Re: [Boost-bugs] [Boost C++ Libraries] #8261: lexical_cast<unsigned> returns unexpected result when using with split_iterator<std::wstring::iterator>

Subject: Re: [Boost-bugs] [Boost C++ Libraries] #8261: lexical_cast<unsigned> returns unexpected result when using with split_iterator<std::wstring::iterator>
From: Boost C++ Libraries (noreply_at_[hidden])
Date: 2014-02-02 07:35:14

#8261: lexical_cast<unsigned> returns unexpected result when using with
split_iterator<std::wstring::iterator>
-------------------------------+--------------------------
  Reporter:  s.a.moreno.a.s@…  |      Owner:  apolukhin
      Type:  Bugs              |     Status:  new
 Milestone:  To Be Determined  |  Component:  lexical_cast
   Version:  Boost 1.52.0      |   Severity:  Problem
Resolution:                    |   Keywords:
-------------------------------+--------------------------

Comment (by apolukhin):

 Replying to [comment:4 Troy Korjuslommi <troykor@…>]:
 > I didn't get a failure on any of the tests on 1_55_0.
 > Must a dev branch new test or something.

 Just out of curiosity downloaded the 1.55. Test is in there:
 `boost_1_55_0/libs/conversion/lexical_cast_test.cpp`. With your patches
 test `void test_getting_pointer_to_function()` fail to compile. Breaking
 existing use cases is unfordable.

 > My understanding of the problem is as follows.
 > The eventual problem is that the code interprets wchar_t* as char*.
 > The reason this happens is that the traits template says the char type
 for CharT is char, not wchar_t. The reason it gets it wrong is that it
 doesn't have any checks for iterator_range types, so it just fumbles.
 > It catches those iterator_range types which have a parameter char or
 wchar_t, but iterator_range<iterator/const_iterator> types have one more
 level of embedded types, so it misses those.
 >
 > One could add those checks, but that would just add complexity without
 any real benefit. A specialization for iterator_range would just get added
 somewhere else anyway. It is much simpler and safer to catch all
 iterator_range types in their own lexical_cast function.
 > If something breaks, then write more special cases.
 > I am attaching another patch which adds another level of indirection, so
 specializations can be written as needed.
 >
 > I know this is slightly different take on what you mentioned. I do see
 your point too with the stream types. I ended up taking a different route.
 Please give it some thought.

 Your approach can be used with developer branch. `try_lexical_convert`
 function does not guarantee unambiguity while getting pointer to function,
 so it can be specialized in your way. Fork the
 [https://github.com/boostorg/conversion conversion repo], apply your
 patch, add tests to ensure that new feature work and make sure that
 '''all''' tests pass.

 > Iter <wchar_t*, std::basic_string<wchar_t,Traits,Alloc> >

 This is not a portable solution (it won't work with SCARY iterators).

-- 
Ticket URL: <https://svn.boost.org/trac/boost/ticket/8261#comment:6>
Boost C++ Libraries <http://www.boost.org/>
Boost provides free peer-reviewed portable C++ source libraries.