bring scan_cescape up to speed

textcode/scan_cescape.3

@@ -0,0 +1,35 @@
+.TH scan_cescape 3
+.SH NAME
+scan_cescape \- parse C escaped string
+.SH SYNTAX
+.B #include <libowfat/textcode.h>
+
+size_t \fBscan_cescape\fP(const char *\fIsrc\fR,char *\fIdest\fR,size_t* \fIdestlen\fR);
+
+.SH DESCRIPTION
+scan_cescape parses C escaped text from src into dest.
+Parsing stops at the 0 terminator, invalid input characters or a double
+quote that was not escaped.
+
+C escape sequences like \\n and \\x0a and \\012 are translated into
+their binary counterparts. The C99 escape sequences \\u and \\U are
+supported and lead to UTF-8 sequences being output.
+
+scan_cescape will then write the number of bytes in dest into *destlen,
+and return the number of bytes decoded from src.
+
+dest can be NULL. destlen can be NULL.
+
+To make sure dest is large enough, either allocate strlen(src)+1 bytes
+or call scan_cescape twice, the first time with dest == NULL (*destlen
+will still be written).
+
+.SH "RETURN VALUE"
+scan_cescape returns the number of bytes successfully parsed
+from src.
+
+.SH EXAMPLES
+scan_cescape("test\\n\");",buf,&i) -> return 7, i=5, buf="test\n".
+
+.SH "SEE ALSO"
+fmt_jsonescape(3), fmt_cescape(3), scan_ldapescape(3)

textcode/scan_cescape.c

@@ -18,25 +18,84 @@ size_t scan_cescape(const char *src,char *dest,size_t *destlen) {
       case 't': c='\t'; break;
       case 'v': c='\v';
       case '\\': break;
-      case 'x':
+      case 'x':	// hex escape; \x0a -> 10
 	{
 	  unsigned char a,b;
 	  a=scan_fromhex(s[i+2]);
 	  b=scan_fromhex(s[i+3]);
-	  if (a<16 && b<16) {
+	  if (a<16) {
+	    if (b<16) {
 	      c=(a<<4)+b;
 	      i+=2;
+	    } else {
+	      c=a;
+	      i+=1;
+	    }
 	  }
 	}
 	break;
+      case 'u': // C99 unicode escape: \u000a -> 10
+      case 'U': // C99 unicode escape: \U0000000a -> 10
+	{
+	  unsigned int j,k=0,l=(s[i+1]=='U'?10:6);
+	  for (j=2; j<l; ++j) {
+	    unsigned char c=scan_fromhex(s[i+j]);
+	    if (c>=16) // error
+	      goto error;	// don't allow short sequences
+	    k=k*16+c;
+	  }
+	  written+=fmt_utf8(dest?dest+written:0,k);
+	  i+=j-1;
+	  continue;
+	}
       default:
+	if (s[i+1]>='0' && s[i+1]<='7') {	// octal escape; \012 -> 10
+	  unsigned int j,k;
+	  for (k=0,j=1; j<4; ++j) {
+	    unsigned int l=s[i+j]-'0';
+	    if (l<8)
+	      k=k*8+l;
+	    else
+	      break;
+	  }
+	  if (dest) dest[written++]=k;
+	  i+=j-1;
+	  continue;
+	}
 	--i;
       }
       ++i;
-    }
-    dest[written]=c;
+    } else if (c=='"')
+      break;
+    if (dest) dest[written]=c;
     ++written;
   }
-  *destlen=written;
+error:
+  if (destlen) *destlen=written;
   return i;
 }
+
+#ifdef UNITTEST
+#include <assert.h>
+#undef UNITTEST
+#include <scan/scan_fromhex.c>
+#include <fmt/fmt_utf8.c>
+
+int main() {
+  size_t dl;
+  char buf[100];
+  assert(scan_cescape("test\\n\");",buf,&dl)==6 && dl==5 && !memcmp(buf,"test\n",5));
+  /* check hex and octal escaping */
+  assert(scan_cescape("test\\x0a\");",buf,&dl)==8 && dl==5 && !memcmp(buf,"test\n",5));
+  assert(scan_cescape("test\\012\");",buf,&dl)==8 && dl==5 && !memcmp(buf,"test\n",5));
+  /* check short escape sequences */
+  assert(scan_cescape("test\\xa\");",buf,&dl)==7 && dl==5 && !memcmp(buf,"test\n",5));
+  assert(scan_cescape("test\\12\");",buf,&dl)==7 && dl==5 && !memcmp(buf,"test\n",5));
+  assert(scan_cescape("test\\1\");",buf,&dl)==6 && dl==5 && !memcmp(buf,"test\1",5));
+  /* check unicode */
+  assert(scan_cescape("test\\u000a\");",buf,&dl)==10 && dl==5 && !memcmp(buf,"test\n",5));
+  assert(scan_cescape("test\\U0000000a\");",buf,&dl)==14 && dl==5 && !memcmp(buf,"test\n",5));
+  /* check that short sequences are rejected */
+  assert(scan_cescape("test\\Ua\");",buf,&dl)==4 && dl==4 && !memcmp(buf,"test",4));
+}
+#endif

textcode/scan_jsonescape.3

@@ -0,0 +1,31 @@
+.TH scan_jsonescape 3
+.SH NAME
+scan_jsonescape \- parse JSON escaped string
+.SH SYNTAX
+.B #include <libowfat/textcode.h>
+
+size_t \fBscan_jsonescape\fP(const char *\fIsrc\fR,char *\fIdest\fR,size_t* \fIdestlen\fR);
+
+.SH DESCRIPTION
+scan_jsonescape parses JSON escaped text from src into dest, leaving a
+UTF-8 string in dest. Parsing stops at the 0 terminator, invalid input
+characters or a double quote that was not escaped.
+
+It will then write the number of bytes in dest into *destlen,
+and return the number of bytes decoded from src.
+
+dest can be NULL. destlen can be NULL.
+
+To make sure dest is large enough, either allocate strlen(src)+1 bytes
+or call scan_jsonescape twice, the first time with dest == NULL (*destlen
+will still be written).
+
+.SH "RETURN VALUE"
+scan_jsonescape returns the number of bytes successfully parsed
+from src.
+
+.SH EXAMPLES
+scan_jsonescape("test\n\");",buf,&i) -> return 6, i=5, buf="test\n".
+
+.SH "SEE ALSO"
+fmt_jsonescape(3), scan_cescape(3), scan_ldapescape(3)

textcode/scan_jsonescape.c

@@ -73,6 +73,10 @@ abort:
 #ifdef UNITTEST
 #include <assert.h>
 #include <string.h>
+#undef UNITTEST
+#include <scan/scan_fromhex.c>
+#include <scan/scan_utf8.c>
+#include <fmt/fmt_utf8.c>
 
 int main() {
   char buf[100];
@@ -105,3 +109,5 @@ int main() {
   return 0;
 }
 #endif
+
+