1berry/ruby
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

move dump format validation into parsing epilogue

Browse Source 
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61380 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This commit is contained in:
naruse 2017-12-21 05:09:16 +00:00
parent 29c6ca423c
commit 188d85934b
2 changed files with 77 additions and 98 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

165
string.c
View File

@ -6081,60 +6081,6 @@ rb_str_dump(VALUE str)
return result; return result;
} }
enum undump_source_format {
UNDUMP_SOURCE_SIMPLE, /* "..." */
UNDUMP_SOURCE_FORCE_ENCODING, /* "...".force_encoding("...") */
UNDUMP_SOURCE_INVALID
};
static enum undump_source_format
check_undump_source_format(const char *s, const char *s_end, long len, rb_encoding *enc,
VALUE *forced_enc_str, long *forced_enc_str_len)
{
unsigned int cbeg, cend;
const char *prev;
static const long force_encoding_minimum_len = rb_strlen_lit("\"\".force_encoding(\"\")");
static const char force_encoding_middle_part[] = "\".force_encoding(\"";
static const long force_encoding_middle_part_len = rb_strlen_lit("\".force_encoding(\"");
static const char force_encoding_end_part[] = "\")";
static const long force_encoding_end_part_len = rb_strlen_lit("\")");
long pos_before_middle_part, pos_before_end_part, pos_after_middle_part;
if (len < 2) return UNDUMP_SOURCE_INVALID;
cbeg = rb_enc_mbc_to_codepoint(s, s_end, enc);
if (cbeg != '"') return UNDUMP_SOURCE_INVALID;
prev = rb_enc_prev_char(s, s_end, s_end, enc);
cend = rb_enc_mbc_to_codepoint(prev, s_end, enc);
if (cend == '"') return UNDUMP_SOURCE_SIMPLE;
if (cend != ')' || len < force_encoding_minimum_len) {
return UNDUMP_SOURCE_INVALID;
}
/* find '".force_encoding("' */
pos_before_middle_part = strseq_core(s, s_end, len,
force_encoding_middle_part, force_encoding_middle_part_len,
0, enc);
if (pos_before_middle_part <= 0) {
return UNDUMP_SOURCE_INVALID;
}
pos_after_middle_part = pos_before_middle_part + force_encoding_middle_part_len;
/* find '")' */
pos_before_end_part = strseq_core(s + pos_after_middle_part, s_end, len - pos_after_middle_part,
force_encoding_end_part, force_encoding_end_part_len,
0, enc);
if (pos_before_end_part < 0 || pos_after_middle_part + pos_before_end_part + 2 != len) {
return UNDUMP_SOURCE_INVALID;
}
*forced_enc_str_len = pos_before_end_part;
*forced_enc_str = rb_str_new(s + pos_after_middle_part, *forced_enc_str_len);
return UNDUMP_SOURCE_FORCE_ENCODING;
}
static int static int
unescape_ascii(unsigned int c) unescape_ascii(unsigned int c)
{ {
@ -6160,10 +6106,10 @@ unescape_ascii(unsigned int c)
} }
} }
static long static void
undump_after_backslash(VALUE undumped, const char *s, const char *s_end, rb_encoding **penc) undump_after_backslash(VALUE undumped, const char **ss, const char *s_end, rb_encoding **penc, bool *utf8)
{ {
const char *s0 = s; const char *s = *ss;
unsigned int c; unsigned int c;
int codelen; int codelen;
size_t hexlen; size_t hexlen;
@ -6190,6 +6136,7 @@ undump_after_backslash(VALUE undumped, const char *s, const char *s_end, rb_enco
s++; s++;
break; break;
case 'u': case 'u':
*utf8 = true;
if (++s >= s_end) { if (++s >= s_end) {
rb_raise(rb_eRuntimeError, "invalid Unicode escape"); rb_raise(rb_eRuntimeError, "invalid Unicode escape");
} }
@ -6256,7 +6203,7 @@ undump_after_backslash(VALUE undumped, const char *s, const char *s_end, rb_enco
s++; s++;
} }
return s - s0 + 1; *ss = s;
} }
static VALUE rb_str_is_ascii_only_p(VALUE str); static VALUE rb_str_is_ascii_only_p(VALUE str);
@ -6276,13 +6223,9 @@ str_undump(VALUE str)
{ {
const char *s = RSTRING_PTR(str); const char *s = RSTRING_PTR(str);
const char *s_end = RSTRING_END(str); const char *s_end = RSTRING_END(str);
long len = RSTRING_LEN(str), n; rb_encoding *enc = rb_enc_get(str);
rb_encoding *enc = rb_enc_get(str), *forced_enc = NULL;
unsigned int c;
enum undump_source_format source_format;
VALUE undumped = rb_enc_str_new(s, 0L, enc); VALUE undumped = rb_enc_str_new(s, 0L, enc);
VALUE forced_enc_str; bool utf8 = false;
long forced_enc_str_len;
int w; int w;
rb_must_asciicompat(str); rb_must_asciicompat(str);
@ -6292,52 +6235,78 @@ str_undump(VALUE str)
if (!str_null_check(str, &w)) { if (!str_null_check(str, &w)) {
rb_raise(rb_eRuntimeError, "string contains null byte"); rb_raise(rb_eRuntimeError, "string contains null byte");
} }
if (RSTRING_LEN(str) < 2) goto invalid_format;
source_format = check_undump_source_format(s, s_end, len, enc, if (*s != '"') goto invalid_format;
&forced_enc_str, &forced_enc_str_len);
if (source_format == UNDUMP_SOURCE_INVALID) {
rb_raise(rb_eRuntimeError, "not wrapped with '\"' nor '\"...\".force_encoding(\"...\")' form");
}
if (source_format == UNDUMP_SOURCE_FORCE_ENCODING) {
forced_enc = rb_find_encoding(forced_enc_str);
if (forced_enc == NULL) {
rb_raise(rb_eRuntimeError, "unknown encoding name - %"PRIsVALUE, forced_enc_str);
}
}
/* strip '"' at the start */ /* strip '"' at the start */
s++; s++;
if (source_format == UNDUMP_SOURCE_SIMPLE) {
/* strip '"' at the end */ for (;;) {
s_end--; if (s >= s_end) {
} else { /* source_format == UNDUMP_SOURCE_FORCE_ENCODING */ rb_raise(rb_eRuntimeError, "unterminated dumped string");
/* strip '".force_encoding("...")' */
s_end -= rb_strlen_lit("\".force_encoding(\"\")") + forced_enc_str_len;
} }
for (; s < s_end; s += n) { if (*s == '"') {
c = rb_enc_codepoint_len(s, s_end, &w, enc); /* epilogue */
if (c == '\\') { s++;
if (s+1 >= s_end) { if (s == s_end) {
rb_raise(rb_eRuntimeError, "invalid escape"); /* ascii compatible dumped string */
} break;
n = undump_after_backslash(undumped, s+1, s_end, &enc);
}
else if (c == '"') {
rb_raise(rb_eRuntimeError, "non-escaped double quote detected");
} }
else { else {
n = w; const char *encname;
rb_str_cat(undumped, s, n); char *buf;
int encidx;
ptrdiff_t size;
if (utf8) {
rb_raise(rb_eRuntimeError, "dumped string contained Unicode escape but used force_encoding");
}
size = rb_strlen_lit(".force_encoding(\"");
if (s_end - s <= size) goto invalid_format;
if (memcmp(s, ".force_encoding(\"", size) != 0) goto invalid_format;
s += size;
encname = s;
s = memchr(s, '"', s_end-s);
size = s - encname;
if (!s) goto invalid_format;
if (size > 100) {
rb_raise(rb_eRuntimeError, "dumped string has unknown encoding name");
}
buf = ALLOC_N(char, size+1);
memcpy(buf, encname, size);
buf[size] = '\0';
encidx = rb_enc_find_index(buf);
if (encidx < 0) {
rb_raise(rb_eRuntimeError, "dumped string has unknown encoding name");
}
rb_enc_associate_index(undumped, encidx);
if (s_end - s != 2 ||
s[0] != '"' ||
s[1] != ')') goto invalid_format;
}
break;
}
if (*s == '\\') {
s++;
if (s >= s_end) {
rb_raise(rb_eRuntimeError, "invalid escape");
}
undump_after_backslash(undumped, &s, s_end, &enc, &utf8);
}
else {
rb_str_cat(undumped, s++, 1);
} }
} }
if (source_format == UNDUMP_SOURCE_FORCE_ENCODING) {
rb_enc_associate(undumped, forced_enc);
ENC_CODERANGE_CLEAR(undumped);
}
OBJ_INFECT(undumped, str); OBJ_INFECT(undumped, str);
return undumped; return undumped;
invalid_format:
rb_raise(rb_eRuntimeError, "invalid dumped string; not wrapped with '\"' nor '\"...\".force_encoding(\"...\")' form");
} }
static void static void

10
test/ruby/test_string.rb
View File

@ -769,6 +769,9 @@ CODE
assert_equal(S(""), S('"\\u{}"').undump) assert_equal(S(""), S('"\\u{}"').undump)
assert_equal(S(""), S('"\\u{ }"').undump) assert_equal(S(""), S('"\\u{ }"').undump)
assert_equal(S("\u3042".encode("sjis")), S('"\x82\xA0"'.force_encoding("sjis")).undump)
assert_equal(S("\u8868".encode("sjis")), S("\"\\x95\\\\\"".force_encoding("sjis")).undump)
assert_equal(S("äöü"), S('"\u00E4\u00F6\u00FC"').undump) assert_equal(S("äöü"), S('"\u00E4\u00F6\u00FC"').undump)
assert_equal(S("äöü"), S('"\xC3\xA4\xC3\xB6\xC3\xBC"').undump) assert_equal(S("äöü"), S('"\xC3\xA4\xC3\xB6\xC3\xBC"').undump)
@ -782,10 +785,17 @@ CODE
assert_raise(RuntimeError) { S('\u3042').undump } assert_raise(RuntimeError) { S('\u3042').undump }
assert_raise(RuntimeError) { S('"".force_encoding()').undump } assert_raise(RuntimeError) { S('"".force_encoding()').undump }
assert_raise(RuntimeError) { S('"".force_encoding("').undump }
assert_raise(RuntimeError) { S('"".force_encoding("UNKNOWN")').undump } assert_raise(RuntimeError) { S('"".force_encoding("UNKNOWN")').undump }
assert_raise(RuntimeError) { S('"\u3042".force_encoding("UTF-16LE")').undump }
assert_raise(RuntimeError) { S('"\x00\x00".force_encoding("UTF-16LE")"').undump }
assert_raise(RuntimeError) { S('"\x00\x00".force_encoding("'+("a"*9999999)+'")"').undump }
assert_raise(RuntimeError) { S(%("\u00E4")).undump } assert_raise(RuntimeError) { S(%("\u00E4")).undump }
assert_raise(RuntimeError) { S('"').undump }
assert_raise(RuntimeError) { S('"""').undump }
assert_raise(RuntimeError) { S('""""').undump } assert_raise(RuntimeError) { S('""""').undump }
assert_raise(RuntimeError) { S('"a').undump }
assert_raise(RuntimeError) { S('"\u"').undump } assert_raise(RuntimeError) { S('"\u"').undump }
assert_raise(RuntimeError) { S('"\u{"').undump } assert_raise(RuntimeError) { S('"\u{"').undump }
assert_raise(RuntimeError) { S('"\u304"').undump } assert_raise(RuntimeError) { S('"\u304"').undump }