PR#6316: Scanf.scanf failure on %u formats when reading big integers.

The approach implemented is the second one suggested by Benoît Vaugon in the PR:
- The int_of_string functions accept a "0u" prefix meaning "decimal unsigned".
- The '%u' format of the scanf functions adds this "0u" prefix before conversion.
This is consistent with the current handling of unsigned hexa, octal, and binary numbers.


git-svn-id: http://caml.inria.fr/svn/ocaml/trunk@16241 f963ae5c-01c2-4b8c-9fe0-0dff7051ff02

Changes

@@ -45,6 +45,8 @@ Runtime system:
   (Peter Zotov, review by Jacques-Henri Jourdan)
 
 Standard library:
+- PR#6316: Scanf.scanf failure on %u formats when reading big integers
+  (Xavier Leroy, Benoît Vaugon)
 - PR#6390, GPR#36: expose Sys.{int_size,max_wosize} for improved js_of_ocaml portability
   (Hugo Heuzard)
 * PR#6494: Add equal function in modules

byterun/ints.c

@@ -23,6 +23,7 @@
 
 static char * parse_sign_and_base(char * p,
                                   /*out*/ int * base,
+                                  /*out*/ int * signedness,
                                   /*out*/ int * sign)
 {
   *sign = 1;
@@ -30,15 +31,17 @@ static char * parse_sign_and_base(char * p,
     *sign = -1;
     p++;
   }
-  *base = 10;
+  *base = 10; *signedness = 1;
   if (*p == '0') {
     switch (p[1]) {
     case 'x': case 'X':
-      *base = 16; p += 2; break;
+      *base = 16; *signedness = 0; p += 2; break;
     case 'o': case 'O':
-      *base = 8; p += 2; break;
+      *base = 8; *signedness = 0; p += 2; break;
     case 'b': case 'B':
-      *base = 2; p += 2; break;
+      *base = 2; *signedness = 0; p += 2; break;
+    case 'u': case 'U':
+      *signedness = 0; p += 2; break;
     }
   }
   return p;
@@ -65,9 +68,9 @@ static intnat parse_intnat(value s, int nbits, const char *errmsg)
 {
   char * p;
   uintnat res, threshold;
-  int sign, base, d;
+  int sign, base, signedness, d;
 
-  p = parse_sign_and_base(String_val(s), &base, &sign);
+  p = parse_sign_and_base(String_val(s), &base, &signedness, &sign);
   threshold = ((uintnat) -1) / base;
   d = parse_digit(*p);
   if (d < 0 || d >= base) caml_failwith(errmsg);
@@ -85,7 +88,7 @@ static intnat parse_intnat(value s, int nbits, const char *errmsg)
   if (p != String_val(s) + caml_string_length(s)){
     caml_failwith(errmsg);
   }
-  if (base == 10) {
+  if (signedness) {
     /* Signed representation expected, allow -2^(nbits-1) to 2^(nbits-1) - 1 */
     if (sign >= 0) {
       if (res >= (uintnat)1 << (nbits - 1)) caml_failwith(errmsg);
@@ -530,9 +533,9 @@ CAMLprim value caml_int64_of_string(value s)
 {
   char * p;
   uint64_t res, threshold;
-  int sign, base, d;
+  int sign, base, signedness, d;
 
-  p = parse_sign_and_base(String_val(s), &base, &sign);
+  p = parse_sign_and_base(String_val(s), &base, &signedness, &sign);
   threshold = ((uint64_t) -1) / base;
   d = parse_digit(*p);
   if (d < 0 || d >= base) caml_failwith(INT64_ERRMSG);
@@ -551,7 +554,7 @@ CAMLprim value caml_int64_of_string(value s)
   if (p != String_val(s) + caml_string_length(s)){
     caml_failwith(INT64_ERRMSG);
   }
-  if (base == 10) {
+  if (signedness) {
     /* Signed representation expected, allow -2^63 to 2^63 - 1 only */
     if (sign >= 0) {
       if (res >= (uint64_t)1 << 63) caml_failwith(INT64_ERRMSG);

stdlib/scanf.ml

@@ -500,7 +500,8 @@ let token_bool ib =
 let token_int_literal conv ib =
   let tok =
     match conv with
-    | 'd' | 'i' | 'u' -> Scanning.token ib
+    | 'd' | 'i' -> Scanning.token ib
+    | 'u' -> "0u" ^ Scanning.token ib
     | 'o' -> "0o" ^ Scanning.token ib
     | 'x' | 'X' -> "0x" ^ Scanning.token ib
     | 'b' -> "0b" ^ Scanning.token ib