Add control character representations

Signed-off-by: Mattias Andrée <m@maandree.se>

6 files changed, 116 insertions, 2 deletions

diff --git a/Makefile b/Makefile
index c11b562..2777447 100644
--- a/Makefile
+++ b/Makefile
@@ -73,7 +73,8 @@ BIN =\
 	convert-to-negative\
 	convert-to-symbols\
 	convert-to-control-characters\
-	convert-to-xiangqi
+	convert-to-xiangqi\
+	convert-to-control-character-representations
 
 LIBOBJ =\
 	libcharconv_decode_utf8_.o\
@@ -136,7 +137,8 @@ LIBOBJ =\
 	libcharconv_symbols.o\
 	libcharconv_control_characters.o\
 	libcharconv_xiangqi_red.o\
-	libcharconv_xiangqi_black.o
+	libcharconv_xiangqi_black.o\
+	libcharconv_control_character_representations.o
 
 LOBJ = $(LIBOBJ:.o=.lo)
 
diff --git a/convert-to-control-character-representations.c b/convert-to-control-character-representations.c
new file mode 100644
index 0000000..1ea01c9
--- /dev/null
+++ b/convert-to-control-character-representations.c
@@ -0,0 +1,4 @@
+/* See LICENSE file for copyright and license details. */
+#include "common.h"
+
+SIMPLE(libcharconv_control_character_representations)
diff --git a/libcharconv.h b/libcharconv.h
index 629fd39..14627c6 100644
--- a/libcharconv.h
+++ b/libcharconv.h
@@ -386,6 +386,7 @@ LIBCHARCONV_FUNC_(libcharconv_symbols);
 /**
  * Convert characters and character sequences to
  * control characters, spaces, and SOFT HYPHENs
+ * (textual representation symbols can be converted)
  */
 LIBCHARCONV_FUNC_(libcharconv_control_characters);
 
@@ -401,6 +402,13 @@ LIBCHARCONV_FUNC_(libcharconv_xiangqi_red);
  */
 LIBCHARCONV_FUNC_(libcharconv_xiangqi_black);
 
+/**
+ * Convert ASCII control characters, and SPACE,
+ * and their abbrevations to textual representation
+ * symbols
+ */
+LIBCHARCONV_FUNC_(libcharconv_control_character_representations);
+
 
 #undef LIBCHARCONV_FUNC_
 #endif
diff --git a/libcharconv_control_character_representations.c b/libcharconv_control_character_representations.c
new file mode 100644
index 0000000..57c1c93
--- /dev/null
+++ b/libcharconv_control_character_representations.c
@@ -0,0 +1,65 @@
+/* See LICENSE file for copyright and license details. */
+#include "lib-common.h"
+#include <string.h>
+
+
+static const char *texts[] = {
+	"NUL", "SOH", "STX", "ETX", "EOT", "ENQ", "ACK", "BEL",
+	"BS",  "HT",  "LF", " VT",  "FF",  "CR",  "SS",  "SI",
+	"DLE", "DC1", "DC2", "DC3", "DC4", "NAK", "SYN", "ETB",
+	"CAN", "EM",  "SUB", "ESC", "FS",  "GS",  "RS",  "US",
+	"SP",  "DEL"
+};
+
+
+enum libcharconv_result
+libcharconv_control_character_representations(const char *s, size_t slen, size_t *n, uint_least32_t *cp, size_t *ncp)
+{
+	size_t i, len, found, found_len;
+	int indeterminate;
+	*n = 0;
+	for (; slen; s++, slen--, ++*n) {
+		if (*(const unsigned char *)s <= ' ') {
+			found = *(const unsigned char *)s;
+			goto conv_byte;
+		} else if (*(const unsigned char *)s == 0x7Fu) {
+			found = 0x21u;
+			goto conv_byte;
+		}
+		indeterminate = 0;
+		found = SIZE_MAX;
+		found_len = 0u;
+		for (i = 0u; i < sizeof(texts) / sizeof(*texts); i++) {
+			len = strlen(texts[i]);
+			if (strncmp(s, texts[i], len < slen ? len : slen))
+				continue;
+			if (slen < len) {
+				indeterminate = 1;
+				continue;
+			}
+			if (len > found_len) {
+				found = i;
+				found_len = len;
+			}
+		}
+		if (found_len)
+			goto conv;
+		if (*n)
+			goto no_conv;
+		if (indeterminate)
+			return LIBCHARCONV_INDETERMINATE;
+	}
+no_conv:
+	return LIBCHARCONV_NO_CONVERT;
+
+conv_byte:
+	found_len = 1u;
+conv:
+	if (*n)
+		goto no_conv;
+	if (*ncp)
+		*cp = (uint_least32_t)UINT32_C(0x2400) | (uint_least32_t)found;
+	*n += found_len;
+	*ncp = 1u;
+	return indeterminate ? LIBCHARCONV_CONVERT_IF_END : LIBCHARCONV_CONVERTED;
+}
diff --git a/libcharconv_control_characters.c b/libcharconv_control_characters.c
index 55e4bbf..c45f03e 100644
--- a/libcharconv_control_characters.c
+++ b/libcharconv_control_characters.c
@@ -132,6 +132,21 @@ libcharconv_control_characters(const char *s, size_t slen, size_t *n, uint_least
 	int indeterminate;
 	*n = 0;
 	for (; slen; s++, slen--, ++*n) {
+		if (s[0] == '\xE2') {
+			if (slen == 1u)
+				return LIBCHARCONV_INDETERMINATE;
+			if (s[1] != '\x90')
+				goto search;
+			if (slen == 2u)
+				return LIBCHARCONV_INDETERMINATE;
+			i = ((const unsigned char *)s)[2];
+			if (0x80u > i || i > 0xA1u)
+				goto search;
+			i &= 0x3Fu;
+			i = i == 0x21u ? 0x7Fu : i;
+			goto conv_repr;
+		}
+	search:
 		indeterminate = 0;
 		found = SIZE_MAX;
 		found_len = 0u;
@@ -158,6 +173,15 @@ libcharconv_control_characters(const char *s, size_t slen, size_t *n, uint_least
 no_conv:
 	return LIBCHARCONV_NO_CONVERT;
 
+conv_repr:
+	if (*n)
+		goto no_conv;
+	if (*ncp)
+		*cp = (uint_least32_t)i;
+	*n += 3u;
+	*ncp = 1u;
+	return LIBCHARCONV_CONVERTED;
+
 conv:
 	if (*n)
 		goto no_conv;
diff --git a/libcharconv_latin.c b/libcharconv_latin.c
index 360b3dd..047dee5 100644
--- a/libcharconv_latin.c
+++ b/libcharconv_latin.c
@@ -449,6 +449,17 @@ libcharconv_latin(const char *s, size_t slen, size_t *n, uint_least32_t *cp, siz
 			c = (uint_least32_t)"gaehrcsGAEHRCS"[c - UINT32_C(0x1FA60)];
 			goto conv;
 
+		} else if (UINT32_C(0x2400) <= c && c <= UINT32_C(0x2421)) {
+			/* control character representations */
+			cs = ((const char *[]){
+				"NUL", "SOH", "STX", "ETX", "EOT", "ENQ", "ACK", "BEL",
+				"BS",  "HT",  "LF", " VT",  "FF",  "CR",  "SS",  "SI",
+				"DLE", "DC1", "DC2", "DC3", "DC4", "NAK", "SYN", "ETB",
+				"CAN", "EM",  "SUB", "ESC", "FS",  "GS",  "RS",  "US",
+				"SP",  "DEL"
+			})[c & 0xFF];
+			goto conv_str;
+
 		} else {
 		use_switch:
 			switch (c) {