1 files changed, 142 insertions, 0 deletions
diff --git a/app/tools/halibut/charset/hz.c b/app/tools/halibut/charset/hz.c
new file mode 100644
index 0000000..1266e7e
--- /dev/null
+++ b/app/tools/halibut/charset/hz.c
@@ -0,0 +1,142 @@
+/*
+ * hz.c - HZ textual encoding of ASCII and GB2312, as defined in RFC 1843.
+ */
+
+#ifndef ENUM_CHARSETS
+
+#include <assert.h>
+
+#include "charset.h"
+#include "internal.h"
+
+static void read_hz(charset_spec const *charset, long int input_chr,
+		    charset_state *state,
+		    void (*emit)(void *ctx, long int output), void *emitctx)
+{
+    /*
+     * When reading, our state variables are:
+     * 
+     *  - s0 is 0 in ASCII mode, 1 in GB2312 mode.
+     * 
+     * 	- s1 stores a character we have just seen but not fully
+     * 	  processed. So in ASCII mode, this can only ever be zero
+     * 	  (no character) or 0x7E (~); in GB2312 mode it can be
+     * 	  anything from 0x21-0x7E.
+     */
+
+    UNUSEDARG(charset);
+
+    if (state->s0 == 0) {
+	/*
+	 * ASCII mode.
+	 */
+
+	if (state->s1) {
+	    assert(state->s1 == '~');
+	    state->s1 = 0;
+	    /* Process the character after a tilde. */
+	    switch (input_chr) {
+	      case '~':
+		emit(emitctx, input_chr);
+		return;
+	      case '\n':
+		return;		       /* ~\n is ignored */
+	      case '{':
+		state->s0 = 1;	       /* switch to GB2312 mode */
+		return;
+	    }
+	} else if (input_chr == '~') {
+	    state->s1 = '~';
+	    return;
+	} else {
+	    /* In ASCII mode, any non-tildes go straight  */
+	    emit(emitctx, input_chr);
+	    return;
+	}
+    } else {
+	/*
+	 * GB2312 mode. As I understand it, we expect never to see
+	 * anything in this mode that isn't 0x21-0x7E. So if we do,
+	 * we'll simply throw an error and return to ASCII mode.
+	 */
+	if (input_chr < 0x21 || input_chr > 0x7E) {
+	    emit(emitctx, ERROR);
+	    state->s0 = state->s1 = 0;
+	    return;
+	}
+
+	/*
+	 * So if we don't have a character stored already, store
+	 * this one...
+	 */
+	if (!state->s1) {
+	    state->s1 = input_chr;
+	    return;
+	}
+
+	/*
+	 * ... otherwise, combine the stored char with this one.
+	 * This will give either `~}', the escape sequence to
+	 * return to ASCII mode, or something which we translate
+	 * through GB2312.
+	 */
+	if (state->s1 == '~' && input_chr == '}') {
+	    state->s1 = state->s0 = 0;
+	    return;
+	}
+
+	emit(emitctx, gb2312_to_unicode(state->s1 - 0x21, input_chr - 0x21));
+	state->s1 = 0;
+    }
+}
+
+static int write_hz(charset_spec const *charset, long int input_chr,
+		    charset_state *state,
+		    void (*emit)(void *ctx, long int output), void *emitctx)
+{
+    int desired_state, r, c;
+
+    UNUSEDARG(charset);
+
+    /*
+     * Analyse the input char.
+     */
+    if (input_chr < 0x80) {
+	desired_state = 0;
+	c = input_chr;
+    } else if (unicode_to_gb2312(input_chr, &r, &c)) {
+	desired_state = 1;
+    } else {
+	return FALSE;
+    }
+
+    if (state->s0 != (unsigned)desired_state) {
+	emit(emitctx, '~');
+	emit(emitctx, desired_state ? '{' : '}');
+	state->s0 = desired_state;
+    }
+
+    if (input_chr < 0)
+	return TRUE;		       /* special case: just reset state */
+
+    if (state->s0) {
+	/*
+	 * GB mode.
+	 */
+	emit(emitctx, 0x21 + r);
+	emit(emitctx, 0x21 + c);
+    } else {
+	emit(emitctx, c);
+    }
+    return TRUE;
+}
+
+const charset_spec charset_CS_HZ = {
+    CS_HZ, read_hz, write_hz, NULL
+};
+
+#else /* ENUM_CHARSETS */
+
+ENUM_CHARSET(CS_HZ)
+
+#endif /* ENUM_CHARSETS */