suzanne.soy/racket
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix reading of badly encoded symbols from bytecode

Browse Source 
This is the third try to fix the bug exposed by "fuzz.rkt". Previous
repairs addressed a symptom at the point of printing bad symbols,
instead of the cause at a failure to validate a symbol's encoding
when reading bytecode. This one fixes reading.
This commit is contained in:
Matthew Flatt 2014-06-23 16:10:04 +01:00
parent 408d6bb773
commit 58eab92dc3
2 changed files with 21 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
racket/src/racket/src/read.c
View File

@ -4474,6 +4474,11 @@ static Scheme_Object *read_compact_svector(CPort *port, int l)
return o;
}
static int valid_utf8(const char *s, int l)
{
return (scheme_utf8_decode((const unsigned char *)s, 0, l, NULL, 0, -1, NULL, 0, 0) >= 0);
}
static Scheme_Object *read_escape_from_string(char *s, intptr_t len,
Scheme_Object *rel_to,
@ -4577,6 +4582,8 @@ static Scheme_Object *read_compact(CPort *port, int use_stack)
l = read_compact_number(port);
RANGE_CHECK_GETS(l);
s = read_compact_chars(port, buffer, BLK_BUF_SIZE, l);
if (!valid_utf8(s, l))
scheme_ill_formed_code(port);
v = scheme_intern_exact_symbol(s, l);
if (SAME_OBJ(v, port->magic_sym))
@ -4609,6 +4616,9 @@ static Scheme_Object *read_compact(CPort *port, int use_stack)
RANGE_CHECK_GETS(l);
s = read_compact_chars(port, buffer, BLK_BUF_SIZE, l);
if (!valid_utf8(s, l))
scheme_ill_formed_code(port);
if (uninterned)
v = scheme_make_exact_symbol(s, l);
else
@ -4623,6 +4633,8 @@ static Scheme_Object *read_compact(CPort *port, int use_stack)
l = read_compact_number(port);
RANGE_CHECK_GETS(l);
s = read_compact_chars(port, buffer, BLK_BUF_SIZE, l);
if (!valid_utf8(s, l))
scheme_ill_formed_code(port);
v = scheme_intern_exact_keyword(s, l);
break;
case CPT_BYTE_STRING:
@ -5037,6 +5049,8 @@ static Scheme_Object *read_compact(CPort *port, int use_stack)
l = ch - CPT_SMALL_SYMBOL_START;
RANGE_CHECK_GETS(l);
s = read_compact_chars(port, buffer, BLK_BUF_SIZE, l);
if (!valid_utf8(s, l))
scheme_ill_formed_code(port);
v = scheme_intern_exact_symbol(s, l);
if (SAME_OBJ(v, port->magic_sym))
@ -5301,6 +5315,9 @@ Scheme_Object *scheme_string_to_submodule_path(char *_s, intptr_t len)
e[l] = 0;
pos += l;
if (!valid_utf8(e, l))
return scheme_null;
pr = scheme_make_pair(scheme_intern_exact_symbol(e, l), scheme_null);
if (last)
SCHEME_CDR(last) = pr;

8
racket/src/racket/src/symbol.c
View File

@ -611,11 +611,11 @@ const char *scheme_symbol_name_and_size(Scheme_Object *sym, uintptr_t *length, i
if (ch > 127) {
/* Decode UTF-8. */
mzchar buf[2];
int ul = 1;
int ul = 2;
while (1) {
if (scheme_utf8_decode_offset_prefix((unsigned char *)s, i, i + ul,
buf, 0, 1,
NULL, 0, '?') > 0)
if (scheme_utf8_decode((unsigned char *)s, i, i + ul,
buf, 0, 1,
NULL, 0, 0) > 0)
break;
ul++;
}