cdecl99: Fall back to getline, instead of Gnulib's readline.

[cdecl99.git] / src / scan.l

diff --git a/src/scan.l b/src/scan.l
index 5610dc30fb7c5b891c59fd4c042ab05b85cde04c..0a4db93bb2ec74e4b64b7ca8085cd82873c7225b 100644 (file)
--- a/src/scan.l
+++ b/src/scan.l
@@ -53,6 +53,67 @@
        } \
        strcpy(yylval->strval, yytext); \
 } while(0)
+
+static char *to_octal(char *dst, unsigned val)
+{
+       unsigned i;
+
+       for (i = 0; i < 3; i++) {
+               *dst++ = '0' + ((val >> 6) & 7u);
+               val <<= 3;
+       }
+
+       return dst;
+}
+
+/*
+ * Convert a single character to a C-style character constant, including quote
+ * characters.  At most 7 bytes are written to the buffer for the longest
+ * octal encoding, e.g., '\177'
+ */
+static void to_readable_ch(char *dst, char c)
+{
+       unsigned char uc = c;
+       unsigned i;
+       char esc;
+
+       /*
+        * The 7 standard C control characters are contiguous in ASCII,
+        * permitting a simple and compact lookup table; separating their
+        * handling from backslash and quote characters hopefully allows
+        * the compiler to recognize that.
+        */
+       switch (c) {
+       case '\a': i = 0; break;
+       case '\b': i = 1; break;
+       case '\t': i = 2; break;
+       case '\n': i = 3; break;
+       case '\v': i = 4; break;
+       case '\f': i = 5; break;
+       case '\r': i = 6; break;
+       default:   i = 7; break;
+       }
+       esc = "abtnvfr"[i];
+
+       /* Otherwise printable characters that should still be escaped. */
+       switch (c) {
+       case '\\': case '\'': esc = c; break;
+       }
+
+       *dst++ = '\'';
+       if (esc) {
+               *dst++ = '\\';
+               *dst++ = esc;
+       } else if (isprint(uc)) {
+               *dst++ = c;
+       } else {
+               *dst++ = '\\';
+               dst = to_octal(dst, uc);
+       }
+       *dst++ = '\'';
+       *dst++ = 0;
+}
+
 %}
 
 %s ENGLISH
@@ -142,29 +203,9 @@ INTEGER 0x[[:xdigit:]]+|0[0-7]+|[[:digit:]]+
 
 [[:space:]]+
 . {
-       char buf[5] = { yytext[0] };
-       unsigned char c = buf[0];
-
-       if (!isprint(c) || c == '\\' || c == '\'') {
-               /* Encode nonprinting characters with C-style escapes */
-               buf[0] = '\\';
-               switch (c) {
-               case '\a': buf[1] = 'a'; break;
-               case '\b': buf[1] = 'b'; break;
-               case '\f': buf[1] = 'f'; break;
-               case '\n': buf[1] = 'n'; break;
-               case '\r': buf[1] = 'r'; break;
-               case '\t': buf[1] = 't'; break;
-               case '\v': buf[1] = 'v'; break;
-               case '\\': buf[1] = '\\'; break;
-               case '\'': buf[1] = '\''; break;
-               default:
-                       buf[1] = '0' + ((c >> 6) & 3);
-                       buf[2] = '0' + ((c >> 3) & 7);
-                       buf[3] = '0' + ((c >> 0) & 7);
-               }
-       }
+       char buf[8];
 
-       cdecl__err(CDECL_ENOPARSE, _("syntax error, unexpected '%s'"), buf);
+       to_readable_ch(buf, yytext[0]);
+       cdecl__err(CDECL_ENOPARSE, _("syntax error, unexpected %s"), buf);
        return T_LEX_ERROR;
 }