qstring: add nasm_unquote() supporting `...`

Add a nasm_unquote() function supporting the intended `...` syntax.

1 files changed, 205 insertions, 0 deletions

diff --git a/nasmlib.c b/nasmlib.c
index 8e56bf4..8cd41cf 100644
--- a/nasmlib.c
+++ b/nasmlib.c
@@ -950,6 +950,211 @@ void nasm_quote(char **str)
     *str = p;
 }
 
+static char *emit_utf8(char *q, int32_t v)
+{
+    if (v < 0) {
+	/* Impossible - do nothing */
+    } else if (v <= 0x7f) {
+	*q++ = v;
+    } else if (v <= 0x000007ff) {
+	*q++ = 0xc0 | (v >> 6);
+	*q++ = 0x80 | (v & 63);
+    } else if (v <= 0x0000ffff) {
+	*q++ = 0xe0 | (v >> 12);
+	*q++ = 0x80 | ((v >> 6) & 63);
+	*q++ = 0x80 | (v & 63);
+    } else if (v <= 0x001fffff) {
+	*q++ = 0xf0 | (v >> 18);
+	*q++ = 0x80 | ((v >> 12) & 63);
+	*q++ = 0x80 | ((v >> 6) & 63);
+	*q++ = 0x80 | (v & 63);
+    } else if (v <= 0x03ffffff) {
+	*q++ = 0xf8 | (v >> 24);
+	*q++ = 0x80 | ((v >> 18) & 63);
+	*q++ = 0x80 | ((v >> 12) & 63);
+	*q++ = 0x80 | ((v >> 6) & 63);
+	*q++ = 0x80 | (v & 63);
+    } else {
+	*q++ = 0xfc | (v >> 30);
+	*q++ = 0x80 | ((v >> 24) & 63);
+	*q++ = 0x80 | ((v >> 18) & 63);
+	*q++ = 0x80 | ((v >> 12) & 63);
+	*q++ = 0x80 | ((v >> 6) & 63);
+	*q++ = 0x80 | (v & 63);
+    }
+    return q;
+}
+
+/*
+ * Do an *in-place* dequoting of the specified string, returning the
+ * resulting length (which may be containing embedded nulls.)
+ *
+ * In-place replacement is possible since the unquoted length is always
+ * shorter than or equal to the quoted length.
+ */
+size_t nasm_unquote(char *str)
+{
+    size_t ln;
+    char bq, eq;
+    char *p, *q, *ep, *escp;
+    char c;
+    enum unq_state {
+	st_start,
+	st_backslash,
+	st_hex,
+	st_oct,
+	st_ucs,
+    } state;
+    int ndig = 0;
+    int32_t nval = 0;
+
+    bq = str[0];
+    if (!bq)
+	return 0;
+    ln = strlen(str);
+    eq = str[ln-1];
+
+    if ((bq == '\'' || bq == '\"') && bq == eq) {
+	/* '...' or "..." string */
+	memmove(str, str+1, ln-2);
+	str[ln-2] = '\0';
+	return ln-2;
+    }
+    if (bq == '`' || eq == '`') {
+	/* `...` string */
+	q = str;
+	p = str+1;
+	ep = str+ln-1;
+	state = st_start;
+
+	while (p < ep) {
+	    c = *p++;
+	    switch (state) {
+	    case st_start:
+		if (c == '\\')
+		    state = st_backslash;
+		else
+		    *q++ = c;
+		break;
+
+	    case st_backslash:
+		state = st_start;
+		escp = p-1;
+		switch (c) {
+		case 'a':
+		    *q++ = 7;
+		    break;
+		case 'b':
+		    *q++ = 8;
+		    break;
+		case 'e':
+		    *q++ = 27;
+		    break;
+		case 'f':
+		    *q++ = 12;
+		    break;
+		case 'n':
+		    *q++ = 10;
+		    break;
+		case 'r':
+		    *q++ = 13;
+		    break;
+		case 't':
+		    *q++ = 9;
+		    break;
+		case 'u':
+		    state = st_ucs;
+		    ndig = 4;
+		    nval = 0;
+		    break;
+		case 'U':
+		    state = st_ucs;
+		    ndig = 8;
+		    nval = 0;
+		    break;
+		case 'v':
+		    *q++ = 11;
+		case 'x':
+		case 'X':
+		    state = st_hex;
+		    ndig = nval = 0;
+		    break;
+		case '0':
+		case '1':
+		case '2':
+		case '3':
+		case '4':
+		case '5':
+		case '6':
+		case '7':
+		    state = st_oct;
+		    ndig = 1;
+		    nval = c - '0';
+		    break;
+		default:
+		    *q++ = c;
+		    break;
+		}
+		break;
+
+	    case st_oct:
+		if (c >= '0' && c <= '7') {
+		    nval = (nval << 3) + (c - '0');
+		    if (++ndig >= 3) {
+			*q++ = nval;
+			state = st_start;
+		    }
+		} else {
+		    p--;	/* Process this character again */
+		    *q++ = nval;
+		    state = st_start;
+		}
+		break;
+
+	    case st_hex:
+		if ((c >= '0' && c <= '9') ||
+		    (c >= 'A' && c <= 'F') ||
+		    (c >= 'a' && c <= 'f')) {
+		    nval = (nval << 4) + numvalue(c);
+		    if (++ndig >= 2) {
+			*q++ = nval;
+			state = st_start;
+		    }
+		} else {
+		    p--;	/* Process this character again */
+		    *q++ = ndig ? nval : *escp;
+		    state = st_start;
+		}
+		break;
+
+	    case st_ucs:
+		if ((c >= '0' && c <= '9') ||
+		    (c >= 'A' && c <= 'F') ||
+		    (c >= 'a' && c <= 'f')) {
+		    nval = (nval << 4) + numvalue(c);
+		    if (!--ndig) {
+			q = emit_utf8(q, nval);
+			state = st_start;
+		    }
+		} else {
+		    p--;	/* Process this character again */
+		    if (p > escp+1)
+			q = emit_utf8(q, nval);
+		    else
+			*q++ = *escp;
+		    state = st_start;
+		}
+		break;
+	    }
+	}
+	*q = '\0';
+	return q-str;
+    }
+
+    /* Otherwise, just return the input... */
+    return ln;
+}
+
 char *nasm_strcat(char *one, char *two)
 {
     char *rslt;