From: rafaelmfranca@...
Date: 2014-04-02T22:21:37+00:00
Subject: [ruby-core:61812] [ruby-trunk - Bug #9694] Bad regexp hangs ruby

Issue #9694 has been updated by Rafael Fran��a.


I tried to reproduce with your script and could not:

```
$ ruby -v
ruby 2.1.1p76 (2014-02-24 revision 45161) [x86_64-darwin13.0]

$ cat foo.rb
str = ('a' * ARGV[0].to_i) + '?'
re = /(\w)$/
re.match(str)

$ time ruby foo.rb 14

real	0m0.134s
user	0m0.075s
sys	0m0.050s

$ time ruby foo.rb 15

real	0m0.145s
user	0m0.082s
sys	0m0.054s

$ time ruby foo.rb 16

real	0m0.137s
user	0m0.076s
sys	0m0.052s

$ time ruby foo.rb 50

real	0m0.142s
user	0m0.080s
sys	0m0.053s
```

----------------------------------------
Bug #9694: Bad regexp hangs ruby
https://bugs.ruby-lang.org/issues/9694#change-46045

* Author: Nikolay Markov
* Status: Open
* Priority: Normal
* Assignee: 
* Category: regexp
* Target version: 
* ruby -v: ruby 2.1.1p76 (2014-02-24 revision 45161) [x86_64-darwin13.0]
* Backport: 2.0.0: UNKNOWN, 2.1: UNKNOWN
----------------------------------------
Here is an extracted problem i ran into recently:
~~~
$ cat test.rb

str = ('a' * ARGV[0].to_i) + '?'
re = /(\w*)*$/
re.match(str)
~~~

On few chars match returns quite fast, but here's what happens on 14 'a'-s and up:
~~~
$ time RBENV_VERSION=2.1.1 ruby test.rb 14

real 1.392	user 1.364	sys 0.026	pcpu 99.83

$ time RBENV_VERSION=2.1.1 ruby test.rb 15

real 3.979	user 3.949	sys 0.026	pcpu 99.89

$ time RBENV_VERSION=2.1.1 ruby test.rb 16

real 11.995	user 11.954	sys 0.031	pcpu 99.92
~~~

Ruby versions 1.9.3 and 2.0 behave similarly. 

I ran into the problem, because one of my colleagues copy-pasted this regexp to test url's somewhere from stackoverflow:
/^(https?:\/\/)?([\da-z\.-]+)\.([a-z\.]{2,6})([\/\w \.-:]*)*\/?$/

I know the regexp is useless, however i think it's still a problem if a bad regexp can hang ruby.
Python (2.7) says that this regexp is bad and does not compile it.
Perl matches without any performance issues





-- 
https://bugs.ruby-lang.org/