On Sun, May 08, 2005 at 11:45:08PM +0200, Jens Rieks wrote:
> Hi,
>
> C<is_cclass> is now implemented. The second part (find_(not_)cclass) will
> follow tomorrow.
The attached patch file adjusts C<is_cclass> to always return false
for offsets beyond the end of the string, and updates
t/op/string_cclass.t to test this.
Pm
Index: charset/iso-8859-1.c
===================================================================
--- charset/iso-8859-1.c (revision 8017)
+++ charset/iso-8859-1.c (working copy)
@@ -322,8 +322,9 @@
is_cclass(Interp *interpreter, PARROT_CCLASS_FLAGS flags, STRING
*source_string, UINTVAL offset)
{
UINTVAL codepoint;
+ if (offset >= source_string->strlen) return 0;
+
codepoint = ENCODING_GET_CODEPOINT(interpreter, source_string, offset);
-
if (codepoint >= sizeof(Parrot_ascii_typetable) /
sizeof(Parrot_ascii_typetable[0])) {
return 0;
}
Index: charset/ascii.c
===================================================================
--- charset/ascii.c (revision 8017)
+++ charset/ascii.c (working copy)
@@ -555,8 +555,9 @@
is_cclass(Interp *interpreter, PARROT_CCLASS_FLAGS flags, STRING
*source_string, UINTVAL offset)
{
UINTVAL codepoint;
+ if (offset >= source_string->strlen) return 0;
+
codepoint = ENCODING_GET_CODEPOINT(interpreter, source_string, offset);
-
if (codepoint >= sizeof(Parrot_ascii_typetable) /
sizeof(Parrot_ascii_typetable[0])) {
return 0;
}
Index: t/op/string_cclass.t
===================================================================
--- t/op/string_cclass.t (revision 8017)
+++ t/op/string_cclass.t (working copy)
@@ -61,40 +61,40 @@
$I2 = is_cclass code, str, $I0
print $I2
inc $I0
- if $I0 < $I1 goto loop
+ if $I0 <= $I1 goto loop
print "\n"
.end
CODE
-000110000000
-110000000000
-110110000000
-000001100000
-110101100000
-001000000110
-110111110101
-110111110001
-000000000110
-001000001010
-000000010001
-110111100000
-001000000000
-001000000110
-111000000110
-000110010001
-000110000000
-110000000000
-110110000000
-000001100000
-110101100000
-001000000110
-110111110101
-110111110001
-000000000110
-001000001010
-000000010001
-110111100000
-001000000000
-001000000110
-111000000110
-000110010001
+0001100000000
+1100000000000
+1101100000000
+0000011000000
+1101011000000
+0010000001100
+1101111101010
+1101111100010
+0000000001100
+0010000010100
+0000000100010
+1101111000000
+0010000000000
+0010000001100
+1110000001100
+0001100100010
+0001100000000
+1100000000000
+1101100000000
+0000011000000
+1101011000000
+0010000001100
+1101111101010
+1101111100010
+0000000001100
+0010000010100
+0000000100010
+1101111000000
+0010000000000
+0010000001100
+1110000001100
+0001100100010
OUT
