[Bug #19784] Fix behaviors against prefix with broken encoding

- String#start_with?
- String#delete_prefix
- String#delete_prefix!
This commit is contained in:
Nobuyoshi Nakada 2023-08-25 00:52:53 +09:00
parent 808b067088
commit b054c2fe06
Notes: git 2023-08-25 23:58:22 +00:00
4 changed files with 59 additions and 19 deletions

View File

@ -7,12 +7,14 @@ describe "String#start_with?" do
it_behaves_like :start_with, :to_s it_behaves_like :start_with, :to_s
# Here and not in the shared examples because this is invalid as a Symbol # Here and not in the shared examples because this is invalid as a Symbol
it "does not check that we are not starting to match at the head of a character" do it "matches part of a character with the same part" do
"\xA9".should.start_with?("\xA9") # A9 is not a character head for UTF-8 "\xA9".should.start_with?("\xA9") # A9 is not a character head for UTF-8
end end
it "does not check we are matching only part of a character" do ruby_bug "#19784", ""..."3.3" do
"\xe3\x81\x82".size.should == 1 it "checks we are matching only part of a character" do
"\xe3\x81\x82".should.start_with?("\xe3") "\xe3\x81\x82".size.should == 1
"\xe3\x81\x82".should_not.start_with?("\xe3")
end
end end
end end

View File

@ -70,7 +70,9 @@ describe :start_with, shared: true do
$1.should be_nil $1.should be_nil
end end
it "does not check that we are not matching part of a character" do ruby_bug "#19784", ""..."3.3" do
"\xC3\xA9".send(@method).should.start_with?("\xC3") it "checks that we are not matching part of a character" do
"\xC3\xA9".send(@method).should_not.start_with?("\xC3")
end
end end
end end

View File

@ -10461,10 +10461,20 @@ rb_str_start_with(int argc, VALUE *argv, VALUE str)
return Qtrue; return Qtrue;
} }
else { else {
const char *p, *s, *e;
long slen, tlen;
rb_encoding *enc;
StringValue(tmp); StringValue(tmp);
rb_enc_check(str, tmp); enc = rb_enc_check(str, tmp);
if (RSTRING_LEN(str) < RSTRING_LEN(tmp)) continue; if ((tlen = RSTRING_LEN(tmp)) == 0) return Qtrue;
if (memcmp(RSTRING_PTR(str), RSTRING_PTR(tmp), RSTRING_LEN(tmp)) == 0) if ((slen = RSTRING_LEN(str)) < tlen) continue;
p = RSTRING_PTR(str);
e = p + slen;
s = p + tlen;
if (!at_char_boundary(p, s, e, enc))
continue;
if (memcmp(p, RSTRING_PTR(tmp), tlen) == 0)
return Qtrue; return Qtrue;
} }
} }
@ -10483,12 +10493,13 @@ static VALUE
rb_str_end_with(int argc, VALUE *argv, VALUE str) rb_str_end_with(int argc, VALUE *argv, VALUE str)
{ {
int i; int i;
char *p, *s, *e;
rb_encoding *enc;
for (i=0; i<argc; i++) { for (i=0; i<argc; i++) {
VALUE tmp = argv[i]; VALUE tmp = argv[i];
const char *p, *s, *e;
long slen, tlen; long slen, tlen;
rb_encoding *enc;
StringValue(tmp); StringValue(tmp);
enc = rb_enc_check(str, tmp); enc = rb_enc_check(str, tmp);
if ((tlen = RSTRING_LEN(tmp)) == 0) return Qtrue; if ((tlen = RSTRING_LEN(tmp)) == 0) return Qtrue;
@ -10498,7 +10509,7 @@ rb_str_end_with(int argc, VALUE *argv, VALUE str)
s = e - tlen; s = e - tlen;
if (!at_char_boundary(p, s, e, enc)) if (!at_char_boundary(p, s, e, enc))
continue; continue;
if (memcmp(s, RSTRING_PTR(tmp), RSTRING_LEN(tmp)) == 0) if (memcmp(s, RSTRING_PTR(tmp), tlen) == 0)
return Qtrue; return Qtrue;
} }
return Qfalse; return Qfalse;
@ -10516,12 +10527,17 @@ rb_str_end_with(int argc, VALUE *argv, VALUE str)
static long static long
deleted_prefix_length(VALUE str, VALUE prefix) deleted_prefix_length(VALUE str, VALUE prefix)
{ {
char *strptr, *prefixptr; const char *strptr, *prefixptr;
long olen, prefixlen; long olen, prefixlen;
rb_encoding *enc = rb_enc_get(str);
StringValue(prefix); StringValue(prefix);
if (is_broken_string(prefix)) return 0;
rb_enc_check(str, prefix); if (!is_broken_string(prefix) ||
!rb_enc_asciicompat(enc) ||
!rb_enc_asciicompat(rb_enc_get(prefix))) {
enc = rb_enc_check(str, prefix);
}
/* return 0 if not start with prefix */ /* return 0 if not start with prefix */
prefixlen = RSTRING_LEN(prefix); prefixlen = RSTRING_LEN(prefix);
@ -10531,6 +10547,19 @@ deleted_prefix_length(VALUE str, VALUE prefix)
strptr = RSTRING_PTR(str); strptr = RSTRING_PTR(str);
prefixptr = RSTRING_PTR(prefix); prefixptr = RSTRING_PTR(prefix);
if (memcmp(strptr, prefixptr, prefixlen) != 0) return 0; if (memcmp(strptr, prefixptr, prefixlen) != 0) return 0;
if (is_broken_string(prefix)) {
if (!is_broken_string(str)) {
/* prefix in a valid string cannot be broken */
return 0;
}
const char *strend = strptr + olen;
const char *after_prefix = strptr + prefixlen;
if (!at_char_boundary(strptr, after_prefix, strend, enc)) {
/* prefix does not end at char-boundary */
return 0;
}
}
/* prefix part in `str` also should be valid. */
return prefixlen; return prefixlen;
} }
@ -10587,7 +10616,7 @@ rb_str_delete_prefix(VALUE str, VALUE prefix)
static long static long
deleted_suffix_length(VALUE str, VALUE suffix) deleted_suffix_length(VALUE str, VALUE suffix)
{ {
char *strptr, *suffixptr, *s; const char *strptr, *suffixptr;
long olen, suffixlen; long olen, suffixlen;
rb_encoding *enc; rb_encoding *enc;
@ -10602,9 +10631,10 @@ deleted_suffix_length(VALUE str, VALUE suffix)
if (olen < suffixlen) return 0; if (olen < suffixlen) return 0;
strptr = RSTRING_PTR(str); strptr = RSTRING_PTR(str);
suffixptr = RSTRING_PTR(suffix); suffixptr = RSTRING_PTR(suffix);
s = strptr + olen - suffixlen; const char *strend = strptr + olen;
if (memcmp(s, suffixptr, suffixlen) != 0) return 0; const char *before_suffix = strend - suffixlen;
if (!at_char_boundary(strptr, s, strptr + olen, enc)) return 0; if (memcmp(before_suffix, suffixptr, suffixlen) != 0) return 0;
if (!at_char_boundary(strptr, before_suffix, strend, enc)) return 0;
return suffixlen; return suffixlen;
} }

View File

@ -1938,6 +1938,8 @@ CODE
assert_send([S("hello"), :start_with?, S("hel")]) assert_send([S("hello"), :start_with?, S("hel")])
assert_not_send([S("hello"), :start_with?, S("el")]) assert_not_send([S("hello"), :start_with?, S("el")])
assert_send([S("hello"), :start_with?, S("el"), S("he")]) assert_send([S("hello"), :start_with?, S("el"), S("he")])
assert_send([S("\xFF\xFE"), :start_with?, S("\xFF")])
assert_not_send([S("\u{c4}"), :start_with?, S("\xC3")])
bug5536 = '[ruby-core:40623]' bug5536 = '[ruby-core:40623]'
assert_raise(TypeError, bug5536) {S("str").start_with? :not_convertible_to_string} assert_raise(TypeError, bug5536) {S("str").start_with? :not_convertible_to_string}
@ -2930,6 +2932,7 @@ CODE
assert_equal("\x95\x5c".force_encoding("Shift_JIS"), s.delete_prefix("\x95")) assert_equal("\x95\x5c".force_encoding("Shift_JIS"), s.delete_prefix("\x95"))
assert_equal("\x95\x5c".force_encoding("Shift_JIS"), s) assert_equal("\x95\x5c".force_encoding("Shift_JIS"), s)
assert_equal("\xFE", S("\xFF\xFE").delete_prefix("\xFF"))
end end
def test_delete_prefix_clear_coderange def test_delete_prefix_clear_coderange
@ -2978,6 +2981,9 @@ CODE
assert_equal(nil, s.delete_prefix!("\xe3")) assert_equal(nil, s.delete_prefix!("\xe3"))
assert_equal("\xe3\x81\x82", s) assert_equal("\xe3\x81\x82", s)
s = S("\xFF\xFE")
assert_equal("\xFE", s.delete_prefix!("\xFF"))
assert_equal("\xFE", s)
end end
def test_delete_prefix_bang_clear_coderange def test_delete_prefix_bang_clear_coderange