std::regex undefined behavior

Question

Сould anyone please describe such behavior of std::regex library

string a{"ERROR"};
regex r1{"errOR",0};
cout<<regex_search(a,r1)<<endl;
regex r2{"errOR"};
cout<<regex_search(a,r2)<<endl;
regex r3{"errOR",regex::ECMAScript};
cout<<regex_search(a,r3)<<endl;

cout<<r1.flags()<<endl;
cout<<r2.flags()<<endl;
cout<<r3.flags()<<endl;

gives output

so 1st example became implicitly ignore casing with sum of flags corresponding to default construction of re i.e 16, by the way in std::regex there is no constant that match 0 value, but there is regex::icase == 1;

Is that conceived behavior of std library or I should not to feed to constructor values that explicitly not supported?

I'm not sure if it is documented what happens if you don't use a proper flag but you should assume nothing good will happen. — NathanOliver
– NathanOliver, Commented Feb 26, 2019 at 21:08
the values of the flags are unspecified (see eg here), but I doubt there is undefined behaviour, anyhow of course you should not pass parameter when they are "explicitily not supported" — 463035818_is_not_an_ai
– 463035818_is_not_an_ai, Commented Feb 26, 2019 at 21:09
"by the way in std::regex there is no constant that match 0 value, but there is regex::icase == 1;" where did you take this from? — 463035818_is_not_an_ai
– 463035818_is_not_an_ai, Commented Feb 26, 2019 at 21:11

J. Calleja · Accepted Answer · 2019-02-28 18:48:32Z

3

std::regex has several constructors. Two of them are:

explicit basic_regex (const charT* str, flag_type flags = ECMAScript);
basic_regex (const charT* str, size_t len, flag_type flags = ECMAScript);

The first constructor allows to construct the regex using a null terminated string. The second constructor allows to construct the regex using an array of characters and the array length.

The line: regex r1{"errOR",0}; uses the second one. So, you are not using zero for the flags but for the string length. As regex_search is able to match your empty string, it returns true.

You can modify your experiment to force the first constructor using:

regex r1{ "errOR", regex_constants::syntax_option_type(0) };

The standard states that:

A valid value of type syntax_option_type shall have at most one of the grammar elements ECMAScript, basic, extended, awk, grep, egrep, set. If no grammar element is set, the default grammar is ECMAScript.

So, now it is using ECMASScript and value returned by regex_search is false.

Regarding your last question, I would try to not use an unsupported parameter, because the standandard also states [defns.undefined]:

Undeﬁned behavior may be expected when this International Standard omits any explicit deﬁnition of behavior or when a program uses an erroneous construct or erroneous data.

answered Feb 28, 2019 at 18:48

J. Calleja

4,9352 gold badges36 silver badges57 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

DSU Over a year ago

Thank you very much for such detailed answer!

Collectives™ on Stack Overflow

std::regex undefined behavior

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related