November 2011 - Boost - lists.boost.org

Standard library Technical Report 2 proposals for Boost libraries
by Beman Dawes 18 Dec '11

18 Dec '11

The C++ committee hasn't decided whether to do a TR2 or to move directly on to the next standard. Either way, there will be a call for library proposals, but the details will be up in the air at least until the next meeting, which will be in Kona, Hawaii, in February. For the moment, the term "TR2" is being used as a surrogate for moving the standard library forward regardless of exact details. In discussions, library working group (LWG) members have made it clear they want a filesystem library based on Boost.Filesystem, and a networking library based on Boost Asio. Individual members of the committee have voiced interest in other Boost libraries, large and small. If you are a Boost developer (or user), and there is a Boost library you would like to see proposed for TR2, now is a good time to start thinking about writing a preliminary proposal to the committee. The point of such a proposal is to gauge interest before putting in a lot of work on a full-blown proposal only to find the LWG isn't interested. See http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2005/n1810.html for the original TR2 call for proposals. While the details will change, it is still a reasonable introduction to what is required. A preliminary proposal would just say "To be supplied" for the "Proposed Text for the Standard" section. It is very helpful to do at least the initial presentation of such a proposal in person. That said, John Maddock and Peter Dimov had many proposals accepted for TR1 without ever attending a standards committee meeting. If you don't plan to present a proposal in person, the best bet is to ask someone who is familiar with the library to present it. Feel free to contact me publicly or privately with questions about proposals or information about attending committee meetings, --Beman

7 18

[local] "true" lambdas without C++11 but only on GCC
by Lorenzo Caminiti 17 Dec '11

17 Dec '11

Hello all, A curiosity: On GCC compilers (but _only_ on GCC compilers) it is possible to combine Boost.Local and the non-standard GCC "expression statements" feature to implement lambda functions. In other words, it is possible to define a Boost.Local closure within a GCC expression statement: std::vector<int> v(3); v[0] = 1; v[1] = 2; v[2] = 3; int factor = 10; std::for_each(v.begin(), v.end(), LAMBDA(int x, const bind& factor) { std::cout << x * factor << std::endl; } LAMBDA_END ); This is just a curiosity because it is NOT standard C++03 and it will ONLY work with GCC compilers. #include <boost/local/function.hpp> #include <vector> #include <algorithm> #include <iostream> // WARNING: This macros only work on GCC!!! #define LAMBDA_WITH_RESULT(result_type, ...) \ ({ /* open GCC statement expression */ \ result_type BOOST_LOCAL_FUNCTION_PARAMS(__VA_ARGS__) #define LAMBDA(...) \ LAMBDA_WITH_RESULT(void, __VA_ARGS__) #define LAMBDA_END \ BOOST_LOCAL_FUNCTION_NAME(BOOST_PP_CAT(lambda, __LINE__)) \ BOOST_PP_CAT(lambda, __LINE__); \ }) /* close GCC statement expression */ int main() { std::vector<int> v(3); v[0] = 1; v[1] = 2; v[2] = 3; int factor = 10; std::for_each(v.begin(), v.end(), LAMBDA(int x, const bind& factor) { std::cout << x * factor << std::endl; } LAMBDA_END ); return 0; } I might mention this code example as a curiosity in Boost.Local docs, currently it's not mentioned there. Thanks. --Lorenzo

2 3

New libraries implementing C++11 features in C++03
by Jeffrey Lee Hellrung, Jr. 17 Dec '11

17 Dec '11

Within the discussion for the review of the proposed Boost.Local library, Hartmut Kaiser raised a concern that I think should be addressed more broadly by the community, if possible. I quote Hartmut: > Ok. However this raises a more serious question. Should we as the > Boost community still encourage solutions and libraries solely for > portability with ancient compilers? I'd say no, but YMMV. Boost will > be still around 2, 5, or 10 years from now. What's the utility of > adding such a _solely_ backwards oriented library from this POV? Indeed, I ask the community, do such libraries belong in Boost? If so, is the bar for acceptance of such libraries automatically and necessarily higher than libraries that introduce genuinely new capabilities in C++11? If so, what additional criteria must such a library meet? I add two additional notes to help put the discussion in context: - Within the proposed Boost.Local review discussion, several individuals pointed out that they will likely have to work professionally on C++03 compilers for at least a few more years. - Recall that Boost has several (recent) libraries that could easily be labelled as "transitional" libraries: Move, Atomic (at least proposed, not sure if accepted), Container. Further, several older libraries are now part of the C++11 standard (e.g., Thread (right?)). - Jeff

20 45

[chrono] i/o to locale or not to locale
by Vicente J. Botet Escriba 10 Dec '11

10 Dec '11

Hi all, currently Chrono provides a limited localization interface for duration and time_point I/O that is not satisfactory. The current implementation Boost.Chrono V1.2 (Boost 1.48) is based on the alternative A1 (see below) in which the hour,minute,second and the epoch can be localized. This could give the impression that chrono i/o is localized, but as Aryton as signaled it is not enough. I'm working on redesigning the chrono i/o in order to make a proposal with Howard Hinnant to the C++ committee. Two main alternatives A1- follow the formatting of recognized international body as ISO. A2- Add localization of specific units and epoch to the preceding format. It seems that the localized alternative seems to have less chances to be adopted, as no other library in the standard is providing these kind of facilities, and locale message facility is not completely satisfying. Things could change if we had a concrete proposal for improving the standard locale library (in which Aryton seems to be working IUC). More detail on the two alternatives: A1- Provide formatted i/o following the ISO 80000 http://en.wikipedia.org/wiki/ISO/IEC_80000, http://en.wikipedia.org/wiki/ISO_31-1, The format for duration could be <value> <unit> With unit been day, hour, minute, second, decisecond, centisecond, ..., decasecond, hectosecond, ... Note that the unit is always in singular form and in English. Next follows several valid inputs for duration 3 day 9 hour 1 second 0 second -1 second 3.77 millisecond 34 [1/30]second time_points could be formatted as <duration> <since epoch> where <since epoch> is a string that identifies the starting point of the epoch's clock, e.g. since boot, since process startup, .... The string associated to each clock are fixed and configurable specializing the trait clock_string Next follows several valid inputs for time_point. 123456 second since boot A1.a We can also consider ISO 8601. IMO, an acceptable format for durations could be a variation of the extended format independently of the date. <hours>:<minutes>:<seconds>[.fractional_seconds] Note that <hours> here is not limited to 24, and gives the total number of hours. A1.b An additional alternative could be to represent a duration as following a variant of the period representation: P[n]Y[n]M[n]DT[n]H[n]M[n]S in which the year and month are represented by days. E.g. P[n]DT[n]H[n]M[n]S[n]F P1234DT12H30M5S555F A2 - Make durations and time point i/o locale dependent. Based on the previous alternative, add the possibility to add specific locale facets that could format a duration following the pattern <value> <unit> <unit> <value> <hours><:><minutes><:><seconds>[<.>fractional_seconds] Here <value> and <unit> will be formatted following the specific locale and in this case the unit could have specific plural forms. For example in French the user could specialize the facets so that she could get 1 jour 3 jours 9 heures 1 seconde 0 secondes -1 seconde 3.77 milliseconds 34 [1/30] de seconde and in Spanish 1 segundo 0 segundo -1 segundo Add the possibility to add specific locale facets that could format a time_point following the pattern <duration> <epoch> <epoch> <duration> system_clock::time_point is a special case that should be formatted as a localized date, relative to the epoch 1900/01/01. A3 - Add on top of A2 some _byname facets. The fist alternative is in line with the money standard locale category that uses standard symbols, as $ or USD. It has the drawback that the <value> could be localized while the <unit> is not, e.g. 1,234 second Note the use of ','. The third alternative is IMO closer to the user expectations, but needs a better locale messages/by_name facet. The second alternative is a pragmatic one that left the possibility to build A3 when a satisfying messages/by_name facet is available. Boost.Chrono 2.0 in the trunk is a prototype of A2, that doesn't provides facets by locale name, only a default English implementation. That means that for the time been, the user needs to imbue a locale with her specific facet to get the expected localization (See french.cpp example in the SVN repository) What the Boost community think about this? Towards which alternative should Boost.Chrono move? Should Boost.Chrono wait until a concrete standard proposal is accepted? Thanks in advance, Vicente

2 6

[msm] deferral stack overfolw
by Richard Szabo 08 Dec '11

08 Dec '11

Hi Christophe just playing around with event deferral I have found a problem which leads to a stack overflow it ends in an infinite loop if I want to defer 2 events in the same state. see attached example it somehow similar to the problem we had earlier but that was fixed ... I have ran this test on MSM trunk version from yesterday. Cheers Richie

2 8

[stl_ext_adv] ] Interest in advanced data structures
by Vadim Stadnik 08 Dec '11

08 Dec '11

Hi all, Advanced data structures (ADS) offer useful STL extensions that are not possible or not efficient with basic data structures, such as arrays, linked lists and red-black trees. The project https://github.com/vstadnik/stl_ext_adv http://dl.dropbox.com/u/51041088/stl_ext_adv_doc/index.html focuses mainly on a generalized and unified computer representation of sequences using the augmented B+ trees. It provides the following STL extensions: - The sequence container supports random access iterators and a union of interfaces of STL containers: vector, list and deque. This sequence offers the advantage of efficient update operations with logarithmic running time compared with linear running time of the same operations of std::vector. - The associative containers (set, multiset, map and multimap) with random access iterators. - The enhanced random access iterators effectively address the fundamental problem of the iterator invalidation. - The function accumulate() with logarithmic computational complexity can significantly improve performance of many practically important algorithms and applications, such as the numerical integration, the statistics, the data analysis etc. The examples demonstrate how to improve the computational complexity from O(N) to O(log N) in the following applications and problems: - numerical integration. - calculation of mean value and standard deviation; - calculation of weighted mean value; - testing if any number of consecutive elements are equal; - counting intersections of one interval with sequence intervals; Is there an interest in this project and implemented STL extensions? Interest in potentially useful, but not yet implemented algorithms and specialized containers: 1. The function accumulate() with constant computational cost. With this function some user algorithms, such as the numerical integration, can come close to the absolute theoretical limit of the computational cost. The function can be implemented through iterators enhanced with functionality of accumulators. For more details, see the section “Random access iterators” in the documentation. 2. The functions sequence::splice() with logarithmic running time in the worst case. Note that functions std::list::splice() provide in many cases constant computational cost, however, in the worst case the cost is linear. 3. Any other potentially useful specialized containers, iterators or algorithms? Augmented B+ trees are versatile and powerful data structures. They can be designed to maximize efficiency of particular algorithms and programming solutions. The limitation of specialized containers is that the computational cost of some other operations may increase. About augmented red black trees. The augmented red black trees can also support various STL extensions. The red black trees are yet out of the scope of this project for the reasons explained in the documentation. However, historically augmented red-black trees are probably the first ADS that could extend STL, since the method of augmenting red black trees was described as early as in 1990 [T.H. Cormen, C.E. Leiserson and R.L. Rivest.” Introduction to Algorithms” (1st ed.). MIT Press and McGraw-Hill, 1990]. This means, in particular, that the very first version of STL could provide associative containers with random access iterators. Also, as I understand the previously submitted project https://github.com/boost-vault/Containers/blob/master/countertree.zipimplem… containers based on the augmented red black trees. It is hardly possible to make an exact prediction, but without a doubt future versions of STL will benefit from advanced data structures. Regards, Vadim Stadnik

11 29

[chrono] steady_clock efficiency
by Kenneth Porter 06 Dec '11

06 Dec '11

I expect most people will use boost::chrono::steady_clock to measure performance. I've been considering switching from my raw calls to the Windows peformance counter APIs to the steady_clock. I'm a little worried about the overhead of what should be a non-intrusive facility. In examing the source, I see that steady_clock::now() reads the performance counter and then multiplies it by the reciprocal of the counter's frequency (acquired on first use of the clock) to return the time in nanoseconds. How expensive is this 64 bit multiply on a 32 bit system? How hard would it be to have now() return a time_point that stores the raw tick value and only converts it to nanoseconds when a calculation in real units is required? (Comparison to another time_point could still be done in performance counter ticks, saving the multiply.) Is it worth the extra complexity to demand-convert ticks to nanoseconds? Source is in boost_1_48_0\boost\chrono\detail\inlined\win\chrono.cpp

5 15

[units] - learning to use
by Janek Kozicki 05 Dec '11

05 Dec '11

Hi, I wrote a short function that integrates Huyghens spherical wave for diffraction. (This is a single slit experiment btw). I have some questions, marked with (*) double integral( const quantity<length> x // x - position where light intensity is calculated , const quantity<length> R // R - distance from slit to screen , const quantity<length> D // D - slit width , const quantity<length> lambda // lambda - wavelength , double steps // steps - number of integration steps // (*) ) { complex<double> i(0,1); complex<double> a(0); static const double pi = 3.14159265358979323846; // (**) quantity<wavenumber,complex<double> > k=2*pi/lambda; quantity<length> d=-0.5*D; // d - position on the slit quantity<length> dd=D/steps; // dd - width of integral step // integrate from -d to d: exp(i*k*r)/r for( ; d<= 0.5*D ; d+=dd ) { quantity<length> r=sqrt(pow<2>(R)+pow<2>(x-d)); a += (dd*exp(i*k*r)/r).value(); // (***) } return pow<2>(abs(a)); } (*) it should be `int steps` but if I write it this way I get a compilation error here: quantity<length> dd=D/steps; // dd - width of integral step That looks like a conversion problem. (**) why do I need to define pi ? I expected to find it somewhere, but I couldn't. (***) it seems strange that I need to take .value() here. The result is dimensionless, why can't it directly convert to complex<double> ? There is a risk that .value() would be expressed in millimeters or angstroms or whatever (maybe depending on dd), and then .value() would return a wrongly scaled number? Another question, why this works: quantity<length> wavelength ( 632.8 *nano*meters); and this does not? quantity<length> wavelength = 632.8 *nano*meters; Of course I know about extra copying, and waste of CPU resources on execution, when compiler isn't able to optimize this away (and usually it is able to do that). But why the heck this simple assignment can't work? PS: I am attaching a short test version of this program, it's in polish, but I hope you don't mind. best regards and thanks in advance for your help -- Janek Kozicki http://janek.kozicki.pl/ |

5 22

[1.49.0] New libraries?
by Beman Dawes 03 Dec '11

03 Dec '11

What new libraries are ready for 1.49.0? What is that state of any new libraries that want to go in 1.49.0? Are they in trunk and do the developers think they are ready to be merged into release? The release managers would like to have a four week beta period of 1.49.0. That means we have to get going on new libraries right way if we are going meet our early February release schedule. --The release team

6 12

[thread] patch to allow custom stack size.
by Brett Lentz 03 Dec '11

03 Dec '11

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Patch and (very) rudimentary usage example is attached. A couple months ago, I sent mail to the list looking to see if there's interest in getting some patches merged from Phusion Passenger. (See: http://marc.info/?t=131722868200003&r=1&w=2 ) So, here's the first patch I'd like to try to get committed. On pthread-based systems, it allows for the user to request a custom stack size. This is useful in Passenger's case where they want to reduce the VM size without requiring the user to hassle with ulimit settings. Passenger spawns many threads rather than using a thread pool for performance reasons. As a result of this design decision, they also need the ability to tune the stack size. Ticket #5956 already exists for this patch. I've verified that the patch works against the current (1.49) codebase out of SVN. What next steps need to be taken before this is committed? - -- - ---Brett. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQEcBAEBAgAGBQJO1lZEAAoJEEAzW/nB31+3fx4H/iulwAM7vGlZW56xsuezTXPC 6SSuYzUlkyVeCngpireVyD0+Um/ACFBM8FSf76cRnewX5449T01IycmX3fUe+z7g DjjUCQAHJ/9pdfrqEHz0d7wZcZbeBHUm2zggdDGcn703JQSagg2Q5d5mz48RU8/J ixTe/oAOAaXk5TQo/92fHrAmtpwfQON4tPsS7Om0kJ5vcv3PbIqQ0Q4eBKRJCVHI /CFfc3xXwSaR1xUkwhPHtjydlIWAMo0+bdKgJE5HGnn6fQSU5jlZ+ihDyajOmM88 ueaIcWi7pa3OZwLzrUNG0xIH4m7wtk0ga8edym4+ZvuovUi4vYRA1k4dIzBOtvg= =0dQJ -----END PGP SIGNATURE-----

6 14