February 2014 - Boost-users

interprocess library deadlocked
by Hernán Leoni 10 Feb '14

10 Feb '14

Hi everyone, I'm using boost interprocess, and in some cases I get the interprocess library deadlocked, now I'm using internal timed locks, so, I get timeout exception Fortunately it is not that frequent, but it is an issue I need to resolve before going live, When this happens the shared memory cannot be used at all, so it needs to be removed and a new one must be created. I'm not sure yet why these internal mutexes are getting locked and never unlocked, it may be happening with some process being killed unexpectedly. I was wanting to see if anybody has already figured out which approach is better for this case. Thanks a lot, Hernan

1 0

Regarding post: Using Boost.Test in Iteration
by Deepa S 10 Feb '14

10 Feb '14

Hi Gennadiy/Gaurav, Wanted to check if there was any resolution to the error indicated in this post: http://lists.boost.org/boost-users/2010/06/59564.php I am running into a similar issue and would like to know how to resolve it. Thanks, Deepa

2 1

Attn: We need GSoC 2014 mentors for Boost!
by Niall Douglas 09 Feb '14

09 Feb '14

Dear Boost community, We are now just five days away from the closing date for Google Summer of Code 2014 registration. Last year's GSoC was a huge success, with Google funding work on Boost to the tune of $35,000 and paying over $8,000 in costs for three representatives of Boost to attend the GSoC mentor's summit at Google HQ. As part of filling in our application for 2014, we must supply to Google a list of potential GSoC mentors and potential GSoC projects for summer 2014. We do NOT want a repeat of 2012 when our application was rejected due to an insufficiently long list of prospective mentors and projects! At the time of writing, we have just *eight* prospective mentors, and we need more to be sure of being approved by Google! Therefore, if you think yourself able to mentor a student doing some work on Boost this summer, *please* consider adding a description of the proposed work item and your name to the list at https://svn.boost.org/trac/boost/wiki/SoC2014. If you want to know more about mentoring a Google Summer of Code funded student work before you nominate yourself, please feel free to ask on the main Boost developers mailing list boost(a)lists.boost.org. Thank you in advance for your time. Niall Douglas (Boost Google Summer of Code admin) Boris Schäling (Boost Google Summer of Code backup admin)

2 4

Boost.Asio ssl session reuse
by Krzysztof Trzepla 09 Feb '14

09 Feb '14

Hello, Is there a way to cache ssl session using Boost.Asio library, so that it can be later on reused with an abbreviated handshake? If so, I will appreciate any example. -- Regards, Krzysztof Trzepla

1 0

[Boost.Python] Wrapping a heterogeneous container
by Mario Lang 08 Feb '14

08 Feb '14

Hi. I have finally figured out how to (mostly) wrap a container of boost::variant<> in boost::python: http://blind.guru/boost_python-and-boost_variant.html The only problem left is the fact that I have to disable the use of reference semantics (proxies) in vector_indexing_suite. If I pass false for the NoProxy template parameter of vector_indexing_suite<>, I can no longer dereference container elements: >>> import bpv >>> bpv.make_vector()[0] Traceback (most recent call last): File "<stdin>", line 1, in <module> TypeError: No Python class registered for C++ class boost::variant<a, b, boost::detail::variant::void_, ...> However, I have a converter registered with to_python_converter<> that handles extracting the proper Python object from a variant. This works if NoProxy is true. However, if NoProxy is true, container elements do not have reference semantics, so in-place modifications will be lost. To me, it also looks like the TypeError raise above is sort of bogus, since I actually used to_python_converter to register conversion for this particular type. So I guess there are different types of registries involved, but the message does not indicate which registry is did refer to. Does anyone know if I can enable reference semantics for a heterogeneous container wrapped with indexing_suite? -- CYa, ⡍⠁⠗⠊⠕

1 0

Apology: RE: FileSystem: crash on OSX 10.8 with create_directory
by Tim Burgess 08 Feb '14

08 Feb '14

Hi, I must apologise for the stupidity of this question - if I'd have surrounded the call with a try/catch block I'd have figured it out sooner, so it's my own bad coding. Now all I need to figure out is how to programmatically create a directory with admin rights to get around the permission denied exception when trying to create files/directories within /library/application support/ on the Mac. Best wishes. Tim Burgess From: Boost-users [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Tim Burgess Sent: 07 February 2014 16:32 To: boost-users list Subject: [Boost-users] FileSystem: crash on OSX 10.8 with create_directory Hi, My app crashes out when I run the following: boost::filesystem::path myPath = wxStandardPaths::Get().GetConfigDir().ToStdWstring(); if (myPath.has_filename()) { myPath.remove_filename(); } // Add the company name and check for directory existence myPath /=wstrRBCompanyName; if (!boost::filesystem::exists(myPath)) { // The following test displays the path I'd expect to see (/library/Raised Bar in my case) wxMessageBox( myPath.native(), wstrErrorTitle, wxOK | wxICON_EXCLAMATION); // Create the required directory - displays a message box indicating deletion of a lock file here and the app exits boost::filesystem::create_directory(myPath); } This code appears to work fine on Windows, but fails on OSX 10.8. I should point out that I'm a blind programmer and I'm struggling to get Apple's VoiceOver utility to interact in a useful manner with Xcode 5's debugger and that's why I'm using a wxMessageBox to try and track down my failure. Best wishes. Tim

1 0

Generating PDF from QuickBook?
by Nat Goodspeed 08 Feb '14

08 Feb '14

I'm sorry, this must be a FAQ, but my Google search turned up mail from 2009 advising me to cd to libs/xxx/doc and run ../../../b2 pdf When I do that in a new modular Boost tree, I get the following errors: /Users/nat/boost-git/tools/build/src/tools/boostbook.jam:704: in construct from module object(boostbook-target-class)@1466 error: Cannot build documentation type 'pdf'. /Users/nat/boost-git/tools/build/src/build/targets.jam:1332: in class(a)basic-target.generate from module object(boostbook-target-class)@1466 /Users/nat/boost-git/tools/build/src/build/targets.jam:757: in generate-really from module object(main-target)@1471 /Users/nat/boost-git/tools/build/src/build/targets.jam:729: in class(a)main-target.generate from module object(main-target)@1471 /Users/nat/boost-git/tools/build/src/build/targets.jam:272: in class(a)project-target.generate from module object(project-target)@1460 /Users/nat/boost-git/tools/build/src/build-system.jam:707: in load from module build-system /Users/nat/boost-git/tools/build/src/kernel/modules.jam:289: in import from module modules /Users/nat/boost-git/tools/build/src/kernel/bootstrap.jam:139: in boost-build from module /Users/nat/boost-git/boost-build.jam:17: in module scope from module Is there some incantation I can apply to produce PDF output?

5 7

Confused with the example students_t_single_sample.cpp
by lizy10b 08 Feb '14

08 Feb '14

Hi there, I got a problem with the example "students_t_single_sample.cpp". The corresponding document is at: http://www.boost.org/doc/libs/1_55_0/libs/math/doc/html/math_toolkit/stat_t… Actally, I am confused with the mathematical logic rather than coding. The document presents a table that shows four hypothesis and tests, I am OK with the frist two, but for the third and the fourth ones: The original text: Hypothesis 3: the sample mean is less than the true mean Test 3: Reject if CDF of t > significance level, cdf(dist, t) > alpha I think p-value should be the integral from t to +Infinite: cdf(complement(dist, t)), So reject Hypothesis 3 when cdf(complement(dist, t)) < alpha, or say cdf(dist, t) > 1-alpha The original text: Hypothesis 4: the sample mean is greater than the true mean Reject if complement of CDF of t > significance level: cdf(complement(dist, t)) > alpha Similarly, the p-value should be the integral from -Infinite to t: cdf(dist, t), so reject Hypothesis 4 when cdf(dist, t) < alpha Is that right? Thanks. Zhiyu Li 2014-01-11 lizy10b

3 3

Newbie math question - roots of a distribution
by Benjamin Ward (ENV) 08 Feb '14

08 Feb '14

Hi Boost list, I've just installed and started to use boost for various problems and I've hit on a math problem I'd like to solve with it but I'm unsure on how to achieve it. I've gone through the documentation and have worked out how to easily compute the value from a cdf of a binomial distribution: #include <boost/math/distributions/binomial.hpp> #include <iostream> int main(){ int numbertrials = 4164; double successprob = 10e-6; int numbersuccesses = 75 boost::math::binomial mydistribution(numbertrials, successprob); double ans = cdf(mydistribution, numbersuccesses); std::cout << "The answer is " << ans << std::endl; return 0; } But now I'd like to find the root of the cdf, in other words, I want to find the value of successprob, from numbertrials and numbersuccesses and ans. Let's define ans as 0.95, by finding the root of cdf(mydistribution, numbersuccesses) - 1, using a root finding algorithm, I should be able to get the value of successprob. I've tried reading the roots documentation and looking at some examples but they use simple functions like x * sin(x) - there is only one variable and it is x, the binomial cdf has quite a few variables - number of trials, number of successes, probability of a single success. How can I define a function with boost::math::binomial that could be used with boosts root libraries to find the probability of success from the cdf, given I know ans, the number of successes, and the number of trials? Thanks, Ben W.

2 3

intrusive data structures with custom pointer class
by Matei David 07 Feb '14

07 Feb '14

Hi, I would like to use some boost::intrusive data structures (lists and trees) in a project I'm working on, but I ran into trouble trying to customize the existing classes to my needs. I decided to post my problem here, in case somebody knowledgeable of boost::intrusive *implementation details* can assist me. Thank you in advance. Background: In the project I'm working on, I need to manage a large number of objects in memory. The data is dynamic, so I need to be able to alloc&dealloc objects, in addition to keeping the objects in various lists&trees (for which I'm trying to use boost::intrusive). The RAM usage is large (think, >100GB), but the number of objects is smaller than 2^32. So one thing I'd like to do is to replace 8-byte raw pointers with 4-byte offsets into a deque. Doing this would solve the alloc/dealloc part. However, I now need to customize boost::intrusive containers to work with these 4-byte offsets instead of raw pointers. Initial solution that doesn't work: The most important limitation of these 4-byte indirect pointers (say to a type T) is that if you fully dereference one to T&, you can no longer reconstruct the indirect pointer (with operator&) in constant time. I.e. converting a real-pointer to indirect-pointer must be avoided. Thus, to make methods such as insert(T&) work, I had to roll out a reference-replacement class in addition to the pointer-replacement class. This is what I initially tried: - Given a type T, I rolled out pointer-replacement class Ptr<T> which holds a 4-byte int, and upon operator*, it produces a Ref<T> object. - A reference-replacement Ref<T> class similarly holds a 4-byte int, it implements operator& which produces the associated Ptr<T> object, and it also provides an operator T& () conversion which dereferences the object from the deque. - With these, I tried a custom traits struct looking like this: value_type = node := T node_ptr = pointer := Ptr<T> reference := Ref<T> Problems: Even though intrusive data structures are designed with the stated goal of leaving memory management to the user, intrusive lists&trees do allocate one extra node object that they use as "header". The code then takes the address (&) of this node and uses it in comparisons as a node_ptr. Because of this, my custom traits above could not compile. The only way I could find to make such containers work for my needs is basically a hack: define "node" to be a "pointer holder" object (neither member nor base of "value"), that allocates a "value" struct from the deque upon construction, that responds to operator& by returning the held pointer, and that otherwise acts as a value&. From what I could gather reading the docs, boost::intrusive containers allow some type of customization via the traits structs. However, I see some drawbacks: A. The traits classes seem to be there to support 2 existing customization types, namely the situation where "node" is a base class of "value" (implementing "base hooks") and the situation where "node" is a member of "value" (implementing "member hooks"). My situation represents a 3rd customization type. B. Even though the traits classes allow one to specify pointer types, it's possible (?) that under the hood the classes were designed to support only raw pointers and smart pointers, not the indirect offset-like pointers I need. Given A&B, it seems to me that in my customization I would need 2 different hacks. In the interest of code stability, I started wondering whether it is feasible to adapt existing boost::intrusive data to my needs (my first instinct) as opposed to simply rolling out my own rbtree. Questions: 0. Obviously: Have you seen a similar situation being dealt with before? Maybe there's an existing solution out there. My reluctance to rolling out my own intrusive rbtrees is due to the fact I'd have to test them, as opposed to using a known implementation. 1. Are boost::intrusive data structures supposed to support custom pointer classes (neither real pointers nor smart pointers)? For instance, in 1.46 I found this code in boost/intrusive/detail/tree_algorithms.hpp static node_ptr uncast(const_node_ptr ptr) { return node_ptr(const_cast<node*>( ::boost::intrusive::detail::boost_intrusive_get_pointer(ptr))); } So despite all the custom traits structs defining node_ptr and const_node_ptr, under the hood you have a line of code which assumes node_ptr can be constructed from node*. This is the expensive real-pointer to offset-pointer no-no I mentioned earlier. However, intrusive lists worked just fine in 1.46, which suggests this was simply a bug. 2. Can you point me to an example of defining and using custom pointer class with boost::intrusive? To enable compilation (of lists only) in 1.46, I had to customize std::iterator_traits (that part I understood), but then also boost::pointer_to_other and boost::intrusive::detail::smart_ptr_type. Even though the code compiled, I don't know what those are doing exactly, so I'm weary of them. 3. In 1.55, I have a few more pages of errors, I think mainly due to these pointer customization issues I don't fully understand. Something seems to have changed from 1.46 to 1.55. Is there a reference about what changed? Or is there a portable way of rolling out such a custom pointer class? Many thanks, M

2 1