Add braille

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

5 files changed, 147 insertions, 3 deletions

diff --git a/Makefile b/Makefile
index f2e0bb9..c1b5690 100644
--- a/Makefile
+++ b/Makefile
@@ -52,7 +52,8 @@ BIN =\
 	convert-to-domino-tiles\
 	convert-to-clock-faces\
 	convert-to-ocr\
-	convert-to-crop-marks
+	convert-to-crop-marks\
+	convert-to-braille
 
 LIBOBJ =\
 	libcharconv_decode_utf8_.o\
@@ -93,7 +94,8 @@ LIBOBJ =\
 	libcharconv_domino_tiles_vertical.o\
 	libcharconv_clock_faces.o\
 	libcharconv_ocr.o\
-	libcharconv_crop_marks.o
+	libcharconv_crop_marks.o\
+	libcharconv_braille.o
 
 LOBJ = $(LIBOBJ:.o=.lo)
 
diff --git a/convert-to-braille.c b/convert-to-braille.c
new file mode 100644
index 0000000..c5513c9
--- /dev/null
+++ b/convert-to-braille.c
@@ -0,0 +1,18 @@
+/* See LICENSE file for copyright and license details. */
+#include "common.h"
+
+USAGE("");
+
+
+int
+main(int argc, char *argv[])
+{
+	ARGBEGIN {
+	default:
+		usage();
+	} ARGEND;
+	if (argc)
+		usage();
+
+	return convert(&libcharconv_braille);
+}
diff --git a/libcharconv.h b/libcharconv.h
index 5b165b6..613ef1b 100644
--- a/libcharconv.h
+++ b/libcharconv.h
@@ -1278,4 +1278,39 @@ enum libcharconv_result libcharconv_ocr(const char *s, size_t slen, size_t *n, u
 enum libcharconv_result libcharconv_crop_marks(const char *s, size_t slen, size_t *n, uint_least32_t *cp, size_t *ncp);
 
 
+/**
+ * Convert sequences of unique [1, 8]-digits to BRAILLE PATTERNs,
+ * a sequence can be empty (for the blank pattern), each
+ * sequence can be (and the empty sequence must be) termianted
+ * by a DIGIT ZERO
+ * 
+ * @param   s     Text to convert
+ * @param   slen  The number of bytes available in `s`
+ * @param   n     Output parameter for the number of consumed bytes
+ * @param   cp    Output buffer for the codepoints
+ * @param   ncp   Input parameter for the number of codepoints that
+ *                fit in `cp`, and output parameter for the number
+ *                of output codepoints (if it exceeds the original
+ *                value of `ncp`, a larger buffer is needed)
+ * @return        LIBCHARCONV_NO_CONVERT:
+ *                  `*n` is the number of bytes from the beginning
+ *                  of `s` that cannot be converted
+ *                LIBCHARCONV_CONVERTED:
+ *                  `*n` is the number of bytes from the beginning
+ *                  of `s` that was converted to a codepoint which
+ *                  is stored in `*cp`
+ *                LIBCHARCONV_INDETERMINATE:
+ *                  If all text has been input, no more can be
+ *                  converted, otherwise more of the text most
+ *                  be made available before the function can
+ *                  determine whether the beginning of `s` can be
+ *                  converted or what it should be converted to
+ *                LIBCHARCONV_CONVERT_IF_END:
+ *                  As LIBCHARCONV_CONVERTED the entire text has
+ *                  been input, as LIBCHARCONV_INDETERMINATE
+ *                  otherwise
+ */
+enum libcharconv_result libcharconv_braille(const char *s, size_t slen, size_t *n, uint_least32_t *cp, size_t *ncp);
+
+
 #endif
diff --git a/libcharconv_braille.c b/libcharconv_braille.c
new file mode 100644
index 0000000..3a9807e
--- /dev/null
+++ b/libcharconv_braille.c
@@ -0,0 +1,62 @@
+/* See LICENSE file for copyright and license details. */
+#include "lib-common.h"
+
+
+enum libcharconv_result
+libcharconv_braille(const char *s, size_t slen, size_t *n, uint_least32_t *cp, size_t *ncp)
+{
+	unsigned pattern, dot;
+	uint_least32_t c;
+	*n = 0;
+	for (; slen--; s++) {
+		if (s[0] == '0') {
+			if (*n)
+				goto no_conv;
+			c = UINT32_C(0x2800);
+			*n = 1u;
+			goto conv;
+		} else if ('1' <= s[0] && s[0] <= '8') {
+			if (*n)
+				goto no_conv;
+			pattern = 1u << (s[0] - '1');
+			*n = 1u;
+			for (;;) {
+				if (!slen--)
+					goto indeterminate;
+				if ('1' <= s[*n] && s[*n] <= '8') {
+					dot = 1u << (s[*n] - '1');
+					if (pattern & dot)
+						break;
+					pattern |= dot;
+					*n += 1u;
+				} else if (s[*n] == '0') {
+					*n += 1u;
+					break;
+				} else {
+					break;
+				}
+			}
+			c = UINT32_C(0x2800) | pattern;
+			goto conv;
+		} else {
+			*n += 1u;
+		}
+	}
+	return LIBCHARCONV_NO_CONVERT;
+
+no_conv:
+	if (!*n)
+		*n = 1u;
+	return LIBCHARCONV_NO_CONVERT;
+
+indeterminate:
+	if (*n)
+		goto no_conv;
+	return LIBCHARCONV_INDETERMINATE;
+
+conv:
+	if (*ncp)
+		*cp = c;
+	*ncp = 1u;
+	return LIBCHARCONV_CONVERTED;
+}
diff --git a/libcharconv_latin.c b/libcharconv_latin.c
index de5b618..5124c53 100644
--- a/libcharconv_latin.c
+++ b/libcharconv_latin.c
@@ -8,7 +8,7 @@ libcharconv_latin(const char *s, size_t slen, size_t *n, uint_least32_t *cp, siz
 	enum libcharconv_result ret = LIBCHARCONV_CONVERTED;
 	uint_least32_t c;
 	char c1, c2, c3, c4, c5;
-	size_t clen;
+	size_t i, clen;
 	unsigned num;
 
 	*n = 0;
@@ -355,6 +355,33 @@ libcharconv_latin(const char *s, size_t slen, size_t *n, uint_least32_t *cp, siz
 			c = (uint_least32_t)"SdYAIXC/P_\\"[c - UINT32_C(0x2440)];
 			goto conv;
 
+		} else if (UINT32_C(0x2800) <= c && c <= UINT32_C(0x28FF)) {
+			/* braille */
+			if (*n)
+				goto no_conv;
+			*n += clen;
+			i = 0u;
+			if ((c & 0x01u) && *ncp > i)
+				cp[i++] = '1';
+			if ((c & 0x02u) && *ncp > i)
+				cp[i++] = '2';
+			if ((c & 0x04u) && *ncp > i)
+				cp[i++] = '3';
+			if ((c & 0x08u) && *ncp > i)
+				cp[i++] = '4';
+			if ((c & 0x10u) && *ncp > i)
+				cp[i++] = '5';
+			if ((c & 0x20u) && *ncp > i)
+				cp[i++] = '6';
+			if ((c & 0x40u) && *ncp > i)
+				cp[i++] = '7';
+			if ((c & 0x80u) && *ncp > i)
+				cp[i++] = '8';
+			if (*ncp > i)
+				cp[i++] = '0';
+			*ncp = i;
+			return ret;
+
 		} else {
 			switch (c) {
 			/* shogi */