July 2009 - Boost - lists.boost.org

[move] moving temporaries of copyable types
by klaus triendl 10 Aug '09

10 Aug '09

Hi, I've recently taken a look at the Boost.Move library and it seems that it is only possible to make 1) either non-copyable types movable or 2) make copyable types move-aware by explicitly moving them. Now, I've discovered for myself that it is possible for pre-c++09 compilers to make copyable types *implicitly* movable by detecting temporaries that can be moved. So, my question is whether this is well-known and this feature was deliberately excluded from Boost.Move or nobody has discovered it yet. It would be great to have support for copy- and (implicit) move-semantics. Klaus Triendl

5 9

[proto]: rewriting POOMA and Blitz++
by Chandrashekhar Kumar 10 Aug '09

10 Aug '09

Hello, At some time Eric expressed that all the expression template based libraries of boost (notably Phoenix, Spirit) has to be re-written using proto. I think that this target is kind of achieved by now. Can something be done along the similar lines for re-writing the top-notch libraries like POOMA and Blitz++ as a complete examples of proto applications? Does anyone has idea about how huge this undertaking can be? Thanks, Chandra

4 7

"I want everything" should build "out of the box" (on Windows)
by KTC 09 Aug '09

09 Aug '09

Users on Microsoft Windows should be able to build boost with everything, both in terms of all supported variants and all boost libraries, with a single command to bjam assuming they have all the optional supporting libraries available. An user download boost, and want to build all the libraries that require building go read the Getting Started Guide <.../getting_started/windows.html#invoke-bjam> and see the after getting bjam they use the command bjam --build-dir=build-directory toolset=toolset-name --build-type=complete stage So they enter that into the command prompt and see lots of warnings and notes about missing optional components. The warnings and notes itself is fine and informative. Actually, it's incomplete as it stand as it doesn't tell the user about Boost.IOStreams optional component requirement of bzip2 & zlib headers. Okay, the user go and install all the optional components the warnings and notes tell them. After installing Python, ICU (building it first), Expat (building it first), and MPI (say MS-MPI), they try again. This time with the command becoming bjam --toolset=... --build-dir=... --build-type=complete -sHAVE_ICU=1 -sICU_PATH="..." -sEXPAT_INCLUDE="..." -sEXPAT_LIBPATH="..." -sBZIP2_SOURCE="..." -sZLIB_SOURCE="..." stage Nope, MPI requires modification of user-config.jam or building it on its own with --with-mpi. Why? No other optional components build that way. So they look to see how they do that, and does it. Second attempt and ... nope. New to 1.39 is the errors error: link=shared together with runtime-link=static is not allowed error: such property combination is either impossible error: or too dangerious to be of any use due to the changes made as a result of <http://lists.boost.org/Archives/boost/2009/04/150675.php>. (Recently noted on <http://groups.google.com/group/boost-list/browse_thread/thread/2a7bd07468b5…>.) While catching those errors is good, maybe it should be fixed as well such that it doesn't come up in the first place if someone do a --build-type=complete with ICU. So now the users have to build Boost.Regex separately from the rest of the library (after finding how to pass the link= & runtime-link= options). By the way, at this point I would like to point out how it's strange that -sHAVE_ICU=1 mean to build with ICU, but so is -sHAVE_ICU=0 .... :S And ... nope, even separately Boost.Regex still wouldn't build without errors. No one told them they need to rename icuin.lib, icuind.lib, icuind.pdb & icudt.lib from ICU to icui18n.lib, icui18nd.lib, icui18nd.pdb & icudata.lib respectively. With Boost.Regex built, we turn to the rest of the library with --without-regex. After waiting a while, we discover ... nope, Most of the library is built but not Boost.Graph. Sorry, this time no one told them they need to change Boost.Graph's Jamfile.v2 file with "<find-shared-library>expat" to "<find-shared-library>libexpatwMT" and adding "<define>XML_STATIC". Boost is great, it's free, it's flexible and I understand it's developed for free by volunteers, but its "installation" for someone that want every libraries could be improved. :-) Regards, KTC -- Only two things are infinite, the Universe and Stupidity. And I'm not quite sure about the former. - Albert Einstein

15 56

[Interprocess::Semaphore] Deadlock on more producers - one consumer problem
by Cosimo Calabrese 09 Aug '09

09 Aug '09

Hi to all, I've a buffer of 10 positions; 4 producers put elements on the buffer and only 1 consumer pick them; the producers generate elements very fastly, but the consumer picks the element very slowly. I've decided to synchronize the producers and the consumer with a boost::interprocess semaphore initialized to 10. Before to put an element in the buffer, the producer execute a wait() in the semaphore; the 4 producers put quickly 10 elements in the buffer, so they all block on the wait(). When the consumer picks an element from the buffer, it performs a post() on the semaphore. Here is the problem: the post() unlocks only one of the 4 producers. Consider this scenario: - the producers are all in wait state; the semaphore counter is 0; - the consumer picks an element and perform a post(), that unlocked the producer_1; the semaphore counter is 1; - the producer_1 is unlocked but it doesn't produce any element; the other producers are locked; - the consumer picks the other 9 elements in the buffer and perform the 9 post() on the semaphore, but the semaphore counter becomes 2, 3, 4, ... 10, and so the semaphore doesn't signal the other producers, that remains in lock state(). --> DEADLOCK. Is there a way to unlock all the producers with the boost::interprocess::semaphore? Or I have mistaken the approach of the problem? Thanks to all, Cosimo Calabrese.

5 27

[1.40] timeline?
by John Maddock 08 Aug '09

08 Aug '09

Do we have a revised timeline for 1.40 yet? I'm asking because the regex lib is ready for a substantial update (provided trunk tests cycle OK obviously!). Thanks, John.

6 13

Boost.Plot? - Scalable Vector Graphics
by Paul A. Bristow 07 Aug '09

07 Aug '09

Jake Voytko started this library during a 2007 GSoC and I have now expanded it quite a bit, and begun to use it for my own purposes (including preparing the plot for the Boost.Math library functions and statistical distributions). Before I do any more work on it, I'd appreciate feedback from Boosters on how useful they think it might be, perhaps as a Boost library? Although it still has plenty of wrinklies (the calculation of positioning of titles, legends, axes labels and ticks labels is intricate - to say the least). And there are not-a-few nasty warts (this may be the 1st version to throw away?), so it is in no way review-ready. But I believe it is working well enough for you to try to use it - and to give me feedback. There are of course hundreds of programs for plotting. What is special about *this* library is that it is in C++ and it has a Boost license. It allows you to add just a few lines of code to the end of your C++ program to take data from any STL container (all or part) and produce a plot as a Scale Vector Graph (.svg) file. (No need to export data to a file and suck into Excel ;-) Applause!) .svg files can be displayed by all good Internet Browsers (and even by IE8 with an Adobe add-in). They are very small but high quality, highly zip compressible, and equally suitable for printing to viewing on mobiles (only the features of Tiny SVG are used.). It takes as little code as svg_2d_plot my_plot; my_plot .plot(map1, "Series 1").stroke_color(blue) .plot(map2, "Series 2").stroke_color(red") .autoscale(map1); my_plot.write("./2d_simple.svg"); but you can have hours of fun trying out hundreds of options to produce some rather fancy plots. For your amusement, attached are a few (mostly contrived) examples (with tasteless garish colors just to highlight the possible options). There are also lots and lots of other examples. (Note that the 20 plots occupy a mere 30 kb when zipped). Unusually, there is lots (too much? - 6Mbyte hyperlinked and indexed PDF!) of documentation, produced using the C++ in QuickBook with Doxygen reference info and John Maddock's auto-indexing. But there are hundreds of options, so you will need all the help you can get from tutorials, examples and indexes. A novel feature is the built-in handling of an 'uncertain' type (based on Evan Marshal Manning, C/C++ Users Journal, March 1996 page 29 to 38) including a double 'most likely' value, an 'uncertainty' estimate, roughly standard deviation, and an estimate of 'degrees of freedom'. This can be used to display confidence ellipses and label values "1.23 +-0.01 (21)", for example. (Of course, you can make much better confidence estimates using the Boost.Math library). You can view and can get code from the Boost sandbox https://svn.boost.org/svn/boost/sandbox/SOC/2007/visualization Docs are available to build as html and pre-built as pdf (patience - 6 Mbyte!) https://svn.boost.org/svn/boost/sandbox/SOC/2007/visualization/libs/svg_plot /doc/pdf/svg_plot.pdf or http://tinyurl.com/coqlbg (or ask if you want zips instead). Paul PS Thanks to all those who have offered support, especially John Maddock of course. It has stress-tested the Quickbook - Doxygen indexing system ;-) I haven't managed to make the index work perfectly yet and I'm sure it can be improved. PPS If I were doing this again, I would not use 'derived' trick to permit re-use of functions by 1D, 2D and boxplots - it causes the Intellisense of MS VisualStudio IDE 'Intelligence Circuits' to melt (and sometimes crash - workaround - delete the .ncb file) making the nice 'goto definition' feature ineffective and makes debugging difficult (as does the use of set and get functions). I suspect it makes compile times longer too. There must be a better way - suggestions welcome. But I *would* use the chaining feature - I think it makes a very nice syntax for the user - but tell me what you think. --- Paul A. Bristow Prizet Farmhouse Kendal, UK LA8 8AB +44 1539 561830, mobile +44 7714330204 pbristow(a)hetp.u-net.com

14 36

[BGL] To customize the graph traversing
by Cosimo Calabrese 06 Aug '09

06 Aug '09

Hi to all, I'm studying the BGL, and I've a question. I've noticed that: - the traversing algorithms, like breadth first search, only decide how to traverse the graph, but do not perform specific calculations; - if I want to perform any calculation during the traverse, I can do it passing a custom visitor to the algorithm. But the question is: how can I customize the traverse method? The algorithms always explore enterely the graph. EXAMPLE: every edge (u, v) of a graph have a property "edge_can_traverse_t" that says on which of adjacent edges starting from v I can continue the traverse. _C _F o| /| o / o / A---->B---->D o o o >G \ \ \ \ _\| _\| E H ----------------------------------------- | edge | property edge_can_traverse_t | | | (vector<edge>) | ----------------------------------------- | (A,B) | (B,D) , (B,E) | | (B,C) | empty | | (B,D) | (D,F) , (D,H) | | (B,E) | empty | | (D,F) | empty | | (D,G) | empty | | (D,H) | empty | ----------------------------------------- For example, if I come from the (A,B) edge, I want that the algorithm consider only the (B,D) and (B,E) edge, and skip the (B,C) edge. The solution I found is to modify the breadth_first_search algorithm, adding a predicate to the visitor and performing a control when the algorithm examines the outer edges of the current vertex. Something like this: //============================================================================== // PREDICATE (in the visitor) //============================================================================== template <class IncidenceGraph, class PredecessorMap> bool can_traverse_edge( edge_descriptor e, IncidenceGraph &g, PredecessorMap p ) { typedef graph_traits<IncidenceGraph>::vertex_descriptor VertexDescriptor; VertexDescriptor u = source( e, g ); // select the source of the edge EdgeDescriptor eFrom = edge( p(u), u ); // select the previous visited edge if ( e is contained in the edge_can_traverse_t property of eFrom ) return true; else return false; } //------------------------------------------------------------------------------ //============================================================================== // breadth_first_visit //============================================================================== template <class IncidenceGraph, class Buffer, class BFSVisitor, class ColorMap> void customized_breadth_first_visit (const IncidenceGraph& g, typename graph_traits<IncidenceGraph>::vertex_descriptor s, Buffer& Q, BFSVisitor vis, ColorMap color) { [.....] put(color, s, Color::gray()); vis.discover_vertex(s, g); Q.push(s); while (! Q.empty()) { Vertex u = Q.top(); Q.pop(); vis.examine_vertex(u, g); for (tie(ei, ei_end) = out_edges(u, g); ei != ei_end; ++ei) { if ( !vis.can_traverse_edge( *ei ) ) continue; [.....] } //------------------------------------------------------------------------------ Does it exist a better mechanism to control the traverse with the BGL? Or I must implement my solution? Best Regards, Cosimo Calabrese.

3 13

Interest in a "Heterogenous Container" system?
by Christian Schladetsch 05 Aug '09

05 Aug '09

Hello, After a discussion on a different thread, I came up with an implementation allowing containers to store and retrieve heterogenous instances, and provide correct deep cloning of those containers. A working motivational example is here http://tinyurl.com/larqyy. I have yet to do a comparison with container<boost::any>. This is just an initial proof-of-concept, and there remains a lot of work to do if it were to be a complete proposal. Before I did that work (especially since I have another proposal in the wings), I wanted to ask if there would be any interest in a completed system? The system is based on a modified ptr_container branch (also in the sandbox - I've contacted Thorsten about this), and and allocator adaptor system to deal with cloning. It allows one to place object pointers into a container using syntax such as: /// a vector of heterogenous objects template <class Alloc = std::allocator<char>, class Base = common_base > struct vector; heterogenous::vector<> v; v.push_back<T0>(); // add a default-constructed object of type T0 v.push_back<T1>(a0,a1,...,an); // add an object of type T1, passig a0..an as ctor arguments T1 &obj = v.ref_at<T1>(1); // get a reference of type T1& to the object at index 1 or throw bad_cast heterogenous::vector<> v2 = v; // copy the container; concrete types are preserved assert(v2.is_type_at<T1>(1)); There are many outstanding issues; it may not fly in the end, it may not work for all container types and semantics, but I would like some early opinion on whether it is even worth my pursuing further. One initial issue that may break any interest is that instances put into the container must be of a type T that derives from heterogenous::base<T>. Regards, Christian.

11 29

[Range] Boost and N2930?
by Beman Dawes 04 Aug '09

04 Aug '09

N2930, Range-Based For Loop Wording (Without Concepts), provides begin() and end() non-member function templates for standard library types. It also provides general range access non-member templates: // 24.7 range access [iterator.range] template<typename C> auto begin(C& c) -> decltype(c.begin()); template<typename C> auto begin(const C& c) -> decltype(c.begin()); template<typename C> auto end(C& c) -> decltype(c.end()); template<typename C> auto end(const C& c) -> decltype(c.end()); template<typename T, size_t N> T* begin(T (&array)[N]); template<typename T, size_t N> T* end(T (&array)[N]); (Until the official WG21 site gets updated, you can see a copy of n2930 at http://mysite.verizon.net/beman/n2930.html) Should Boost.Range and/or RangeEx be modified and/or extended in light of N2930? Can Boost.Range, RangeEx, or some other Boost component help developers and users who want to take advantage of N2930 in C++03 now, and do so in a way that transitions smoothly to C++0x as language and library available? --Beman

3 2

[function] should it require rtti?
by Emil Dotchevski 04 Aug '09

04 Aug '09

Line 223 in function_base.hpp uses typeid. Can this be avoided? If not, perhaps it can be refactored to use the workaround found in sp_typeinfo.hpp? Emil Dotchevski Reverge Studios, Inc. http://www.revergestudios.com/reblog/index.php?n=ReCode

5 5