Btukfyl

It is really simple. I am trying to use regex to indentify certain property values in a line of a vcard string.

So, here's the code:

int main(string args){



string input = "TEL;VALUE=uri;PREF=1;TYPE="voice,home":tel:+1-555-555-5555;ext=5555";







string regString = "(tel:(?<phnum>.*);)*(?<pref>PREF=1;)*";



Regex regex = new Regex(regString);

MatchInfo match;



regex.match(input_end, 0, out match);



stdout.printf(match.fetch_named("phnum"));



stdout.printf(match.fetch_named(pref));



return 0;

}

What i want to do, really, is to say that the tag phnum applies to a subgroup of characters, when it appears in the input (hence the *, at the outside of the group). so when i do: match.fetch_named("phnum") the returned value would be: "+1-555-555-5555".

I am just getting segmentation faults, even though regex tester apps seem to accept the pattern well enough.

There are a number of things that can be done to improve the Vala code:

• GLib's Regex binding to PCRE will return an error message giving some details about an invalid regular expression. In Vala this message can be read by putting new Regex () in a try...catch block.


• 
  regex.match() returns true when a match is found, so wrapping regex.match() in an if statement makes the program more robust


• Vala has the null coalescing operator, ??, which is a convenient way of providing an alternative value when there is a null value


• 
  MatchInfo has the next() method and when combined with Vala's do {} when () loop gives a good way of retrieving multiple matches safely

The regex you are using needs to exclude the terminating character, ;. So tel:(?<phnum>[^;|.]*); would match all characters excluding ; after tel: until ; is reached.

Here is a working example putting all that together:

int main () {



  string input = "TEL;VALUE=uri;PREF=1;TYPE="voice,home":tel:+1-555-555-5555;ext=5555";



  string regString = "tel:(?<phnum>[^;|.]*);|PREF=(?<pref>[0-9]*);";

  Regex regex;

  MatchInfo match;

  try {

    regex = new Regex(regString);

    if (regex.match(input, 0, out match)) {

      do {

        stdout.printf("Phone number: %sn", match.fetch_named("phnum") ?? "None");

        stdout.printf("Preference: %sn", match.fetch_named("pref") ?? "None");

      }

      while (match.next());

    }

  }

  catch (Error error) {

    print (@"$(error.message)n");

    return 1;

  }



  return 0;

}

This outputs:

Phone number: 

Preference: 1

Phone number: +1-555-555-5555

Preference: None

There are two matches. What is interesting is the first match returns empty string for the phone number. This is because it is a valid sub-pattern, but didn't match anything. Why pref is null for the second match is unclear to me. That needs some more investigation as to what is going on in the regex engine, but this hopefully gives you enough to get on with.

printf takes a format string first. You need to change those to:

stdout.printf("%sn", match.fetch_named("phnum"));



stdout.printf("%sn", match.fetch_named("pref"));

If the format string is null, printf will segfault.

If you don't want to bother with a format string, you can use FileStream.puts, but you still need a null check:

if (match.fetch_named("phnum")!=null)

  stdout.puts(match.fetch_named("phnum"));

In think regex.match(input_end should also be regex.match(input

Without taking the exact format of the phone number into account, one possible solution could be to match the allowed characters in a character class and get the value from the group named phnum

tel:(?<phnum>[0-9+-]+)

Regex demo | Test it online

Or a bit more broader match could be to use a negated character class [^ to match what you don't want and get the value from group named phnum

tel:(?<phnum>[^rn;]+)

Regex demo

Your regex should be shortened to:

tel:(?<phnum>.*);$

Now 'phnum' Group will contain the phone number.