Blocks.txt
-shineget libc/unicode/unicodedata.txt    https://www.unicode.org/Public/UCD/latest/ucd/UnicodeData.txt
-shineget libc/unicode/eastasianwidth.txt https://www.unicode.org/Public/UCD/latest/ucd/EastAsianWidth.txt
-shineget libc/unicode/SpecialCasing.txt  https://www.unicode.org/Public/UCD/latest/ucd/SpecialCasing.txt