December 2012 - Boost - lists.preview.boost.org

Use of deprecated macros in Boost
by Marshall Clow 13 Dec '12

13 Dec '12

In Boost 1.50, we (Boost) marked a bunch of macros as "deprecated" In Boost 1.51, we deprecated a bunch more - replaced them with new, shiny macros. I've been working with many of you (library maintainers) to remove all the uses of deprecated macros in Boost, and we've been almost completely successful .. on the trunk. We're a week away from the freezing of the release branch for the 1.53 release, and I'd like to make sure that all those macro changes get merged over to the release branch. http://www.boost.org/development/ I will be happy to help - just let me know. Thanks in advance! -- Marshall

3 11

Heads up - string_ref landing
by Marshall Clow 12 Dec '12

12 Dec '12

I'm about to check in some new functionality into the string_algo library; an implementation of string_ref. A string_ref is a non-owning reference to a string. It is implemented as a {pointer, length} pair, and is exceedingly useful when parsing, and manipulating strings in "read-only" ways. The header will be in boost/algorithm/string_ref.hpp, but I welcome discussion about where it should live. Ideally, I think it should be just "boost/string_ref.hpp". Basic sanity tests will be checked in with the header file - I have tested on clang, clang11, and gcc. One reason for checking this in now is to get the tests run on lots of different systems. Docs will be coming soon, but http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2012/n3442.html is the proposal for "a future version of C++" ;-) Please let me know what you think! -- Marshall Marshall Clow Idio Software <mailto:mclow.lists@gmail.com> A.D. 1517: Martin Luther nails his 95 Theses to the church door and is promptly moderated down to (-1, Flamebait). -- Yu Suzuki

18 45

[smart_ptr][thread][test] breaking change on shared_ptr operator bool.
by Vicente J. Botet Escriba 12 Dec '12

12 Dec '12

Hi, https://svn.boost.org/trac/boost/changeset/81780 " Use explicit operator bool when available; ..." introduces a breaking change on shared_ptr in Boost.Thread and Boost.Test at least. I would preferred if this breaking change was announced and why not added conditionally so that the user could move smoothly to the new semantic. Anyway, I wanted to make this evident so that the Boost maintainer using this facility fixes their code as soon as possible. Best, Vicente

7 19

[atomic] Header rearrangement
by Andrey Semashev 12 Dec '12

12 Dec '12

Hi, Before merging Boost.Atomic to release I would like to propose to rearrange library headers a bit: 1. The current boost/atomic/config.hpp and platform.hpp seem to be not intended for public use as they don't provide any public components. I propose to move them to boost/atomic/detail. 2. The boost/atomic.hpp header defines the atomic<> template. I propose to move its contents to boost/atomic/atomic.hpp. The boost/atomic.hpp will be left as the "include all" header and will only include boost/atomic/atomic.hpp for now. The motivation for the proposal is to cleanup the public library interface (header layout) and simplify extension later. For instance, function-style interface can be added as a new header in boost/atomic, it would fit naturally in the proposed layout. Also, with the upcoming transition to git it seems that all library headers should reside in the library's own subdirectory and not directly in boost/. AFAICT, the change will not affect any other libraries since only boost/atomic.hpp is included currently. If noone objects (Tim, I would very much like to hear from you on this), I can do the necessary changes.

3 2

[thread 1.48] Multiple interrupt/timed_join leads to deadlock
by Gaetano Mendola 12 Dec '12

12 Dec '12

Hi all, I was investigating a rare deadlock when issuing an interrupt and a timed_join in parallel. I come out with the the following code showing the behavior. This is the backtrace of it: http://pastebin.com/xArHgHXf The deadlock is rare so sometime you need to wait a bit. I couldn't try it with boost 1.52 because the code is invalid due the precondition of "thread joinable" when issuing the timed_join. Is the code not valid or a real bug? Regards Gaetano Mendola ////////////////////// CODE ////////////////////// #include <iostream> #include <boost/thread.hpp> struct Runner { void operator()() { const boost::posix_time::time_duration mySeconds = boost::posix_time::seconds(100); boost::this_thread::sleep(mySeconds); } }; struct Interrupter { Interrupter(boost::thread& aThread, boost::barrier& aBarrier) : theThread(aThread), theSeconds(boost::posix_time::seconds(1)), theBarrier(aBarrier) { } void operator()() { theBarrier.wait(); for (int i=0; i<1000;++i) { theThread.interrupt(); theThread.timed_join(theSeconds); } } boost::thread& theThread; boost::posix_time::time_duration theSeconds; boost::barrier& theBarrier; }; int main() { for (size_t i = 0; i < 1000000; ++i) { Runner myRunner; boost::thread myRunnerThread(myRunner); boost::barrier myBarrier(2); Interrupter myInterrupter1(myRunnerThread, myBarrier); Interrupter myInterrupter2(myRunnerThread, myBarrier); boost::thread myInterrupterThread1(myInterrupter1); boost::thread myInterrupterThread2(myInterrupter2); myInterrupterThread1.join(); myInterrupterThread2.join(); myRunnerThread.join(); std::cout << "."; std::cout.flush(); } }

5 34

[thread] [Boost-testing] Trunk regression failed on missing 'virtual_noexcept.cpp'
by Rene Rivera 11 Dec '12

11 Dec '12

Seems the thread lib has some missing test files? ---------- Forwarded message ---------- From: Jim Bell <Jim(a)jc-bell.com> Date: Tue, Dec 11, 2012 at 10:37 AM Subject: [Boost-testing] Trunk regression failed on missing 'virtual_noexcept.cpp' To: Running Boost regression tests <boost-testing(a)lists.boost.org> error: at G:/boost/GCC/trunk/tools_bb/**kernel\modules.jam:107 error: Unable to find file or target named error: 'virtual_noexcept.cpp' error: referred to from project at error: '../libs/thread/test' ______________________________**_________________ Boost-Testing mailing list Boost-Testing(a)lists.boost.org http://lists.boost.org/**mailman/listinfo.cgi/boost-**testing<http://lists.boost.org/mailman/listinfo.cgi/boost-testing> -- -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org - grafik/redshift-software.com -- 102708583/icq - grafikrobot/aim - grafikrobot/yahoo

2 1

Request for interest in a library to ease double dispatching
by Per Karlström 11 Dec '12

11 Dec '12

Hi, I have been tinkering on a template library to ease the use of double dispatching. Now I know there are other techniques such as visitor patterns that might solve this problem in a more general way. However I believe that there are occasions when using double dispatch is a good way forward. I have some excerpts of example code at the end of this mail. Of course if I somehow have missed a boost library that already handles this I would be happy to know about it. So if there is even any interest for this I would be happy if someone would like to cooperate with me to make this library better and in the end so good so that it can be part of boost. I feel I do not have the time to spare to finalize this on my own, plus I feel I still have things to learn and would be happy to cooperate with someone more experienced in template programming. What I in that case need help with is: *How (or is it) possible to make the library more transparent to the number of types that should be in the double dispatch hierarchy. Right now, as I have shown in the code at the end of this mail, you have to put a post-fix number telling how many types to use. *How (if possible) can a class be double dispatchable with different return types. The way I solved it was with a union. But would it somehow be possible to do like this struct Val : public DoubleDispatchable<bool,A,B>, public DoubleDispatchable<int,A,B> *What else would be needed to start the journey to make this a library fit for boost (is it even good enough to ever make into boost?) If anyone is interested in a cooperation I am happy to provide my working draft of this library and to have a more private talk about how we could move forward. Best Regards Per Karlström For now a union is used to hold the return value to make this simmilar for all types of returns. union ReturnVal { float float_type_m; bool bool_type_m; ReturnVal(const bool& bool_type_):bool_type_m(bool_type_){} ReturnVal(const float& float_type_):float_type_m(float_type_){} ReturnVal():bool_type_m(false){} }; In essence a library user can define handler classes akin to this: struct less : public DoubleDispatchHandler_3<ReturnVal,A,B,C> { return_type handle_alpha_alpha(const alpha_type& arg1, const alpha_type& arg2) const {return arg1 < arg2} return_type handle_alpha_beta (const alpha_type& arg1, const beta_type& arg2) const {return true;} return_type handle_alpha_gamma (const alpha_type& arg1, const gamma_type& arg2) const {return true;} //And all other combinations }; Or with a default function struct equal : public DoubleDispatchHandler_3<ReturnVal,A,B,C> { return_type handle_default() const {return false;} return_type handle_alpha_alpha(const alpha_type& arg1, const alpha_type& arg2) const; return_type handle_beta_beta (const beta_type& arg1, const beta_type& arg2) const; return_type handle_gamma_gamma (const gamma_type& arg1, const gamma_type& arg2) const; }; Then the base class is derived from DoubleDispatchable class Val : public DoubleDispatchable_3<ReturnVal,A,B,C> { public: virtual ~Val(){} virtual std::ostream& print(std::ostream& ost) const=0; virtual bool operator<(const Val& other) const =0; virtual bool operator==(const Val& other) const =0; virtual float combine(const Val& other) const =0; }; While Child classes had to define the member function dispatch_<type> struct A : public Val { int val_m; private: //Handles the first double dispatch call return_type dispatch_alpha(const alpha_type& alpha, const handler_type& handler) const { return double_dispatch(alpha,*this,handler); } return_type dispatch_beta(const beta_type& beta, const handler_type& handler) const { return double_dispatch(beta,*this,handler); } return_type dispatch_gamma(const gamma_type& gamma, const handler_type& handler) const { return double_dispatch(gamma,*this,handler); } public: A(const int& val_=0):val_m(val_){} //Normal calls to the operators via a template function bool operator<(const Val& other) const { return double_dispatch(*this,other,less()).bool_type_m; } bool operator==(const Val& other) const { return double_dispatch(*this,other,equal()).bool_type_m; } float combine(const Val& other) const { return double_dispatch(*this,other,add()).float_type_m; } std::ostream& print(std::ostream& ost) const { ost<<"A("<<val_m<<")"; return ost; } }; class B{...}; class C{...};

3 5

Registration for C++Now 2013 is now open
by Boris Kolpackov 11 Dec '12

11 Dec '12

The seventh annual C++Now Conference (formerly BoostCon) will be held at the Aspen Center for Physics in Aspen, Colorado, May 12th to 17th, 2013. "We are thrilled to announce the second annual C++Now conference, the whole-language edition of BoostCon covering all the coolest topics in C++," said Dave Abrahams, Conference Co-Chair. "In 2012, we broadened the conference scope by adding a third track and offering more C++11 coverage than any other event, and the community responded with an unprecedented number of registrations. In 2013, we are going to build on that success with foundational sessions integrating what we've all learned about using C++11 during the past year, while continuing the exploration of cutting-edge topics that BoostCon attendees have come to expect." Early Bird Savings Deadlines! Early Bird conference registration, which ends April 14th, 2013, costs $599. After that date, the registration fee is $699. Register now at our registration page <http://www.eventbrite.com/event/2717332615>. Early Bird hotel registration ends December 31st, 2012 and saves $20 per night. Please reserve your room using the Aspen Meadows online reservation system <https://resweb.passkey.com/go/cplusplusnow2013> Speakers If you are interested in presenting, we are currently accepting proposals <http://cppnow.org/2013-call-for-submissions/>. Registration fee is waived for one speaker of every standard session presentation. Shorter sessions are prorated. Student/Volunteers Registration fees will be waived for a limited number of individuals that wish to attend as volunteers. Volunteer work consists of helping to run the conference and will not prevent volunteers from attending sessions. This is the first year this opportunity is being offered. If you are interested in applying to attend as a volunteer, contact us at <organizers(a)cppnow.org>. Sponsors For a copy of the conference Sponsorship Prospectus contact sponsorship(a)cppnow.org. C++Now is presented by Boost in-cooperation with ACM. Permanent link for this announcement is: http://cppnow.org/2013-registration-announcement/

1 0

Using c++11 features in clang
by Marshall Clow 10 Dec '12

10 Dec '12

In response to a some questions on the boost-users list (and boost bug reports), I wrote up this: http://marshall.calepin.co/llvmclang-and-standard-libraries-on-mac-os-x.html I don't know if this should be a FAQ entry for Boost, or for clang. Suggestions/comments/edits welcome. -- Marshall Marshall Clow Idio Software <mailto:mclow.lists@gmail.com> A.D. 1517: Martin Luther nails his 95 Theses to the church door and is promptly moderated down to (-1, Flamebait). -- Yu Suzuki

3 3

[hash] Tests fail on Intel compiler
by Andrey Semashev 10 Dec '12

10 Dec '12

Hi, I've been looking through Boost.Atomic and Boost.Thread tests on the "lazy piranha fedora 17 64" tester, intel-linux-13.0.1 compiler. The tests fail because Boost.Thread does not compile because of Boost.Hash. Here's an example of build log: http://www.boost.org/development/tests/trunk/developer/output/lazy%20piranh… The problem is that apparently BOOST_HASH_CONFORMANT_FLOATS is not defined on this compiler while there are both ::fpclassify and std::fpclassify functions. Are there valid reasons for not defining the macro on this compiler?

2 1

[trac] Add new components
by Michel Morin 10 Dec '12

10 Dec '12

The following four libraries do not have "components" in trac (lockfree already has the component in trac): - atomic - coroutine - multiprecision - odeint (numeric/odeint) Regards, Michel

4 4

[algorithm][hex] function hex_char_to_int in unnamed namespace in header file
by Felipe Magno de Almeida 10 Dec '12

10 Dec '12

Hello, AFAIK, unnamed namespaces are dangerous in headers. There is a commentary just above the function definition: // this needs to be in an un-named namespace because it is not a template // and might get included in several compilation units. This could cause // multiple definition errors at link time. But it seems the better thing would be to define it inline. It is exactly because it can be used in several translation units that it can cause an ODR violation if defined in a unnamed namespace AFAIK. Regards, -- Felipe Magno de Almeida

3 8

Boost new library policy
by islam y. yasar 10 Dec '12

10 Dec '12

Hi Boost Developers; i and my friend want to develope a library for boost. Is there any policy document for that ? i want to learn this process step by step. thank you.

6 9

[boost::asio] SCTP support with boost::asio
by Ajay Yadav 10 Dec '12

10 Dec '12

Hi, I am newbie to the boost library. I would like to develop the SCTP client-server application. Is boost::asio supported SCTP. Please guide me how I can make the application using SCTP. Regards, Ajay Yadav DISCLAIMER: ----------------------------------------------------------------------------------------------------------------------- The contents of this e-mail and any attachment(s) are confidential and intended for the named recipient(s) only. It shall not attach any liability on the originator or NECHCL or its affiliates. Any views or opinions presented in this email are solely those of the author and may not necessarily reflect the opinions of NECHCL or its affiliates. Any form of reproduction, dissemination, copying, disclosure, modification, distribution and / or publication of this message without the prior written consent of the author of this e-mail is strictly prohibited. If you have received this email in error please delete it and notify the sender immediately. . -----------------------------------------------------------------------------------------------------------------------

2 1

[quickbook] template vs block vs phrase issues
by Andreas Pokorny 09 Dec '12

09 Dec '12

Hi, I use quickbook to write software design documentation targeting pdf. We now introduce a requirements tracking system that includes both source code and design documents. To describe or refer to requirements in docbook docments special xml markup is needed. I.e. something like <reqfooo><status>approved</status><version>1</version><id>DoOnlyNeededStuff</id><description></description></regfoo>. Since I do not want to write XML I added quickbook templates: [template reqfoo[id contents]'''<reqcfoo><id>'''[id]'''</id>'''[contents]'''</reqfoo>'''] [template sdesc[contents]'''<shortdesc>'''[contents]'''</shortdesc>'''] [template verification[contents]'''<verifycrit>'''[contents]'''</verifycrit>'''] [template source[contents]'''<source>'''[contents]'''</source>'''] [template usecase[contents]'''<usecase>'''[contents]'''</usecase>'''] [template status[contents]'''<status>'''[contents]'''</status>'''] [template version[contents]'''<version>'''[contents]'''</version>'''] and finally: [template desc[contents]'''<description>'''[contents] '''</description>'''] Now I want to use block elements inside the contents parameter of desc. Is there any way to achieve that? The only thing I can think of is: [template desc[]'''<description>'''] [template enddesc[]'''</description>'''] But then quickbook is just as good as xml! I tried to make desc a block level template by adding a line break at different positions, but then I get broken xml code, since para tags are spread around <description> regards Andreas

2 5

[Git] Modularizing Boost libraries with sublibs
by Daniel Pfeifer 09 Dec '12

09 Dec '12

While adding Boost.ODEint to the modularization manifest, I realized some inconsistencies about sublibs. The directories in "libs" usually contain one library each. In this case, we move the content to a dedicated Git repository. Some directories however (algorithm, numeric) host multiple libraries. How should we treat them? Move each one to a dedicated Git repository, or leave them grouped together? What's currently inconsistent, is that "numeric" is split into individual repositories, while "algorithm" is not. Further, I saw that the Git submodules of numeric's sublibraries are generated in "libs" rather than "libs/numeric", which is wrong. However, before I fix it, we need to decide whether these individual submodules should be individual or not. Opinions? Cheers, Daniel

4 5

Re: [boost] [Git] tools/wave missing (was: Documentation for Git and Modular Boost conversion)
by Hartmut Kaiser 09 Dec '12

09 Dec '12

> On Fri, Dec 7, 2012 at 3:42 PM, Eric Niebler <eric(a)boostpro.com> wrote: > > On 12/7/2012 11:44 AM, Oliver Kowalke wrote: > > >> Do I have to clone an additional submodule (which is not mentioned in > >> the docu)? > > > > Beman's docs appear to lack instructions for how to clone the > > super-project and initialize the sub-modules. That would be a nice > > addition. I don't know the steps and can't find them ATM, but I know > > it's documented *somewhere*. Beman? > > Just added: https://svn.boost.org/trac/boost/wiki/TryModBoost > > The original instructions were buried in a C++Now presentation slide. When I follow those steps as far as I can see everything looks fine except for tools/wave, which is missing. Is that intentional? Regards Hartmut --------------- http://boost-spirit.com http://stellar.cct.lsu.edu > > Thanks, > > --Beman > > _______________________________________________ > Unsubscribe & other changes: > http://lists.boost.org/mailman/listinfo.cgi/boost

2 2

[Smart Pointers]
by Rahul Sr 09 Dec '12

09 Dec '12

Hi All, This is my first post though so please bear with my sentences. I have used scoped_ptr but this is too tied with new/delete semantics i.e. can't use it with pointers which gets allocated from specialized memory area. class AManager { public: A * allocate(); void deallocate(A * p); }; class A { ..... }; This specially happens when I use scoped pointer for C like pointers/allocator: Can I modify scoped pointer to have release semantics which is user defined: template<typename T> class DefaultReleasePolicy { static void releaseA(T * p) { delete p; } }; class ReleaseAPolicy { static void releaseA(A * p) { AManager::getInstance()->deallocate(p); } }; template<typename T,typemame ReleasePolicy = DefaultReleasePolicy<T> > class ScopedPtrModified { ~ScopedPtrModified() { ReleasePolicy::release(m_ptr); } }; Is this fine? Let me know your thoughts. Regards, Rahul

3 3

[odeint] Request to merge odeint into release branch
by Karsten Ahnert 09 Dec '12

09 Dec '12

I would like to ask for permission to merge odeint into the release branch. There are currently four failures in the regression tests. Two have already been fixed. The other two originate from the build time out of 180 seconds. These two tests will often require longer build times, since they heavily use Boost.Fusion, Boost.Units and Boost.MPL.

2 5

[multiprecision] Request permision to merge to release.
by John Maddock 09 Dec '12

09 Dec '12

I'd like to ask permission to merge the multiprecision lib to release. Note that as things stand I'm still chasing down test failures, but these are now mostly down to: * Things that work locally, but the testers haven't cycled yet. * 300s timeouts - mostly from the intel compiler :-( Aside: in an effort to reduce the number of timeouts, I've basically gutted some of the tests when the compiler is Intel, and I think I have all the build times under 20s on my Linux box, yet I'm still getting 300s timeouts on some tester machines, either those machines are seriously underpowered or I'm missing something here :-( Cheers, John.

3 5

[smart ptr] [shared] suggestion: single pointer in object returned from make_shared
by Julian Gonggrijp 09 Dec '12

09 Dec '12

Dear Smart Ptr maintainer(s), I recently learned that boost::make_shared allocates the space for the pointee and the reference counter in one go. It makes a lot of sense, and I might have suggested it if it wasn't already the case. Now I'm assuming that a shared_ptr stores two raw pointers internally, one to the pointee and one to the counter. I guess that would be necessary in the "normal" case where the pointee and the counter are allocated separately. Now in the case where the shared_ptr is created by make_shared, ideally you'd need to store only one raw pointer because pointee and counter reside in the same memory object. That would be the finishing touch to a nonintrusive smart pointer that really performs like an intrusive one. To me such a thing seems highly desirable. I realise this would require a new implementation of the shared pointer. Could this somehow be made to cooperate or at least coexist with the old shared_pointer? Would it perhaps even be an option to deprecate the old shared_pointer in order to entirely replace it by the new version in the future? I'll be most interested to read your comments on this idea. -Julian

1 0

Boost.Test updates in trunk: need for (mini) review?
by Gennadiy Rozenal 09 Dec '12

09 Dec '12

Hi, It's been a long while since I merged any changes into boost release and by now there are whole bunch of new features there. Some are small, but some others are big ones such that they can change the way people use the library. I'd like to list them here (excluding bug fixes and removal of deprecated interfaces) and ask what should we do with these. Specifically let me know if you think these require some kind of (mini) review from community, but any other comments are welcome as well. So here we go: I. New testing tool BOOST_CHECKA This tool is based on excellent idea from Kevlin Henney. I chose the name CHECKA for this tool due to the lack of better name, but I am open to suggestions. This testing tool capable of replacing whole bunch of existing other tools like BOOST_CHECK_EQUAL, BOOST_CHECK_GT etc. Usage is the most natural you can wish for: BOOST_CHECKA( var1 - var2 >= 12 ); And the output will include as much information as we can get: error: in "foo": check var1 - var2 >= 12 failed [23-15<12] II. New "data driven test case" subsystem New data driven test case subsystem represent a generalization of parameterized test case and test case template and eventually will replace both. The idea is to allow user to specify an arbitrary (monomorphic or polymorphic) set of samples and run a test case on each of the samples. Samples can have different arity, thus you can have test cases with multiple parameters of different types. For now we support following dataset kinds (aside of generators none of dataset constructions routines performs *any* copying): a) singleton - dataset constructed out of single sample data::make(10) - singleton dataset with integer sample data::make("qwerty") - singleton dataset with string sample b) array - dataset constructed out of C array int a[] = {1,2,3,4,5,6}; data::make(a) - dataset with 6 integer values c) collection - dataset constructed out of C++ forward iterable collection std::vector<double> v{1.2, 2.3, 3.4, 5.6 }; data::make(v) - dataset with 6 double values d) join - dataset constructed by joining 2 datasets of the same type int a[] = {1,2,3}; int b[] = {7,8,9}; data::make(a) + data::make(b) - dataset with 6 integer values e) zip - dataset constructed by zipping 2 datasets of the same size, but not necessarily the same type This dataset has an arity which is sum of argument dataset arities. int a[] = {1,2,3}; char* b[] = {"qwe", "asd", "zxc"}; data::make(a) ^ data::make(b) dataset with 3 samples which are pairs of int and char*. f) grid - dataset constructed by "multiplying" 2 datasets of the same different sizes and types type This dataset has an arity which is sum of argument dataset arities. int a[] = {1,2,3}; char* b[] = {"qwe", "asd"}; double c[] = {1.1, 2.2}; data::make(a) * data::make(b) * data::make(c) dataset with 12 samples which are tuples of int and char* and double. g) xrange generator dataset which produes samples in some range data::xrange( 0., 3., 0.4 ) - dataset with 8 double samples data::xrange<int>((data::begin=9, data::end=15)) - dataset with 6 int samples data::xrange( 1., 7.5 ) - dataset with 7 double samples data::xrange( 5, 0, -1 ) - dataset with 5 int samples h) random - generator dataset with unlimited number of random samples data::random(data::distribution = std::normal_distribution<>(5.,2)) dataset with random double numbers following specified distribution data::random(( data::engine = std::minstd_rand(), data::distribution = std::discrete_distribution<>(), data::seed = 20UL )); dataset with random int numbers following specified distribution While all this interfaces can be used by itself to build complex datasets with various purposes, primary use case it was developped for is new data driven test case interface: BOOST_DATA_TEST_CASE( test_name, dataset, parameter_names... ) Here are couple examples: int samples1[] = {1,2,3}; BOOST_DATA_TEST_CASE( t1, samples1, sample ) { BOOST_CHECKA( foo(sample) > 0 ); } Above test case is going to be executed 3 times with different sample values char* strs[] = {"qwe", "asd", "zxc", "mkl" }; BOOST_DATA_TEST_CASE( t1, data::xrange(4) ^ strs ^ data::random(), intval, str, dblval ) { MyObj obj( dblval, str ); BOOST_CHECKA( obj.goo() == intval ); } Above test case will be executed 4 times with different values of parameters intval, str, and dblval. Polymorphic datasets are still being developed, but should appear soon(ish). III. Auto-registered test unit decorators. Previously is was not possible and/or convenient to assign attributes to the automatically registered test units. To alleviate this I've introduced a notion of test unit decorator. These can be "attached" to any test unit similarly to like it is done in other languages Following decorators already implemented: label - adds labels to a test unit expected_failures - set expected failures for a test unit timeout - sets timeout for a test unit description - sets a test unit description depends_on - sets a test unit dependency enable_if/disable_if - facilitates a test unit status change fixture - assigns fixture to a test units Test unit description is new test unit attribute, which is reported by new list_content command line argument described below. Usage of labels covered below as well. enable_if/disable_if allow new, much more flexible test management. By adding enable_if/disable_if decorators to the test unit you can conditionally select which test units to run at construction time based on some compile time or run-time parameters. And we finally we have suite level fixtures which are set by attaching fixture decorator to a test suite (suite level fixtures executed once per test suite). Attachment of decorator is facilitated by using of BOOST_TEST_DECORATOR. Note that you can use any of '+', '-', '*' symbols to attach decorator (and any number of '*'): BOOST_TEST_DECORATOR( + unittest::fixture<suite_fixture>() ) BOOST_AUTO_TEST_SUITE( my_suite1 ) { } BOOST_TEST_DECORATOR( - unittest::timeout( 100 ) - unittest::expected_failures( 1 ) - unittest::enable_if( 100 < 50 ) ) BOOST_AUTO_TEST_CASE( my_test5 ) BOOST_TEST_DECORATOR( **** unittest::label( "L3" ) **** unittest::description( "suite description" ) **** unittest::depends_on( "my_suite2/my_test7" ) ) BOOST_AUTO_TEST_CASE( my_test11 ) IV Support for "run by label" plus more improvements for filtered runs Previously you only had an ability to filter test unit to execute by name. Now you can attach labels to test units and create collections to run which are located in arbitrary positions in test tree. For example you can have special label for performance test cases, which is attached to all test units which are responsible for testing performance of your various components. Or you might want to collect exception safety tests etc. To filter test units by level you still use --run_test CLA. Labels are dented by @prefix: test.exe --run=@performance You can now repeat --run_test argument to specify multiple conditions: test.exe --run=@performance,@exception_safety --run=prod/subsystem2 In addition run by name/label now recognizes dependencies. So if test unit A depends on test unit B and test unit B is disabled/is not part of current run test unit A will not run as well. Finally you now have an ability to specify "negative" conditions, by prefixing name or label with ! test.exe --run=!@performance This will run all test units which are not labeled with "performance". V. Support for failure context In many scenarios it is desirable to specify an additional information to a failure, but you do not need to see it if run is successful. Thus using regular print statements is undesirable. This becomes especially important when you develop common routine, which performs your testing and invoke it from multiple test units. In this case failure location is no help at all. To alleviate this problem two new tools are introduced: BOOST_TEST_INFO BOOST_TEST_CONTEXT BOOST_TEST_INFO attaches context to next assertion which is executed. For example: BOOST_TEST_INFO( "a=" << a ); BOOST_CHECKA( foo(a) == 0 ); BOOST_CHECK_CONTEXT attaches context to all assertions within a scope: BOOST_CHECK_CONTEXT( "Starting test foo from subset " << ss ) { BOOST_CHECKA( foo(ss, a) == 1 ); BOOST_CHECKA( foo(ss, b) == 2 ); } VI. Two new command line arguments: list_context, wait_for_debugger Using list_context you can see a full or partial test tree of your test module. wait_for_debugger allows to force test module to wait till you attach debugger. VII. Colored output. Using new CLA --color_output you can turn on colored output for error and information messages VIII New production testing tools interface introduced This feature was covered in details in my presentation in BoostCon 2010. The gist if this is an ability to use Boost.Test testing tools interfaces in production (nothing to do with testing) code. User supplied implementation is plugged in IX. Number of smaller improvements: * Notion of framework shutdown. Allows to eliminate some fake memory leaks * Added checkpoints at fixture entry points, test case entry point and test case exit point for auto registered test cases * New FPE portable interfaces introduced and FPE handling is separated from system errors handling. You can detect FPE even if catch_system_error is false * Added an ability to erase registered exception translator * execution_monitor: new interface vexecute - to be used to monitor nullary functions with no result values * test_tree_visitor interface extended to facilitate visitors applying the same action to all test units * execution_monitor use typeid to report "real" exception type if possible. * New ability to redirect leaks report into a file Thank you for your time, Gennadiy

21 66

[config] Macro for null pointer
by Edward Diener 08 Dec '12

08 Dec '12

I have found something like this to be helpful, when working with multiple compilers: #include <boost/config.hpp> #if defined(BOOST_NO_NULLPTR) #define BOOST_XXX_NULLPTR 0 #else #define BOOST_XXX_NULLPTR nullptr #endif where XXX is some local name for my own use. And then use BOOST_XXX_NULLPTR in places where a null pointer is needed. Would this be a candidate for a BOOST_NULLPTR macro in the config library instead ? Edward Diener

16 63

Re: [boost] [boost-steering] Re: [Git] Documentation for Git and Modular Boost conversion
by Brian Wood 08 Dec '12

08 Dec '12

Dave Abrahams: > Boost is making this transition. It *will* cause some disruption. It > will also make things easier for some people. I'm glad to hear this effort is continuing. > IMO we should not burden this move, which is supposed to > make it easier and more efficient overall for Boost to operate, > with any unnecessary obligations. That sounds harsh, but I guess that's the way the cookie crumbles. -- Brian Wood Ebenezer Enterprises -- so far G-d has helped us. http://webEbenezer.net

1 0

[serialization] Proposal: make 'version' trait (and others) enable_if-friendly
by Gabriel Redner 07 Dec '12

07 Dec '12

I have recently run up against a limitation of the serialization library in terms of specializing the 'version' trait. Fortunately, I believe the issue has a simple workaround which I don't believe will break existing code. In general, one can change the version of a serialized class by specializing the 'version' template class: ============================== template <typename T> struct version { BOOST_STATIC_CONSTANT(unsigned int, value = 0); }; ... template <> struct version<MyType> { BOOST_STATIC_CONSTANT(unsigned int, value = 1); }; ============================== However, this approach does not work if 'MyType' is of the form (see [1]): ============================== template <typename T> struct Foo { struct MyType { int data1; }; }; template <typename T> struct version<Foo<T>::MyType> // ERROR { BOOST_STATIC_CONSTANT(unsigned int, value = 1); }; ============================== It's not possible for the compiler to deduce the enclosing type, so this doesn't work. However, if MyType can be identified by a metafunction, then I can use enable_if to select the specialization of 'version' that I want. However, this requires support from 'version' itself: ============================== template <typename T, typename Enable = void> // NOTE new template parameter struct version { BOOST_STATIC_CONSTANT(unsigned int, value = 0); }; ... template <typaneme T> struct version<MyType, typename boost::enable_if<is_mytype<T> >::type> { BOOST_STATIC_CONSTANT(unsigned int, value = 1); }; ============================== So, my proposal is simply to add this extra template parameter to 'version'. I do not believe other uses of 'version' will be affected, but it would not surprise me if I've overlooked something. Presumably the same modification could be made to implementation_level etc. A patch (for 'version' only) against trunk is enclosed below. Comments are welcome! Thanks, -Gabe [1] http://boost.2283326.n4.nabble.com/serialization-versioning-for-class-neste… =================================================================== --- version.hpp (revision 81723) +++ version.hpp (working copy) @@ -32,7 +32,7 @@ // default version number is 0. Override with higher version // when class definition changes. -template<class T> +template<class T, class Enable=void> struct version { template<class U> @@ -53,8 +53,8 @@ }; #ifndef BOOST_NO_INCLASS_MEMBER_INITIALIZATION -template<class T> -const int version<T>::value; +template<class T, class Enable> +const int version<T, Enable>::value; #endif } // namespace serialization ===================================================================

5 12

Inspect Report
by Daniel James 07 Dec '12

07 Dec '12

Hi, The snapshot builds and inspect results haven't been running for some time, so I've added running the inspect script to the documentation build for more up to date results. You can see them at: http://boost.cowic.de/rc/docs-inspect-trunk.html http://boost.cowic.de/rc/docs-inspect-release.html These are generated after building documentation, which is not exactly the same as the final release, but it's close. Daniel

4 10

[context] asm_context_sources breaking regression testing for WinCE
by David Deakins 07 Dec '12

07 Dec '12

Hi Oliver, I finally got a little bit of time to investigate recent strange behavior on the VeecoFTC regression tester and found that the build jamfile for the context library is halting Boost.Build during the regression tests. It appears that, when it can't find an appropriate architecture to use, it actually generates an argument error in try-one-generator-really. I attached the output of the bjam.log and perhaps it will be useful for investigating. Let me know if you would like me to try anything. Thanks, -Dave

2 12

[iterator] UB when implicitly using default constructed counting_iterator<unsigned>
by Peter Sommerlad 07 Dec '12

07 Dec '12

Hi Jeffrey, hi all, I believe the usability of counting iterator could be improved, by value-initializing its Iterator type in the default constructor, because integral types won't be initialized by its default constructor. This can lead to hard to recognize user errors, e.g., when wrapping such a counting iterator with a filter iterator. counting_iterator is special in that sense, because to my understanding it is intended to adapt integral types to become iterators. And a default initialized scalar type, isn't. I suggest a one line change: counting_iterator() { } should become counting_iterator() : super_t(Incrementable()) { } The following example program shows the interesting UB, because the 3rd argument of make_filter_iterator, default constructs a counting_iterator<unsigned>(). Any thoughts? Objections? Teachings... Regards Peter. =====>8 #include <iostream> #include <boost/iterator/filter_iterator.hpp> #include <boost/iterator/counting_iterator.hpp> #include <algorithm> #include <iterator> #include <functional> struct is_prime{ bool operator()(unsigned int val) const { if (val < 2) return false; using std::placeholders::_1; auto ce=boost::make_counting_iterator(val); return ce==std::find_if_not(boost::make_counting_iterator(2u),ce, bind(std::modulus<unsigned>{},val,_1)); } }; int main(){ std::copy_n(boost::make_filter_iterator(is_prime{},boost::make_counting_iterator(1u) // if the following line is omitted we have undefined behavior, because of uninitialized unsigned value // ,boost::make_counting_iterator(std::numeric_limits<unsigned>::max()) // must be different from above value ),40,std::ostream_iterator<unsigned>{std::cout,", "}); } =====8< -- Prof. Peter Sommerlad Institut für Software: Bessere Software - Einfach, Schneller! HSR Hochschule für Technik Rapperswil Oberseestr 10, Postfach 1475, CH-8640 Rapperswil http://ifs.hsr.ch http://cute-test.com http://linticator.com http://includator.com tel:+41 55 222 49 84 == mobile:+41 79 432 23 32 fax:+41 55 222 46 29 == mailto:peter.sommerlad@hsr.ch

3 2

[coroutine] permission to merge to boost-release?
by Oliver Kowalke 07 Dec '12

07 Dec '12

Hello, because the merge window for new libraries is closed on Monday I'd like to ask for permission to merge boost.coroutine into boost-release. regards, Oliver

2 2

Re: [boost] [iterator] UB when implicitly using default constructed counting_iterator<unsigned>
by Peter Sommerlad 07 Dec '12

07 Dec '12

Hi Jeff On 01.12.2012, at 02:53, Jeffrey Lee Hellrung, Jr. wrote: >> And a default >> initialized scalar type, isn't. >> > > "isn't"...what, exactly? it isn't initialized with a defined value. struct X{ X(){} unsigned x; }; void foo(){ X y; // -> y.x has an uninitialized value AFAIK if it is local } that is the situation we run into with the default constructed counting_iterator<unsigned>() > I'm not opposed to giving defined behavior to a default constructed > counting_iterator< [integral type] >, I'm just not sure what that defined > behavior should be. Maybe I'm over-analyzing it... I still believe the most simple change I suggested is enough (at least it would cover my situation well enough). Those input iterators where a default constructed one is the end iterator is IMHO more of an exception, because it is a special case for the stream iterators. Your idea with std::numberic_limits< Iterator >::max() requires a much more complex solution with template meta programming, i.e., enable_if<is_arithmetic<Iterator>> or similar. And it won't provide the default value of Iterator, which is zero in the case of the built-in arithmetic types, i.e., unsigned(). What do you think? Regards Peter. -- Prof. Peter Sommerlad Institut für Software: Bessere Software - Einfach, Schneller! HSR Hochschule für Technik Rapperswil Oberseestr 10, Postfach 1475, CH-8640 Rapperswil http://ifs.hsr.ch http://cute-test.com http://linticator.com http://includator.com tel:+41 55 222 49 84 == mobile:+41 79 432 23 32 fax:+41 55 222 46 29 == mailto:peter.sommerlad@hsr.ch

7 21

[array] last check in break serialization for array
by Robert Ramey 06 Dec '12

06 Dec '12

A particular test in the serialization library spontaneously has broken. By "spontaneously" I mean that I haven't checked anything in lately. The test is test_array_xml_warchive in the serialization library test matrix. Here are a few facts: a) I haven't checked anything in so the serialization library itself hasn't changed. b) It occurs on all GCC platforms. c) it "seems" to have started around 30 November d) it occurs only on the trunk - all these tests have been run since 30 November e) it doesn't occur on the release branch - none of these tests run since 30 November f) it occours on my development machine. Note that I use release branch for all testing - except for the serialization library itself which I use the trunk version. g) The nature of the error suggests that it is related to memory allocation/deallocation > From all of the above I'm looking at change on the release branch #81635 > (or there abouts) by glenfe Merged revision(s) 81608-81610 from trunk: Optimization: Add specializations of make_array_helper, allocate_array_helper, and array_deleter for fixed size arrays to avoid storing size. ........ Consistent formatting across overloads of make_shared and allocate_shared (array forms). ........ More consistency in type parameters in helper details of allocate_shared and make_shared. ... Does anyone have any insight into this? Robert Ramey

3 2

[Interprocess] message_queue should use timed_lock
by Hartmut Kaiser 06 Dec '12

06 Dec '12

Hey, I'm running into a timing issue with interprocess::message_queue even when I use the timed send and received functions: One end: if (!mq_->timed_receive(&msg, sizeof(msg), recvd_size, priority, boost::get_system_time() + boost::posix_time::milliseconds(1))) { // ... return false; } // ... The other end: while (!mq_->timed_send(&msg, sizeof(msg), 0, boost::get_system_time() + boost::posix_time::milliseconds(1))) { break; } (mq_ is a shared_ptr<message_queue>) here both ends sit forever inside the timed_receive/timed_send function, both waiting to acquire a lock (ipc/message_queue.hpp, lines 702 and 799), which is held by a third participating process. Shouldn't this lock be using a timed_lock to allow to maintain the API post condition (i.e. returning after the given amount of time in any case)? Regards Hartmut --------------- http://boost-spirit.com http://stellar.cct.lsu.edu

2 1

[trunk-tests] msvc-8.0 internal compiler error
by Oliver Kowalke 06 Dec '12

06 Dec '12

Hi, msvc-8.0 generates an internal compiler error in the regression tests for boost.coroutine (tests in boost-trunk - http://www.boost.org/development/tests/trunk/developer/output/VC8%20jc-bell… ). Is it possible to prevent this error a code modification? Other versions like msvc-9.0/10.0/11.0/12.0 do pass the regression test. Oliver test_coroutine.cpp ..\boost/mpl/eval_if.hpp(41) : fatal error C1001: An internal error has occurred in the compiler. (compiler file 'msc1.cpp', line 1393) To work around this problem, try simplifying or changing the program near the locations listed above. Please choose the Technical Support command on the Visual C++ Help menu, or open the Technical Support help file for more information ..\boost/foreach.hpp(364) : see reference to class template instantiation 'boost::mpl::eval_if<C,F1,F2>' being compiled with [ C=boost::mpl::false_, F1=boost::range_const_iterator<boost::coroutines::coroutine<int (void)>>, F2=boost::range_mutable_iterator<boost::coroutines::coroutine<int (void)>> ] ..\libs\coroutine\test\test_coroutine.cpp(415) : see reference to class template instantiation 'boost::foreach_detail_::foreach_iterator<T,C>' being compiled with [ T=coro_int_void, C=boost::mpl::false_ ] Internal Compiler Error in C:\Program Files (x86)\MSVC8\VC\BIN\cl.exe. You will be prompted to send an error report to Microsoft later.

1 0

[compond_file_binary] Gauging interest in a possible library submission.
by А В 06 Dec '12

06 Dec '12

Hi, all. Is there any interest in a cross-platform C++ library which allows to create/read/write binary compound files? Briefly, compound file is filesystem for storing files and directories within a single file on a disk. Initially compound file format is developed by Microsoft and now is the part of Microsoft's Open Specifications Documentation. Compound files are used across various platforms and applications, and not restricted by some specific domain. In general, any application can store information in such a way. I've found on the Internet interest in such library and decided to implement it. Additional information can be found here: Article on the Wikipedia: http://en.wikipedia.org/wiki/Compound_File_Binary_Format Benefits of Compound Files: http://msdn.microsoft.com/en-us/library/windows/desktop/aa378938(v=vs.85).a… [MS-CFB] Open specification for compound files: http://msdn.microsoft.com/en-us/library/dd942138.aspx Regards, Alexander Voitenko.

11 26

generating docu fails
by Oliver Kowalke 05 Dec '12

05 Dec '12

Hello, 'quickbook context.qbk' generates context.xml but generating the html via bjam fails: /boost_1_52_0/tools/build/v2/build/generators.jam:1085: in ensure-type from module generators error: target { context.qbk. } has no type /boost_1_52_0/tools/build/v2/build/generators.jam:1319: in generators.construct from module generators /boost_1_52_0/tools/build/v2/build/targets.jam:1495: in construct from module object(typed-target)@927 /boost_1_52_0/tools/build/v2/build/targets.jam:1298: in object(typed-target)(a)927.generate from module object(typed-target)@927 /boost_1_52_0/tools/build/v2/build/targets.jam:732: in generate-really from module object(main-target)@958 /boost_1_52_0/tools/build/v2/build/targets.jam:704: in object(main-target)(a)958.generate from module object(main-target)@958 /boost_1_52_0/tools/build/v2/build/targets.jam:258: in object(project-target)(a)926.generate from module object(project-target)@926 /boost_1_52_0/tools/build/v2/build-system.jam:736: in load from module build-system /boost_1_52_0/tools/build/v2/kernel/modules.jam:283: in import from module modules /boost_1_52_0/tools/build/v2/kernel/bootstrap.jam:142: in boost-build from module /boost_1_52_0/boost-build.jam:17: in module scope from module how can it be fixed? Oliver

2 2

[mpl] [bimap] BOOST_MPL_LIMIT_METAFUNCTION_ARITY issue on g++ 4.6.3 and boost 1.50
by Adam Lach 05 Dec '12

05 Dec '12

Hi, I have already asked for help on Boost.users, but with no luck, see here <http://boost.2283326.n4.nabble.com/Problems-in-boost-mpl-aux-template-arity…> . Since then I have managed to reproduce this problem in a simpler case: #define BOOST_MPL_CFG_NO_PREPROCESSED_HEADERS #define BOOST_MPL_LIMIT_METAFUNCTION_ARITY 10 #include <boost/mpl/inherit.hpp> #include <boost/bimap.hpp> struct S0 {}; struct S1 {}; struct S2 {}; struct S3 {}; struct S4 {}; struct S5 {}; template<typename T0, typename T1, typename T2, typename T3, typename T4, typename T5> struct example : boost::mpl::inherit<T0,T1,T2,T3,T4,T5> {}; int main() { example<S0,S1,S2,S3,S4,S5> pasta; return 0; } The weird thing is that it occurs only if bimap.hpp is included. Please help. Thank you in advance, Adam Lach -- View this message in context: http://boost.2283326.n4.nabble.com/boost-mpl-bimap-BOOST-MPL-LIMIT-METAFUNC… Sent from the Boost - Dev mailing list archive at Nabble.com.

2 2

Looking for someone with MSVC 8 or MSVC 9
by Marshall Clow 04 Dec '12

04 Dec '12

One of the string_ref tests is failing on these two compilers - neither of which I have access to. I would appreciate some help debugging this failure [ FWIW, the same tests pass on MSVC 10 and 11 ] Thanks in advance! -- Marshall Marshall Clow Idio Software <mailto:mclow.lists@gmail.com> A.D. 1517: Martin Luther nails his 95 Theses to the church door and is promptly moderated down to (-1, Flamebait). -- Yu Suzuki

3 4

Re: [boost] [iterator] UB when implicitly using default constructed counting_iterator<unsigned>
by Peter Sommerlad 04 Dec '12

04 Dec '12

Hi Jeff, On 04.12.2012, at 18:59, Jeffrey Lee Hellrung, Jr. wrote: > On Sun, Dec 2, 2012 at 8:41 AM, Peter Sommerlad <peter.sommerlad(a)hsr.ch>wrote: > just consider the following small program: >> >> #include <iostream> >> #include <boost/iterator/counting_iterator.hpp> >> >> int main(){ >> std::cout << unsigned() << '\n'; // zero >> std::cout << *boost::counting_iterator<unsigned>()<< '\n';// >> undefined behavior >> } >> > > Not quite UB :) It just outputs an undefined value. It won't blow up the > moon, for example. That is not true. According to the standard reading an uninitialized value incurs undefined behavior. While in practice your computer might not explode, it is inherently unsafe and unportable to rely on something like this. See below. Therefore my concerns (I am a regular attendant of the standard committee meetings, so I can check with the real experts about that, if you doubt). 4.1 Lvalue-to-rvalue conversion [conv.lval] • 1 A glvalue (3.10) of a non-function, non-array type T can be converted to a prvalue.53 If T is an incomplete type, a program that necessitates this conversion is ill-formed. If the object to which the glvalue refers is not an object of type T and is not an object of a type derived from T, or if the object is uninitialized, a program that necessitates this conversion has undefined behavior. If T is a non-class type, the type of the prvalue is the cv-unqualified version of T. Otherwise, the type of the prvalue is T.54 > > I understand your concerns, I think, but I believe it's somewhat misguided. > I believe your proposal could do at least as much to mask errors as it will > to find errors. Indeed, your original example would've silently worked as > expected (with the present implementation of filter_iterator, anyway) with > your proposed change, even though it's (still) incorrect usage. I still believe it is much better for a library to not incur silently undefined behavior, but to provide a defined default value, even if it is unusable in general, but detectable. if I would have used 0 as the start iterator in the original example instead of 1 I would have seen something is wrong with my unit tests. However, the undefined behavior didn't help me detecting it. Regards Peter. -- Prof. Peter Sommerlad Institut für Software: Bessere Software - Einfach, Schneller! HSR Hochschule für Technik Rapperswil Oberseestr 10, Postfach 1475, CH-8640 Rapperswil http://ifs.hsr.ch http://cute-test.com http://linticator.com http://includator.com tel:+41 55 222 49 84 == mobile:+41 79 432 23 32 fax:+41 55 222 46 29 == mailto:peter.sommerlad@hsr.ch

1 0

[BGL] status of maximum adjacency search
by Fernando Vilas 03 Dec '12

03 Dec '12

Hi, What remains to get ticket 6780 ready to commit? -- Thanks, Fernando Vilas fvilas(a)iname.com

2 2

[functional/hash] error: more than one instance of overloaded function "fpclassify" matches the argument list:
by Vicente J. Botet Escriba 02 Dec '12

02 Dec '12

Hi, there is a compile error that making fail all the tests in this tester. Could someone take care of it? Best, Vicente Test output: lazy piranha fedora 17 64 C++11 - thread - async__async_p / intel-linux-13.0.1 Rev 81655 / Sat, 1 Dec 2012 18:49:57 +0000 *Report Time: *Sun, 2 Dec 2012 12:54:01 +0000 Compile [2012-12-02 03:35:40 UTC]: fail "/opt/intel/bin/icpc" -c -xc++ -std=c++11 -O0 -g -w2 -inline-level=0 -pthread -fPIC -wd1418 -wd193,304,383,444 -wd2415 -wd593,981 -DBOOST_ALL_NO_LIB=1 -DBOOST_CHRONO_DYN_LINK=1 -DBOOST_SYSTEM_DYN_LINK=1 -DBOOST_SYSTEM_NO_DEPRECATED -DBOOST_THREAD_BUILD_DLL=1 -DBOOST_THREAD_POSIX -DBOOST_THREAD_THROW_IF_PRECONDITION_NOT_SATISFIED -DBOOST_THREAD_USE_DLL=1 -I".." -c -o "/home/michael/boost_testing/results/boost/bin.v2/libs/thread/test/async__async_p.test/intel-linux-13.0.1/debug/threading-multi/sync/futures/async/async_pass.o" "../libs/thread/test/sync/futures/async/async_pass.cpp" ../boost/functional/hash/detail/hash_float.hpp(214): error: more than one instance of overloaded function "fpclassify" matches the argument list: function "fpclassify(double)" function "std::fpclassify(double)" argument types are: (double) switch (fpclassify(v)) { ^ detected during: instantiation of "std::size_t={unsigned long} boost::hash_detail::float_hash_value(T) [with T=double]" at line 312 of "../boost/functional/hash/hash.hpp" instantiation of "boost::hash_detail::float_numbers<T>::type boost::hash_value(T) [with T=double]" at line 413 of "../boost/functional/hash/hash.hpp" compilation aborted for ../libs/thread/test/sync/futures/async/async_pass.cpp (code 2) TinyUrl <javascript:make_tinyurl()>

1 0

Fwd: [iterator] UB when implicitly using default constructed counting_iterator<unsigned>
by Jeffrey Lee Hellrung, Jr. 01 Dec '12

01 Dec '12

[Sorry, forgot to reply-all.] ---------- Forwarded message ---------- From: Jeffrey Lee Hellrung, Jr. <jeffrey.hellrung(a)gmail.com> Date: Fri, Nov 30, 2012 at 5:53 PM Subject: Re: [boost][iterator] UB when implicitly using default constructed counting_iterator<unsigned> To: Peter Sommerlad <peter.sommerlad(a)hsr.ch> On Fri, Nov 30, 2012 at 1:14 AM, Peter Sommerlad <peter.sommerlad(a)hsr.ch>wrote: > Hi Jeffrey, hi all, > > I believe the usability of counting iterator could be improved, by > value-initializing its Iterator type in the default constructor, because > integral types won't be initialized by its default constructor. This can > lead to hard to recognize user errors, e.g., when wrapping such a counting > iterator with a filter iterator. > > counting_iterator is special in that sense, because to my understanding it > is intended to adapt integral types to become iterators. And a default > initialized scalar type, isn't. > "isn't"...what, exactly? > I suggest a one line change: > > counting_iterator() { } > > should become > > counting_iterator() : super_t(Incrementable()) { } > Well, that's one thing you can do... > The following example program shows the interesting UB, because the 3rd > argument of make_filter_iterator, default constructs a > counting_iterator<unsigned>(). > > Any thoughts? Objections? Teachings... > > Regards > Peter. > > =====>8 > #include <iostream> > #include <boost/iterator/filter_iterator.hpp> > #include <boost/iterator/counting_iterator.hpp> > #include <algorithm> > #include <iterator> > #include <functional> > > struct is_prime{ > bool operator()(unsigned int val) const { > if (val < 2) return false; > using std::placeholders::_1; > auto ce=boost::make_counting_iterator(val); > return ce==std::find_if_not(boost::make_counting_iterator(2u),ce, > > bind(std::modulus<unsigned>{},val,_1)); > } > }; > int main(){ > > std::copy_n(boost::make_filter_iterator(is_prime{},boost::make_counting_iterator(1u) > // if the following line is omitted we have undefined behavior, > because of uninitialized unsigned value > // > ,boost::make_counting_iterator(std::numeric_limits<unsigned>::max()) // > must be different from above value > ),40,std::ostream_iterator<unsigned>{std::cout,", > "}); > } > > =====8< The thing that bothers me about this is that a default-constructed iterator frequently is an end iterator; indeed, that's the assumption that filter_iterator appears to be making if no end iterator is supplied. So if the default-constructed iterator is not an end iterator, and you're wrapping it with a filter_iterator and not explicitly supplying the end, I'd say you're automatically asking for trouble, regardless of whether you're using a counting_iterator or not. On the other hand, I'm not sure what a sensible default value is for counting_iterator< [integral type] >; 0 is almost never going to be the end value; std::numeric_limits<>::max() might be a better choice, but really the end value is context-dependent. I'm not opposed to giving defined behavior to a default constructed counting_iterator< [integral type] >, I'm just not sure what that defined behavior should be. Maybe I'm over-analyzing it... - Jeff

1 0