seems there is a bug in PowerGREP?

[Затлер Евгений Викторович]

Hello,

it seems there is a bug in powerGREP.
i tried version 3.4.2
I am searching for

(\d+\b).*?\1

through a line of comma separated values:

12, 22, 3, 45, 3, 36 

Here is what i get in PowerGREP

12, 22, 3, 45, 3, 36 

(if you use html enabled mailer you should see the colored matches)

While I agree with "3, 45, 3" match , I definitely can not agree with the first match "2, 2"

The capturing group includes a boundary \b that matches between the first "2" and a comma

And the backreference \1 should as well account for that boundary.

In my target string, there is no boundary here

22

 ^

between the neughbouring "twos"

Do you agree that it is a bug?

==================================

P.S.

if i change the regexp to (\d+\b).*\1

PowerGREP reports that "3, 45, 3, 3" matched. Again it considers there is a boundary inside "36".

And that is strange.