S_grok_bslash_[ox]: assert char fits

S_grok_bslash_o() and S_grok_bslash_x() scan \o{...} and \x{...} in a
string, and return a UV codepoint. Some callers assume that if that
codepoint is converted into a series of utf8 bytes and re-inserted into
the string, that it will tale up less space than the original "\x{...}"
sequence. Add asserts that this is indeed so.

1 files changed, 26 insertions, 4 deletions

diff --git a/dquote_static.c b/dquote_static.c
index 93c023b01e..401bf478eb 100644
--- a/dquote_static.c
+++ b/dquote_static.c
@@ -88,7 +88,9 @@ S_grok_bslash_o(pTHX_ char **s, UV *uv, const char** error_msg,
 
 /*  Documentation to be supplied when interface nailed down finally
  *  This returns FALSE if there is an error which the caller need not recover
- *  from; , otherwise TRUE.  In either case the caller should look at *len
+ *  from; otherwise TRUE.  In either case the caller should look at *len [???].
+ *  It guarantees that the returned codepoint, *uv, when expressed as
+ *  utf8 bytes, would fit within the skipped "\x{...}" bytes.
  *  On input:
  *	s   is the address of a pointer to a NULL terminated string that begins
  *	    with 'o', and the previous character was a backslash.  At exit, *s
@@ -118,6 +120,11 @@ S_grok_bslash_o(pTHX_ char **s, UV *uv, const char** error_msg,
 		 * ourselves */
 	        | PERL_SCAN_SILENT_ILLDIGIT;
 
+#ifdef DEBUGGING
+    char *start = *s - 1;
+    assert(*start == '\\');
+#endif
+
     PERL_ARGS_ASSERT_GROK_BSLASH_O;
 
 
@@ -176,6 +183,10 @@ S_grok_bslash_o(pTHX_ char **s, UV *uv, const char** error_msg,
     /* Return past the '}' */
     *s = e + 1;
 
+    /* guarantee replacing "\x{...}" with utf8 bytes fits within
+     * existing space */
+    assert(OFFUNISKIP(*uv) < *s - start);
+
     return TRUE;
 }
 
@@ -188,7 +199,10 @@ S_grok_bslash_x(pTHX_ char **s, UV *uv, const char** error_msg,
 
 /*  Documentation to be supplied when interface nailed down finally
  *  This returns FALSE if there is an error which the caller need not recover
- *  from; , otherwise TRUE.
+ *  from; otherwise TRUE.
+ *  It guarantees that the returned codepoint, *uv, when expressed as
+ *  utf8 bytes, would fit within the skipped "\x{...}" bytes.
+ *
  *  On input:
  *	s   is the address of a pointer to a NULL terminated string that begins
  *	    with 'x', and the previous character was a backslash.  At exit, *s
@@ -215,6 +229,10 @@ S_grok_bslash_x(pTHX_ char **s, UV *uv, const char** error_msg,
     char* e;
     STRLEN numbers_len;
     I32 flags = PERL_SCAN_DISALLOW_PREFIX;
+#ifdef DEBUGGING
+    char *start = *s - 1;
+    assert(*start == '\\');
+#endif
 
     PERL_ARGS_ASSERT_GROK_BSLASH_X;
 
@@ -240,7 +258,7 @@ S_grok_bslash_x(pTHX_ char **s, UV *uv, const char** error_msg,
             }
             return FALSE;
         }
-	return TRUE;
+	goto ok;
     }
 
     e = strchr(*s, '}');
@@ -267,7 +285,7 @@ S_grok_bslash_x(pTHX_ char **s, UV *uv, const char** error_msg,
         }
         *s = e + 1;
         *uv = 0;
-        return TRUE;
+        goto ok;
     }
 
     flags |= PERL_SCAN_ALLOW_UNDERSCORES;
@@ -289,6 +307,10 @@ S_grok_bslash_x(pTHX_ char **s, UV *uv, const char** error_msg,
     /* Return past the '}' */
     *s = e + 1;
 
+  ok:
+    /* guarantee replacing "\x{...}" with utf8 bytes fits within
+     * existing space */
+    assert(OFFUNISKIP(*uv) < *s - start);
     return TRUE;
 }