Bug 2108 – regexp.d: The greedy dotstar isn't so greedy

As far as I'm aware, ".*" should be greedy by default and become non-greedy when changed to ".*?". As it stands now, both ".*" and ".*?" are non-greedy when it comes to std.regexp and I have found no way to make ".*" greedy, flags or otherwise. This can be seen by using "<packet>text</packet><packet>text</packet>" as the buffer to match against and "<packet.*/packet>" as the pattern. When I use this with std.regexp.search, it only matches the first opening and closing tag instead of the outer set. I just hope this isn't my lack of regex-fu coming back to haunt me.

*** Issue 2487 has been marked as a duplicate of this issue. ***

This is also an issue in Windows with std.regex using DMD 2.043 But I would like to add that it is always greedy prior to text. The first assert will fail since it was not non-greedy and the second is what it should be. import std.regex; void main() { assert(match("Hello there you silly person you.", regex(r"\b.+? you .+\w")).hit != "Hello there you silly"); assert(match("Hello there you silly person you.", regex(r"\b.+? you .+\w")).hit == "there you silly person"); }

(In reply to comment #2) > This is also an issue in Windows with std.regex using DMD 2.043 > > But I would like to add that it is always greedy prior to text. The first > assert will fail since it was not non-greedy and the second is what it should > be. > > import std.regex; > > void main() { > assert(match("Hello there you silly person you.", > regex(r"\b.+? you .+\w")).hit != "Hello there you silly"); > > assert(match("Hello there you silly person you.", > regex(r"\b.+? you .+\w")).hit == "there you silly person"); > } Actually it should be assert(match("Hello there you silly person you.", regex(r"\b.+? you .+\w")).hit == "Hello there you silly person you"); Two points - \b also matches at the begining of input (if the first char is \w), and the last .+ is greedy, and since '.' is certainly not a \w, we have what we have. Also tested at: http://www.regextester.com/ http://www.regular-expressions.info/javascriptexample.html ... etc. P.S. The patch is coming ;)

Reassigning to Dmitry.

I'd gladly close this issue, since it now works correctly in std.regex. But the report is filed against std.regexP. Should I close it?

Yes. Please also update the changelog.dd file.

Fixed for std.regex: https://github.com/D-Programming-Language/phobos/commit/9afb00e36b625322d7f1d8ec0fbd876c2b5c03fc