- Boost - lists.boost.org

Boost 1.32.0/SUN CC shared_ptr/weak_ptr test failure in shared_ptr_basic_test.cpp
by Timo Geusch 17 Dec '04

17 Dec '04

Ah yes, the neverending saga of the SUN C++ compiler, version 5.5 in this particular case. My apologies if this information is already well-known, I have been off the list for a couple of months, but a quick check of the mailing list archive didn't show any discussion regarding this subject. I had been trying to figure out what's breaking the simple shared_ptr to weak_ptr assignments and copies on this particular compiler. The code shown below is based on the code in the above file, around line 260. Anyway, for reference, the following test fails using the default settings for sunpro-tools.jam: weak_ptr<X> wp2 = static_pointer_cast<X>(p5); BOOST_TEST(wp2.use_count() == 1); as the reference count is 2, not 1 as expected. After looking at the source code for static_pointer_cast<> and a bit of experimentation with different types of scoping produced the following code instead: weak_ptr<X> wp2; { wp2 = static_pointer_cast<X>(p5); } BOOST_TEST(wp2.use_count() == 1); Interestingly, this code works as expected - it appears that this particular compiler doesn't clean up the temporary shared_ptr<X> that's being generated inside static_pointer_cast<> until it hits a closing '}'. Digging around the web confirmed that originally the SUN C++ compilers destroyed temporaries as described above and the newer compilers are still backward-compatible. In order to get standard-compliant behaviour one needs to specify -features=tmplife. One updated boost jamfile and the compiler appears to be much happier - all the smart_ptr appear to pass now, apart from shared_ptr_assign_fail.cpp which I assume isn't supposed compile in the first place. A quick test suggests that the 5.3 also supports the above compiler flag so it may be worth including this setting in the sunpro-tools.jam file, however I haven't tested this yet.

1 0

[multi_index] announce: added serialization support
by Joaquín Mª López Muñoz 16 Dec '04

16 Dec '04

The version currently commited to the CVS includes serialization support, docs, examples, etc. I'll be happy to know about your usage experience, as well as general comments on the new feature. Regards, Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

2 1

Re: [boost] thoughts on a GUI component library
by brangdon＠cix.compulink.co.uk 16 Dec '04

16 Dec '04

In-Reply-To: <BAY101-F20FDD5860C530B8D6ADE17A0AC0(a)phx.gbl> msclrhd(a)hotmail.com (Reece Dunn) wrote (abridged): > struct position{ float x; float y; }; > struct size{ float dx; float dy; }; > struct area > { > float top; float left; > float width; float height; > }; Why float rather than double? I'd want to avoid unnecessary conversions, so I'd want to use the same units all the way through. At the moment my document model uses integers, scaled not to 96 dpi pixels but to logical units which are 72,000 dpi. With 32-bit ints that gives a drawing area of about 30,000 inches. Using floats would give only 23 bits of precision, which reduces the drawing area to around 60 inches, which is getting a bit small. I'd prefer to avoid using fractional units because of exactness issues with 1/3 etc. I don't have a strong opinion on floating point generally. Sometimes I want to use pixel rounding, and sometimes I want to transform coordinates without losing accuracy. Adobe's PDF is an example which uses a floating point coordinates, albeit at 72 dpi rather than 72,000. I do know that most of the time, using integers works fine. You don't lose precision because you only ever transform the coordinates once. For bounding box transforms you round left/top with floor and right/bottom with ceil. For anti-aliasing you transform, eg, from 72,000 dpi to 384 dpi to get 1/4 pixel precision. -- Dave Harris, Nottingham, UK

1 0

MPL Book is on the way!!!
by Jody Hagins 16 Dec '04

16 Dec '04

I recieved notification that it shipped yesterday, and according to the UPS website, it will be here tomorrow, Dec 15!

6 6

[Boost-bugs] [ boost-Bugs-1086305 ] incorrect conditions to #define DISABLE_WIDE_CHAR_SUPPORT
by SourceForge.net 16 Dec '04

16 Dec '04

Bugs item #1086305, was opened at 2004-12-16 11:20 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=107586&aid=1086305&group_… Category: lexical_cast Group: None Status: Open Resolution: None Priority: 5 Submitted By: Sergej Zagursky (seza) Assigned to: Kevlin Henney (kevlin) Summary: incorrect conditions to #define DISABLE_WIDE_CHAR_SUPPORT Initial Comment: I think absence of BOOST_NO_INTRINSIC_WCHAR_T isn't a requirement for DISABLE_WIDE_CHAR_SUPPORT definition. Suggested fix: //old code #if defined(BOOST_NO_STRINGSTREAM) || defined(BOOST_NO_STD_WSTRING) || defined(BOOST_NO_STD_LOCALE) || defined(BOOST_NO_INTRINSIC_WCHAR_T) #define DISABLE_WIDE_CHAR_SUPPORT #endif //new code #if defined(BOOST_NO_STRINGSTREAM) || defined(BOOST_NO_STD_WSTRING) || defined(BOOST_NO_STD_LOCALE) #define DISABLE_WIDE_CHAR_SUPPORT #endif ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=107586&aid=1086305&group_… ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://productguide.itmanagersjournal.com/ _______________________________________________ Boost-bugs mailing list Boost-bugs(a)lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/boost-bugs

1 0

Cross Builder, v1.0
by John Torjo 16 Dec '04

16 Dec '04

Dear boosters, I thought this might be of interested in this, even if it's Windows-only for now. I'm happy to announce the first version of "Cross Builder": http://www.torjo.com/cb/ In short, Cross Builder allows you to build C++ projects very easy. In addition to that, it allows you to modify various compiler settings in a simple and straightforward way. Cross Builder's purpose is to make it as easy as possible for you to build your projects with *any* C++ compiler. Or, put in other words, *one click*. Compiler settings apply hierarchically. This is a very powerful mechanism, and once you see it at work, you'll simply love it! When you modify a setting, you modify it for: * a HDD directory or file * a configuration (such as "Debug", "Release", etc.) * a compiler (such as "All compilers", "MSVC" (all MS compilers), "VC 7.1", etc) From then on, this setting will apply to all: * it sub-configurations * its sub-compilers * its HDD sub-directories ...until it's overridden. Screenshots and possibility to download: http://www.torjo.com/cb/ As a side-node, in the near future, I will also post the source-code for the first version of CB. Feedback is welcome. Best, John -- John Torjo, Contributing editor, C/C++ Users Journal -- "Win32 GUI Generics" -- generics & GUI do mix, after all -- http://www.torjo.com/win32gui/ -- http://www.torjo.com/cb/ - Click, Build, Run!

8 17

program_options & VC6: internal compiler error?
by Olaf van der Spek 16 Dec '04

16 Dec '04

Hi, I was trying to use the program_options library and used the following example. However, I got an internal compiler error. I reduced the example to this minimum. The library build fine, so I assume it's compatible with VC6. What went wrong? Code: #include "stdafx.h" #include <boost/program_options.hpp> #include <iostream> int main(int ac, char* av[]) { boost::program_options::variables_map vm; std::cout << vm["compression"].as<int>(); // line 8 return 0; } Output: Configuration: Program Options Test - Win32 Release Compiling... StdAfx.cpp Compiling... Program Options Test.cpp C:\VC\temp\Program Options Test\Program Options Test.cpp(8) : error C2679: binary '<<' : no operator defined which takes a right-hand operand of type '' (or there is no acceptable conversion) C:\VC\temp\Program Options Test\Program Options Test.cpp(8) : error C2062: type 'int' unexpected Error executing cl.exe. Program Options Test.exe - 2 error(s), 0 warning(s) Configuration: Program Options Test - Win32 Debug Compiling... Program Options Test.cpp C:\VC\temp\Program Options Test\Program Options Test.cpp(8) : fatal error C1001: INTERNAL COMPILER ERROR (compiler file 'msc1.cpp', line 1794) Please choose the Technical Support command on the Visual C++ Help menu, or open the Technical Support help file for more information Error executing cl.exe. Program Options Test.exe - 1 error(s), 0 warning(s) -- Olaf van der Spek http://xccu.sf.net/

2 2

Re: Nasty bug in Boost 1.32.0 bind]
by Peter Dimov 15 Dec '04

15 Dec '04

Sean Parent wrote: > On Dec 6, 2004, at 12:10 PM, Peter Dimov wrote: >> >> This is not a bug. Bind was deliberately changed to return by value >> because of a bug report similar to the following: >> >> pair_t make_pair( int, int ); >> >> bind( &pair_t::first, bind( make_pair, 10, 20 ) )(); >> >> The inner bind returns by value (because make_pair returns a value) >> and the outer bind returns a reference to a temporary. >> >> A similar, but rare, situation can arise when bind is invoked on an >> rvalue: >> >> pair_t const make_pair(); >> >> bind( &pair_t::first, _1 )( make_pair() ); >> >> I'm more interested in the actual code that was broken by this >> change, because the assert() above was never required to pass. >> By-value or by-reference was (and probably still is) an open >> question. Of course I should have gotten it right in the first place. > > The documentation explicitly states that it will return a const > reference: > ----- >> bind(&X::f, args) >> >> is equivalent to >> >> bind<R>(mem_fn(&X::f), args) >> >> where R is the return type of X::f (for member functions) or a const >> reference to the type of the member (for data members.) > ----- Yes, you are right. I missed that. > The actual failure was in conjunction with boost::function() where it > fails silently because of a cast to result type inside boost function. > Here is a snippet that was my first attempt at isolating the problem: > > ---- > #include <boost/bind.hpp> > #include <boost/function.hpp> > #include <utility> > #include <cassert> > > int main() > { > typedef std::pair<int, int> pair_t; > > pair_t pair(10, 20); > > boost::function<const int&(const pair_t&)> > function(boost::bind(&pair_t::first, _1)); > > const int& x = function(pair); > > assert(&pair.first == &x); > > return 0; > } I get warning C4172 from VC++ 7.1, "returning address of local variable or temporary", on this example (in function_template.hpp:111). The example can be further simplified to: #include <boost/function.hpp> int f() { return 5; } int main() { boost::function<int const & ()> g( f ); int const & r = g(); } which produces the same warning. A static_assert in boost::function would probably help on compilers that don't emit a warning. [...] >> The explicit version: >> >> const int& x = boost::bind<int&>(&pair_t::first, _1)(pair); >> >> should work on conforming compilers. > > Good to know - I will review my code and try to be explicit where > possible - however, I use bind() a lot within templates and rely on it > not making an additional (often expensive) copy of my members. I don't > think there there is a correct answer for the default behavior > (between returning a const& and a value) but the docs need to be > consistent - and a change of this nature needs to be called out in > the release notes. Yes, you are right. I underestimated the impact of this change; the new behavior was supposed to eliminate this sort of bug, not introduce it. Breaking existing code is never pleasant. The change has been in the CVS for some time, but unfortunately most real-world testing occurs on the actual releases.

4 4

Cross Builder and pkg-config discussions
by David Abrahams 15 Dec '04

15 Dec '04

These discussions should probably move to the Boost.Build mailing list now. They've drifted a little far from library development issues. http://www.boost.org/more/mailing_lists.htm#jamboost -- Dave Abrahams Boost Consulting http://www.boost-consulting.com

1 0

[PATCH]Enhancing any_cast
by Mikhail Glushenkov 15 Dec '04

15 Dec '04

This is an enhanced version of Unai Uribarri's patch (see http://lists.boost.org/MailArchives/boost/msg42284.php ). Adds support for the following syntax: boost::any a = 7; int& i = any_cast<int&>(a); Current syntax for reference extraction is int& i = *any_cast<int>(&a); Old syntax is preserved. On compilers without support for partial template specialization users are required to provide explicit specializations of boost::remove_reference for their classes (usually via BOOST_BROKEN_COMPILER_TYPE_TRAITS_SPECIALIZATION) if they want to use the new syntax. Patch was tested on msvc 6 and 7.1, intel 8.1 and gcc 3.3.1. If this patch will be accepted, I'll write tests/update documentation. --- any.hpp Thu Sep 16 11:23:30 2004 +++ new_any.hpp Fri Dec 10 19:47:45 2004 @@ -15,6 +15,12 @@ #include "boost/config.hpp" #include <boost/throw_exception.hpp> +#include <boost/type_traits/remove_reference.hpp> + +#ifdef BOOST_NO_TEMPLATE_PARTIAL_SPECIALIZATION +#include <boost/type_traits/is_reference.hpp> +#include <boost/static_assert.hpp> +#endif namespace boost { @@ -168,12 +174,39 @@ } template<typename ValueType> - ValueType any_cast(const any & operand) + const ValueType any_cast(const any & operand) { - const ValueType * result = any_cast<ValueType>(&operand); + #ifdef BOOST_NO_TEMPLATE_PARTIAL_SPECIALIZATION + // Please use BOOST_BROKEN_COMPILER_TYPE_TRAITS_SPECIALIZATION macro + // to generate specialization of remove_reference for your class + // See type traits library documentation for details + BOOST_STATIC_ASSERT(!::boost::is_reference< + boost::remove_reference<ValueType>::type >::value); + + #endif //BOOST_NO_TEMPLATE_PARTIAL_SPECIALIZATION + + typedef typename boost::remove_reference<ValueType>::type Type; + const Type * result = any_cast<Type>(&operand); if(!result) boost::throw_exception(bad_any_cast()); return *result; + } + + template<typename ValueType> + ValueType any_cast(any & operand) + { + #ifdef BOOST_NO_TEMPLATE_PARTIAL_SPECIALIZATION + // Please use BOOST_BROKEN_COMPILER_TYPE_TRAITS_SPECIALIZATION macro + // to generate specialization of remove_reference for your class + // See type traits library documentation for details + BOOST_STATIC_ASSERT(!::boost::is_reference< + boost::remove_reference<ValueType>::type >::value); + + #endif //BOOST_NO_TEMPLATE_PARTIAL_SPECIALIZATION + + typedef typename boost::remove_reference<ValueType>::type Type; + Type * result = any_cast<Type>(&operand); + if(!result) + boost::throw_exception(bad_any_cast()); + return *result; } }

2 3