|
Boost Users : |
Subject: Re: [Boost-users] [xpressive] Accessing named capture from another scope in the semantic action
From: Eric Niebler (eniebler_at_[hidden])
Date: 2014-01-19 18:14:09
On 1/19/2014 11:23 AM, Roger wrote:
> Hi!
>
> Is it possible access captured group from another scope in the
> semantic action? Can I write a lazy function somehow and access
> another scope? The following code results in bad_lexical_cast
> exception, likely due to the named capture "Year" is an empty
> string in the "outer scope".
>
> const sregex yyyymmdd = (Year=_d>>_d>>_d>>_d) >> (Month= _d>>_d)
>>> (Day= _d>>_d);
>
> sregex rx = yyyymmdd [ tm_year(t) = as<int>(Year) - 1900 /*,
> */ ] // bad lexical cast: source type value could not be
> interpreted as target
>>> hhmmss [ tm_hour(t) = as<int>(Hour) /*,
*/ ]
> ;
>
> In my case the regex has no recursion, no captured repeated groups,
> so I can't understand really why I got nested result set.
A nested regex *always* results in nested match results. Xpressive
can't know in general that the regex currently being defined is
non-recursive, and any other behavior would lead to inconsistencies.
> Dumping the nested result tree indicates that the values are
> captured in another scope. But this is in the match_result<> set,
> and I can't figure out if there is any equivalent set available
> inside the semantic action.
>
> 20140119190000 #rx day = hour = minute = month = second = year =
> 20140119 #yyyymmdd day = 19 hour = minute = month = 01 second =
> year = 2014 190000 #hhmmss day = hour = 19 minute = 00 month =
> second = 00 year =
>
> Im asking mostly for educational reason. Thanks for any hints!
I don't think there's a general way to do what you're trying to do.
That's a bummer. But since your regex is non-recursive, you can use
the following hack to flatten your regex out and get the behavior you
want:
const auto yyyymmdd = boost::proto::deep_copy(
(Year=_d>>_d>>_d>>_d) >> (Month= _d>>_d) >> (Day= _d>>_d));
const auto hhmmss = boost::proto::deep_copy(
(Hour=_d>>_d)>>(Minute=_d>>_d)>>(Second=_d>>_d));
Notice the "auto" and the "deep_copy".
HTH,
-- Eric Niebler Boost.org http://www.boost.org
Boost-users list run by williamkempf at hotmail.com, kalb at libertysoft.com, bjorn.karlsson at readsoft.com, gregod at cs.rpi.edu, wekempf at cox.net