- Boost - lists.boost.org

[context] mini-review period ended
by Giovanni Piero Deretta 12 Jan '12

12 Jan '12

Hi all, The boost.context mini-review period ended sometime betweend yesterday and today depending on your timezone. We didn't get many reviews, but we had some interesting discussions and I should be able to reach a conclusion. If anybody is still completing a review, please let me know me ASAP, otherwise I should be able to wrap everything up by this weekend and announce the results next week. Thanks for everybody that partecipated in the review and thanks to Oliver for submitting his library. -- gpd

1 0

Re: [boost] Condition Variables, POSIX signals and system time changes
by Johan Borkhuis 12 Jan '12

12 Jan '12

Vicente, > I'm reworking the timed wait interface using Boost.Chrono. In this case > you could use a steady clock to mean a MONOTONIC clock. Unfortunately my > refactoring did use pthread_condattr_setclock to force a monotonic clock > while waiting on the condition. Even if this should not be the source of > the problem as it works when you don't generate the signals, I think > that this will be better when available. I will need some time to make > this working but I hope I could release it for boost 1.50 :( Looking forward to that. For the moment we have a workaround, but I will keep an eye on changes in this area, so we can move back when this becomes available. >> The demo application shows where things go right and wrong, so hopefully >> you should be able to reproduce the issue. > > I have not understood the example completely. How do you change the time > forward or backward? Could you point me where or how do you do? by hand? Sorry, should have mentioned that. I change the time externally by hand, using the date command. When I change the time backwards (so an earlier time) the problem occurs. When I change the time forward again to a time equal to or before the time I changed the time the application continues. >> >> Again, thank you for your quick response. >> >> > Thanks for reporting this and for the application example. > Vicente Thank you for your work on this. Kind regards, Johan Borkhuis

1 0

problem in trunk with file system library
by Robert Ramey 12 Jan '12

12 Jan '12

I'm seeing this message on bunch of tests "/sierra/Sntools/extras/compilers/gcc-4.3.4/bin/g++" -ftemplate-depth-128 -O0 -fno-inline -Wall -pedantic -g -fPIC -Wno-long-long -ftemplate-depth-255 -DBOOST_ALL_NO_LIB=1 -DBOOST_ARCHIVE_TEST=binary_archive.hpp -DBOOST_FILESYSTEM_DYN_LINK=1 -DBOOST_LIB_DIAGNOSTIC=1 -DBOOST_SYSTEM_DYN_LINK=1 -I".." -c -o "/scratch/kbelco/boost/results/boost/bin.v2/libs/serialization/test/test_array_binary_archive.test/gcc-4.3.4/debug/test_array.o" "../libs/serialization/test/test_array.cpp" In file included from ../boost/filesystem.hpp:36, from ../libs/serialization/test/test_tools.hpp:5, from ../libs/serialization/test/test_array.cpp:22: ../boost/filesystem/v3/operations.hpp:139: error: extra ???;??? ../boost/system/error_code.hpp:214: warning: ???boost::system::posix_category??? defined but not used ../boost/system/error_code.hpp:215: warning: ???boost::system::errno_ecat??? defined but not used ../boost/system/error_code.hpp:216: warning: ???boost::system::native_ecat??? defined but not used This is provoked by the following code:symlink_perms = 0x4000 // on POSIX, don't resolve symlinks; implied on Windows};BOOST_BITMASK(perms); // line 139//--------------------------------------------------------------------------------------//Would it be possible to get this fixed. I don't what the exact fix is. Robert Ramey

2 1

Fwd: How to build the documentation
by Pedro Larroy 11 Jan '12

11 Jan '12

---------- Forwarded message ---------- From: Pedro Larroy <pedro.larroy.lists(a)gmail.com> Date: Sun, Dec 25, 2011 at 7:35 PM Subject: How to build the documentation To: boost-docs(a)lists.boost.org Hi I created a small patch for libs/python/doc/tutorial/doc/tutorial.qbk How can I check that it builds correctly and generates propper html? Regards. Pedro.

5 8

Re: [boost] Condition Variables, POSIX signals and system time changes
by Johan Borkhuis 11 Jan '12

11 Jan '12

Vicente, Thank you for your quick reply. I did not get the message from the mailing list, so I copied your questions into this mail. As requested I opened a ticket (ticket# 6377) and I attached the demo application to it, with a small description. > Do you mean that you are not using Boost.thread here? No, the application is running in one thread. The condition variables are used as an alternative sleep method. We discovered that other sleep-methods are even more vulnerable to time changes, and also have different responses to Posix signals. The condition variables continue to work correctly, even with Posix signals, however in combination with time changes there are some issues. In our real application we are using boost threads quite heavily though. > What do you changed exactly? Could you post a patch? I attached the demo application to the issue. It includes the regular Boost condition variable use and the changed implementation. It also includes the Posix timer code to generate the signals. I did not change the Boost code (yet), as I am not that familiar with it. > Please, could you post the test that fails and the one that succeeds, it > will help me a lot. The demo application shows where things go right and wrong, so hopefully you should be able to reproduce the issue. Again, thank you for your quick response. Regards, Johan Borkhuis

1 0

Re: [boost] Proposal for a Differential Evolution C++ library
by Adrian Michel 11 Jan '12

11 Jan '12

On 1/9/2012 12:09 PM, Simonson, Lucanus J wrote: > > > I think that numerical optimization is more general than its application to > machine learning. While you argue that you'd like to see the library go > deep in machine learning I don't see anything wrong with a general purpose > numerical optimization library. However, DE is more general than numerical > optimization, while the library seems to be an application of DE to > numerical optimization. I'm not sure that DE is even one of the best ways > to do numerical optimization (I'm pretty sure genetic algorithm is a bad > way, so I'm skeptical), and I know I'd like a more general interface to a DE > engine than one designed for numerical optimization problems. I'd like to > see the library go wide for max applicability rather than deep. Hopefully > that will keep it out of all the different cans of worms that could be > opened in each application domain. Put numerical optimization and machine > learning applications in the example code of the library and submit just the > core engine as the library itself. Narrowing the scope of a library submission > is usually the path to success. > I forwarded Ken Price (one of the DE co-authors) some of the comments I received for his view on things and here's his reply, which hopefully should clarify some of the issues that have been raised: "The skeptic is right in believing that GAs have traditionally done poorly at numerical optimization, primarily because they rely on general data manipulation operations (bit flips, exchanges, rotations, insertions, replacements...) that are not appropriate for real-parameter optimization. DE, however, was specifically designed for numerical optimization (although it has since been applied with limited success to combinatorial problems). Instead of relying on general data manipulation operations, DE's unique differential mutation operation is a real-valued vector operation that is appropriate for the numerical optimization problem domain. More than just appropriate, the differential mutation operation at the core of DE is effective because it exploits the tendency of a population to distribute itself around function contours. Randomly sampling the difference in the location of population members when they are distributed along a function's level surfaces turns out to be a very effective way to deal with some of the most challenging aspects of numerical optimization, e.g. multiple optima, parameter-dependence, slow-down at high resolution and disparate parameter sensitivities. As evidence of DE's utility as a numerical optimizer, I would offer that it has been part of Mathematica's numerical optimization toolbox for over a decade. It helped the Mathematica team win the SIAM 100-digit challenge (ten problems to ten-digit accuracy) by solving what the contest participants agreed was the contest's most difficult (numerical optimization) problem. (DE also solved one of the other problems in the contest). In the many other contests that have been held to compare numerical optimizers, DE has consistently placed at or near the top. DE is also part of the OPTIMUS package, which is highly respected and in wide use by several automobile manufactures. Alternatively, skeptics can do an Internet search and discover for themselves the enormous range of engineering problems the DE has successfully solved, nearly all of which are numerical optimization problems, often with multiple, non-linear constraints. DE has also proven effective in the domain of multi-objective numerical optimization. Finally, I would suggest that since DE is so simple and easy to implement, skeptics ought to just try it. Most people are surprised by how effective something as simple as DE can be."

2 1

[concept_check] Should OutputIterator check for output_iterator_tag?
by Ronald Garcia 11 Jan '12

11 Jan '12

Hello, In tracking down a problem with Boost.MultiArray, I'd like to write a test case that exhibits the failure. Specifically, certain iterators should be models of OutputIterator, and it's my impression that an output iterator needs to have an iterator_category that's convertible to output_iterator_tag(). However, it looks like the OutputIterator Concept checker in the concept checker library does not check that the output iterator has such an iterator_category. Is that an oversight, or is it not the case that an output iterator have a category that's convertible to output_iterator_tag? It looks like InputIterator, RandomAccessIterator concept checkers perform such a check. Thanks, Ron

3 4

Re: [boost] Proposal for a Differential Evolution C++ library
by Adrian Michel 11 Jan '12

11 Jan '12

On 1/9/2012 9:41 AM, Brian Schrom wrote: > > On 01/09/2012 09:11 AM, Paul A. Bristow wrote: >>> -----Original Message----- >>> >>> I, for one, think this looks a very useful library. >>> [...] >>> So you may need to find and drum up support from some potential (or > better still, actual) users ? >>> > I will provide a review if this library gets to review, so > far, I've not looked much at this submission. > Thanks. > I think that machine learning is an area where there is enough maturity > that implementations in Boost are desirable. > > A couple of things to consider. > > It would be desirable that the machine learning topic area be > designed a bit more comprehensively than a single library. For example, I > should be able to run a GA on the same data as the DE. I would hope > that they have the same API. This somewhat implies to me that DE/GA are > instances of training algorithms, for which, many are possible. The way I see it, the only common point between different optimization algorithms (if they are generic enough) is the objective function, which should have a very simple prototype: double f( std::vector< double > ); so in theory if different algorithms are implemented to accept the same types (in C++ sense) of objective functions, then you could run them interchangeably between GA, DE etc without any modification to the function or algorithm. > > There are many things that go into a ML solution. Data conditioning, > model training, model validation, model evaluation. Training in > batch modes, instance based, etc. Just to name a couple issues... > > Then the normal list of library desires...completely generic, optimized > for every conceivable configuration (embedded processor, server farm, > etc i.e. single process, multi-threaded, mpi, etc) > > My advice is to be prepared for the big can of worms that this topic could > potentially bring up. > I think the topic area will be a very hard sell, but would be very > useful. > There are many real-life applications that I believe would benefit immediately from a maybe less than perfect implementation, so if it were my choice, I would rather release an initial version sooner and enhance it based on experience gathered from its practical use. Adrian Michel www.tradery.com

2 2

Weak functor: Interested? (How to boost::bind to a weak_ptr)
by Vladimír Třebický 11 Jan '12

11 Jan '12

Hi! I saw people asking how to do boost::bind using a weak pointer. Generally it's not possible without first specifying what should happen if the weak_ptr cannot be converted to shared_ptr. The "weak_fn" I wrote is a small class holding a member pointer, a weak pointer and a "policy": * It is callable so you can insert it to boost::bind. * You can construct it using free function templates. * The policy says what should happen when the object is called yet the weak_ptr cannot be converted to shared_ptr. * It's implemented both using variadic templates and Boost.Preprocessor. Copy of the help comment: /** Returns a callback that can be used eg. in boost::bind. When called, it * tries to lock weak_ptr to get a shared_ptr. If successful, it calls * given member function with given arguments. If not successful, it calls given * policy functor. Built-in policies are: * * "ignore_if_invalid" - does nothing * "throw_if_invalid" - throws "bad_weak_ptr" * "return_default_if_invalid" - returns given value * * Example: * * struct Foo { * void bar(int i) { * std::cout << i << std::endl; * } * }; * * struct do_something { * void operator()() { * std::cout << "outdated reference" << std::endl; * } * }; * * int main() * { * boost::shared_ptr<Foo> sp(new Foo()); * boost::weak_ptr<Foo> wp(sp); * * boost::bind(boost::weak_fn(&Foo::bar, wp), _1)(1); * sp.reset(); * boost::bind(boost::weak_fn(&Foo::bar, wp), 1)(); * boost::bind(boost::weak_fn(&Foo::bar, wp, do_something()), 1)(); * } */ What's missing: * Cannot point to a free function. Only member functions. * ? Wdys? Vladimir.

6 16

Condition Variables, POSIX signals and system time changes
by Johan Borkhuis 11 Jan '12

11 Jan '12

I came across the following issue with condition variables. I don't know if this is an issue with my application or with Boost condition variables. First my setup. I am using Linux on an ARM system, running Boost 1.47. My application needs accurate timing, and for this I use a POSIX timer using SIGRTMIN. This part is not something that I can control, it is an external library that implements this. I use condition variables are a timing mechanism with in my software. When I use condition variables without the POSIX timer there is no problem. The timing continues correctly, even if I change the time forward or backward. When I start the POSIX timer the condition variable does not time-out when I change the time backward. It only times out when the original time (plus timeout) is reached. The code I use to create the condition variable: boost::mutex m; boost::condition_variable cond; boost::unique_lock<boost::mutex> lk(m); cond.timed_wait(lk, duration); I tested this also by extracting the code from Boost and implementing this inline. This gave the same results. When I changed the clock of the pthread condition variable to monotonic (call to pthread_condattr_setclock(&condattr, CLOCK_MONOTONIC)) the result was OK again. Is this an issue with my implementation, can I change some setting or is this a (known) issue with the Boost condition variables? If people are interested, I do have a small test application that I used to test this issue that I can post here. I don't know however what the policy is on this list wrt posting of attachments. Kind regards, Johan Borkhuis

3 2