mismatched paramters in error messages are reported as expressions and not types

[Jonathan Wakely]

On 21 September 2011 12:08, Křištof Želechovski wrote:
> g++ (4.6) recently threw at me the following error message:
>
> doit.cpp:290:129: error: no match for ‘operator<<’ in ‘std::cout << boost::range::equal_range [with ForwardRange = boost::range_detail::transform_range<boost::value_factory<ls::file_rec<char>::as_char0>, std::vector<ls::file_rec<char> > >, Value = int, typename boost::range_iterator<const ForwardRange>::type = boost::transform_iterator<boost::value_factory<ls::file_rec<char>::as_char0>, __gnu_cxx::__normal_iterator<ls::file_rec<char>*, std::vector<ls::file_rec<char> > >, boost::use_default, boost::use_default>]((*(const boost::range_detail::transform_range<boost::value_factory<ls::file_rec<char>::as_char0>, std::vector<ls::file_rec<char> > >*)(& boost::range_detail::operator| [with InputRng = std::vector<ls::file_rec<char> >, UnaryFunction = boost::value_factory<ls::file_rec<char>::as_char0>]((* & lv_fr), (*(const boost::range_detail::transform_holder<boost::value_factory<ls::file_rec<char>::as_char0> >*)(& boost::adaptors::{anonymous}::transformed.boost::range_detail::forwarder<Holder>::operator() [with T = boost::value_factory<ls::file_rec<char>::as_char0>, Holder = boost::range_detail::transform_holder]((boost::value_factory<ls::file_rec<char>::as_char0>(), boost::value_factory<ls::file_rec<char>::as_char0>()))))))), (* &((int)a_n.std::basic_string<_CharT, _Traits, _Alloc>::operator[] [with _CharT = char, _Traits = std::char_traits<char>, _Alloc = std::allocator<char>, std::basic_string<_CharT, _Traits, _Alloc>::reference = char&, std::basic_string<_CharT, _Traits, _Alloc>::size_type = long unsigned int](0ul))))’
>
> Am I suppposed to read and understand this?  It is overwhelming.
>
> The same message, somewhat split, is a bit more readable:
>
> doit.cpp:290:129:
> error: no match for ‘operator<<’ in ‘
> std::cout <<
> boost::range::equal_range
> [with
> ForwardRange = boost::range_detail::transform_range<boost::value_factory<ls::file_rec<char>::as_char0>, std::vector<ls::file_rec<char> > >,
> Value = int,
> typename boost::range_iterator<const ForwardRange>::type =
> boost::transform_iterator<
> boost::value_factory<ls::file_rec<char>::as_char0>, __gnu_cxx::__normal_iterator<ls::file_rec<char>*, std::vector<ls::file_rec<char> > >, boost::use_default, boost::use_default>]
> (
> (
> *
> (const boost::range_detail::transform_range<boost::value_factory<ls::file_rec<char>::as_char0>, std::vector<ls::file_rec<char> > >*)
> (
> & boost::range_detail::operator| [with InputRng = std::vector<ls::file_rec<char> >, UnaryFunction = boost::value_factory<ls::file_rec<char>::as_char0>]
> (
> (* & lv_fr),
> (
> *
> (const boost::range_detail::transform_holder<boost::value_factory<ls::file_rec<char>::as_char0> >*)
> (
> &
> boost::adaptors::{anonymous}::transformed.boost::range_detail::forwarder<Holder>::operator()
> [with T = boost::value_factory<ls::file_rec<char>::as_char0>, Holder = boost::range_detail::transform_holder]
> ((boost::value_factory<ls::file_rec<char>::as_char0>(), boost::value_factory<ls::file_rec<char>::as_char0>()))))))),
> (
> *
> &
> (
> (int)
> a_n.
> std::basic_string<_CharT, _Traits, _Alloc>::operator[]
> [with
> _CharT = char, _Traits = std::char_traits<char>, _Alloc = std::allocator<char>,
> std::basic_string<_CharT, _Traits, _Alloc>::reference = char&, std::basic_string<_CharT, _Traits, _Alloc>::size_type = long unsigned int](0ul))))’
>
> What we can see is that the error message does not tell me the types that do not match but the expressions that do not match, and while subexpressions are sometimes marked with types, the type of the mismatched argument is only mentioned in the following notes.  I mean, is it really necessary to dump the whole expression instead of just reporting the failing argument types?  Is there a switch to turn it off?


I don't think so, although -fno-pretty-templates can make it easier to
read in some cases.

Feel free to add comments or examples to my enhancement request about
this http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49152