March 2013 - Boost - lists.boost.org

[mpi] Boost.MPI and MPI 3!
by Renato Forti 01 Apr '13

01 Apr '13

Hi All, Boost.MPI project will support MPI version 3 with that provide some support to fault tolerance! Have any work in this area? Thanks

2 2

Associative container using a vector
by Stefan Strasser 01 Apr '13

01 Apr '13

Hi everyone, I often find myself using a specific data structure and I'm wondering whether: - it has a wellknown name - there is a generic equivalent in a boost library - or there is interest to add one. The basic idea is that you have a bunch of values stored in one object of type A: struct A{ vector<Value> values; }; and then you have multiple objects of type B that each have to associate data with each value stored in A. you could use: struct B{ map<Value,Data> data; }; but since A::values already provides a contiguous index for the values you can just as well store the data in B in a vector with matching indexes, and make sure that the indexes used in A::values don't change: struct A{ vector<Value> values; vector<std::size_t> freelist; map<Value,std::size_t> index; void erase(...){ push index of unused value to freelist; } void push_back(Value){ reuse a freelist entry or values.push_back; } }; struct B{ vector<optional<Data> > data; }; keep in mind that there is one object A for many objects B. you can now access the associated data of many objects B by looking up the index once in object A, and by wasting some space in B::data for unused entries: b.data[a.index[value]] If Data is a pointer or any other type that knows an empty() state, you can also get rid of optional<> and generalize it as two template like: template<class Value> class index; template<class Mapped,class OptionalTraits=boost_optional_traits> class vector_map; Is there anything like that?

3 4

[config] BOOST_NOEXCEPT on MSVC
by Michael Marcin 01 Apr '13

01 Apr '13

Why isn't BOOST_NOEXCEPT defined to be throw() for msvc? Microsoft Visual Studio 11.0\VC\include\yvals.h defines: #define _NOEXCEPT throw () and the rest of the standard library implementation uses it ex. void swap(thread& _Other) _NOEXCEPT { // swap with _Other _STD swap(_Thr, _Other._Thr); } http://msdn.microsoft.com/en-us/library/wfa0edys.aspx states: throw() The function does not throw an exception. However, if an exception is thrown out of a function marked throw(), the Visual C++ compiler will not call unexpected (see unexpected (CRT) and unexpected (<exception>) for more information). If a function is marked with throw(), the Visual C++ compiler will assume that the function does not throw C++ exceptions and generate code accordingly. Due to code optimizations that might be performed by the C++ compiler (based on the assumption that the function does not throw any C++ exceptions) if a function does throw an exception, the program may not execute correctly. It seems a better choice here than expanding BOOST_NOEXCEPT to nothing.

2 1

[fusion] What do we do with nil ?
by Joel de Guzman 31 Mar '13

31 Mar '13

Hello, This is a constantly nagging issue. There are a couple of places where nil is a PP define especially on the Mac: - objc.h #ifndef nil #define nil __DARWIN_NULL #endif and - MacTypes.h #ifndef nil #define nil NULL #endif So people complain because fusion uses nil as: namespace boost { namespace fusion { struct nil; }} So, nil is a reserved word in Objective-C. But this is C++! And that define is SUCH A BAD PRACTICE in C++! Some folks, like Paul Mensonides maintains a strong stance against this: I would just use #define nil nil as a sanity check and call it done. If the symbol is already defined, compilation error (or at least warning on overly permissive compilers). The symbol will be need to be undefined in between the inclusion of the two libraries. Workarounds like nil_ and (min)(...) just perpetuate the problem. Fix the root problem rather than treating the symptom indefinitely. Otherwise, where is the line drawn? Even if you can't provide a transparent fix (such as defining nil as a constant and min/max as functions), the backwards compatibility argument fails in the long run. The amount of effort required to make the necessary changes to existing code will eventually be outweighed by the effort to implement workarounds put into new code (as well as educate everyone that they are necessary)--especially in cases like these which can mostly be emulated in a way that causes most existing code to work. As I said over and over again, I'm listening and I would want to "fix" this issue if someone comes up with a reasonable (tested) patch that: 1) Detects the platform 2) Renames fusion::nil to something else (only for the affected platform) Yet, no-one wants to do this. Instead we get nasty remarks such as those found here: https://svn.boost.org/trac/boost/ticket/5010#comment:16 What do we do? I'm seeking for a resolution. Regards, -- Joel de Guzman http://www.boostpro.com http://boost-spirit.com

4 5

Re: [boost] Review managers wanted for augmented data structures
by Jonathan Wakely 31 Mar '13

31 Mar '13

8:08 PM, "Vadim Stadnik" wrote: > > Is there an alternative of not integrating advanced data structures into > STL? Why do they need to be put "into STL"? Why can't they exist in addition to it, or alongside it? > Augmented trees are non-standard due to different performance guarantees. They're non-standard because they're not in the standard. > A more reasonable and practical approach to STL would be to provide a wide > set of interchangeable containers and data structures with different > performance guarantees and allow programmers to make decisions about the > most suitable containers for specific applications. That's exactly what the STL is. The STL is about iterators and algorithms and concepts, not a fixed set of predefined containers. You've claimed the STL makes your containers "illegal" which is an odd claim that I can't agree with. Extending the STL with alternative containers is not just acceptable, I'm pretty sure it was intended by Stepanov et al. If your containers don't meet the complexity requirements of an Associative Container as defined in the standard that's fine, just don't claim it is an Associative Container and there shouldn't be a problem, surely. I don't understand why you think there are objections to non-standard containers. The containers in the standard are not suitable for all situations and are not the only choices, if there's a more suitable non-standard container then using it makes sense. This doesn't put the STL "under pressure", it proves its value, because existing algorithms work with both standard and non-standard containers if they model the required concepts.

5 12

[gil] undocumented/unintended concept requirements in color_convert
by Michael Marcin 31 Mar '13

31 Mar '13

I was trying to add other color spaces support and noticed that the rgb to cmyk converter contains: typedef typename channel_type<P2>::type T2; T2 x = channel_traits<T2>::max_value()-get_color(dst,black_t()); if (x>0.0001f) { float x1 = channel_traits<T2>::max_value()/float(x); T2 should model: http://www.boost.org/doc/libs/1_53_0/libs/gil/doc/html/g_i_l_0264.html but this code additionally requires T2 to be greater-than comparable to float and convertible to float.

3 4

[exception] uncaught_exception_count, scope(failure), scope(success)
by Evgeny Panasyuk 31 Mar '13

31 Mar '13

Hello All, There is implementation of uncaught_exception_count - https://github.com/panaseleus/stack_unwinding/blob/master/boost/exception/u… . uncaught_exception_count is a function similar to std::uncaught_exception from standard library, but instead of boolean result it returns unsigned int showing current count of uncaught exceptions. uncaught_exception_count allow us to implement scope(failure) and scope(success) features from D language in C++. ( previous discussion is at http://boost.2283326.n4.nabble.com/scope-exit-D-style-scope-failure-and-sco… ) Implementation of uncaught_exception_count is tested on {Clang 3.2, GCC 3.4.6, GCC 4.1.2, GCC 4.4.6, GCC 4.4.7, MSVC2005SP1, MSVC2008SP1, MSVC2010SP1, MSVC2012} x {x32, x64}. There are also live stress tests - detection of 500 simultaneous uncaught exceptions: GCC 4.6.3: http://liveworkspace.org/code/2qmSso%240 GCC 4.7.2: http://liveworkspace.org/code/dY0VD%240 Clang 3.2: http://liveworkspace.org/code/49MaPk%240 I think that uncaught_exception_count (and uncaught_exception_count_latch primitive) can be added into Boost.Exception library, at least into "detail" section. After that scope(failure) and scope(success) features can be implemented in Boost.ScopeExit library. Note, approximately 14 of 16 Boost.ScopeExit examples in manual are in fact emulations of scope(failure)/scope(success). So, looks like in first place Boost.ScopeExit was introduced to simplify emulation of scope(failure) and scope(success) features. What do you think? Thank you. Best Regards, Evgeny Panasyuk

5 11

Experimenting with modular Boost
by Jonathan Wakely 30 Mar '13

30 Mar '13

The instructions at https://svn.boost.org/trac/boost/wiki/TryModBoost seem to be bogus, apparently because the cloned repo contains no .gitmodules file so the 'git submodule update --init' command does nothing, and I then get an error for the last step: $ cmake -P forward_headers.cmake CMake Error: Error processing file:forward_headers.cmake I'm "only" using Git 1.7.12 but I don't think that's the problem. The older instructions at https://svn.boost.org/trac/boost/wiki/CMakeModularizationStatus refer to a non-existent http://github.com/boost-lib/boost URL, so they don't work either. I've looked at several other wiki pages, but they're also inconsistent or don't help. Is there any accurate documentation for modular Boost? How does someone try modular boost?

2 1

[Filesystem] v3 path separator changes
by Alexander Lamaison 30 Mar '13

30 Mar '13

I'm finally getting round to moving to Filesystem v3 and now my code is breaking all over the place. The cause is the output of string() on Windows which has changed behaviour. I'm manipulating Unix paths for use over SFTP, but doing so on Windows. For instance I might want to append "c" to the Unix path "/a/b". path p("/a/b"); p /= "c"; cout << p.string(); In version 2 this would output "/a/b/c" but now it produces "/a/b\c". Why the breaking change? I'm trying to understand the rationale behind many of the v3 changes which [1] and [2] don't even begin to cover. [1] http://www.boost.org/doc/libs/1_53_0/libs/filesystem/doc/v3.html [2] http://www.boost.org/doc/libs/1_53_0/libs/filesystem/doc/deprecated.html Alex -- Swish - Easy SFTP for Windows Explorer (http://www.swish-sftp.org)

8 33

[switch generator] status
by Jeff Flinn 29 Mar '13

29 Mar '13

Does anyone know the current status of Steven's switch generator library, I see it was reviewed/accepted in 2008, but I don't see it in trunk or any release. Thanks, Jeff

2 1