Re: Inconsistent regexp matching when using quantifiers?

From: Dean (dean_at_[hidden])
Date: 2003-04-25 12:35:06

--- In Boost-Users_at_[hidden], "Joshua B. Smith" <josh_at_n...> 
wrote:
> On Thu, Apr 24, 2003 at 11:37:34PM -0000, Dean wrote:
> > Hi all,
> >
> > ... snip ...
> > 
> > \d{3}-\d{2}-\d{4}
> > 
> > As expected, that pattern was found in "123-12-1234" but not 
in "1234-
> > 12-1234".  However it *was* found in "1234567-12-1234".
> > 
> > Is this behavior by design or is it a bug?
> 
> It is (probably) design.  Intervals can specify a min and a max, 
for example:
> 
> \d{3,3}-\d{2,2}-\d{4,4}  will match "123-12-1234" but NOT "1234567-
12-1234".
> \d{3}-\d{2}-\d{4} will match "123-12-1234" but NOT "1234567-12-
1234" also.

I'm not sure what you were trying to say above, but my understanding 
is that the 2 patterns you just mentioned are equivalent.  The docs 
say "{3}" is equivalent to "{3,3}" not "{3,}".  

> 
> It will, however, yeild a correct search (there is a difference 
between search
> and match).  You didn't mention if you were doing a regex_search or
> regex_match ?

I'm doing a search because I don't want to know whether the whole 
string matches but whether the regex is found in the string.  
Specifically, I'm doing:

m_regex.Search( sampleBody, boost::match_default | boost::match_any)

While I can believe that the design intention was that "\d{3}-" 
should be found in "1234567-" (at the fifth character), it seems 
inconsistent that it is *not* also found in "123456-" and "12345678-
".  I'm seeing that inconsistent behavior.

> > FWIW, it's easy enough for me to workaround the current behavior 
with 
> > a pattern like this:
> > 
> > (^|[^a])a{1}b
> 
> You could use this, but I wouldn't recomend it (but that's just 
me...regex
> construction is deeply personal :)   ).  HTH.

I realize there is more than one way to do it, and I'd be interested 
in what you'd recommend.

FWIW, in our SSN-matching case, we'll probably just use "\b\d{3}-\d
{2}-\d{4}\b".

Thanks for the reply!

--Dean