Opened 16 months ago

Last modified 16 months ago

#13231 new Bugs

u32regex_replace() for std::u32string (UTF-32) won't compile

Reported by: robic@… Owned by: John Maddock
Milestone: To Be Determined Component: regex
Version: Boost 1.64.0 Severity: Showstopper
Keywords: Cc:


Note - This pertains to input UTF-32 strings.

Compiler: VS2015 , VC++ Boost version: 1.64

The u32regex_search() works fine and as expected


Problem description: Intellisence detects proper input
parameters using the u32string parameter, but throws many
errors when compiling.

The errors start with

boost_1_64_0\boost/functional/hash/extensions.hpp(262): error C2664: 'size_t boost::hash_value(std::type_index)': cannot convert argument 1 from 'const char32_t' to 'std::type_index'

and end with

RxReplace?.cpp(9255): note: see reference to function template instantiation 'std::basic_string<char32_t,std::char_traits<char32_t>,std::allocator<char32_t>> boost::u32regex_replace<char32_t>(const std::basic_string<char32_t,std::char_traits<char32_t>,std::allocator<char32_t>> &,const boost::u32regex &,const std::basic_string<char32_t,std::char_traits<char32_t>,std::allocator<char32_t>> &,boost::regex_constants::match_flag_type)' being compiled

with about 20 errors in between.

I hope this is just a problem with extensions.hpp
and maybe there is a quick fix (or it has already been fixed).
I desperately need this to work with std::u32string types
as UnicodeString? just won't cut it for my purposes.

From the documentation:

For each regex_replace algorithm defined by
<boost/regex.hpp>, then <boost/regex/icu.hpp> defines an overloaded algorithm that takes the same arguments, but which is called u32regex_replace, and which will accept UTF-8, UTF-16 or UTF-32 encoded data, as well as an ICU UnicodeString? as input.

The code below reproduces the compile errors

	// rxconst.h
	// ==============================

	typedef std::u32string   X_32string;
	typedef std::wstring     X_string;
	typedef boost::u32regex  U_X_regex;
	typedef boost::u32match  X_u32match;
	typedef std::u32string::const_iterator U32SITR;

	#define U_MAKEREGEX(str,options) make_u32regex(str,options)
	#define U_REGEX_SEARCH boost::u32regex_search
	#define U_REGEX_REPLACE boost::u32regex_replace

	// RxReplace.cpp ( won't compile )
	// ==============================
	#include "rxconst.h";

	U_X_regex uRx = U_MAKEREGEX( _T("[trgt]+"), regex_constants::perl );
	X_32string sTarget, sFmt, sOutput;
	WstrToU32string( X_string(_T("target")), sTarget ); 
	WstrToU32string( X_string(_T("")), sFmt ); 
	sOutput =  U_REGEX_REPLACE( sTarget, uRx, sFmt );	// <-- Won't compile 

	// RxSearch.cpp ( compiles and runs as expected )
	// ==============================
	#include "rxconst.h";

	U_X_regex uRx = U_MAKEREGEX( _T("[abce]+"), regex_constants::perl );
	X_32string sSrc;
	WstrToU32string( X_string(_T("source")), sSrc ); 
	U32SITR SitrStart = sSrc.begin();
	U32SITR start = SitrStart;
	U32SITR end   = sSrc.end();

	X_u32match _M;

	while ( U_REGEX_SEARCH( start, end, _M, uRx, Flags, SitrStart ) )
		// ...

Change History (3)

comment:1 Changed 16 months ago by anonymous

Component: Noneregex
Owner: set to John Maddock

comment:2 Changed 16 months ago by anonymous

Severity: ProblemShowstopper

comment:3 Changed 16 months ago by robic@…

This looks to be fixed in release 1.65.1

I've tested the above code with this release and it works fine.

Details where it was fixed:

Chapter 15. Boost.Functional/Hash
Change log:
Boost 1.65.0 
• Support for char16_t, char32_t, u16string, u32string 




Note: See TracTickets for help on using tickets.