- Boost - lists.boost.org

[iostreams] Cannot compile against zlib 1.2.4
by Lars Viklund 13 Apr '10

13 Apr '10

Hi list. zlib 1.2.4 was released in mid-March, reorganzing the sources a bit. They removed gzio.c and added in a bunch of new source files, something the Jamfile for iostreams does not particularly appreciate. I submitted a ticket [1] with a somewhat tested solution (remove gzio from the list of sources), but I have no idea whether it's the right way to solve it as I'm not aware of how the guts of zlib_compressor and friends work. How does one go forward with something like this, as I don't think I've seen much activity from the iostreams camp at all lately? It would be awfully nice to get some kind of fix into 1.43.0, as it currently prevents anyone interested in building zlib automagically via bjam from doing so. One thing I guess might get in the way is whether this should support older versions of zlib sources. In my opinon no, as this only applies to people who want to build from their own freshly unpacked zlib source tree, not affecting people who use prebuilt binaries from their distro or downloads. [1] https://svn.boost.org/trac/boost/ticket/4091 -- Lars Viklund | zao(a)acc.umu.se

2 1

VS2005 SP1 Hard Debugger Crash in Class Destructor with Boost.Filesystem
by Jack Yee 13 Apr '10

13 Apr '10

Hello, I have a very strange crash involving the use of auto-linking of the boost::filesystem (1.38/1.42) and putting breakpoints in a unmanaged class destructor. It crashes hard enough that the task must be stopped through the Task Manager. Below is the test case that causes the crash: #include "stdafx.h" #include <vcclr.h> #include "boost/filesystem.hpp" // Note 1 public ref class ClassOne { public: ClassOne() {} ~ClassOne() {} }; class ClassTwo { public: ClassTwo(ClassOne^ p) {one = p;} ~ClassTwo() {one = nullptr;} // Note 2 gcroot<ClassOne^> one; }; int _tmain(int argc, _TCHAR* argv[]) { ClassOne ^one = gcnew ClassOne(); ClassTwo *two = new ClassTwo(one); delete two; // Note 3 return 0; } Note 1 -- Commenting out this header and setting breakpoint at Note 2 will be OK Note 2 -- Putting a breakpoint here with "boost/filesystem.hpp" will cause VS2005 to crash Note 3 -- Putting a breakpoint here with "boost/filesystem.hpp" will be OK BUILD CONFIGURATION: * Common Language Runtime Support (/clr) enable * Preprocessor defines: WIN32;_DEBUG;_CONSOLE; * Standard include and library search paths to your Boost installation. Other Useful information: * Running the program without breakpoints doesn't indicate any errors or non-zero exit codes. * Tried both 1.38 and 1.42 boost libraries * Debugger crash does not happen under VS2008 * Same crash results on Windows XP Professional, Vista Business and Windows 7 machines. Event Viewer Output: Faulting application devenv.exe 8.050727.867, time stamp 0x45d2c842, faulting module OLEAUT32.dll, version 6.0.6001.18000, time stamp 0x4791a74f, exception code 0xc0000005, fault offset 0x00003e56, process id 0x8ec, application start time 0x01cadab2a1430262. Hopefully the example is concise enough that I can get some other willing people to try it out and see if they get the same results. Apologies in advance if there is a typo in the source code as the work machines are isolated from connecting from the Internet. Thanks! Jack

1 0

GSOC 2010 - BGL Graph Connectives.
by Tri Nguyen 12 Apr '10

12 Apr '10

Hi, I am Tri Nguyen from South Plains College. I am interested in Graph Connectives Project of BGL. In this writing, I would like to ask two questions, and then propose my idea of doing four Connectives: union, join, intersection, and difference. My first question is: How big is the graph? Or students will try to do as big as possible. My second question is: What type of data-structures (matrix or list) will be used? I attached my ideas in the PDF file below. Please have a look. I appreciate all responses, Thank you, Tri Nguyen.

4 15

GSoC2010 Sweepline Algorithm
by Andriy Sydorchuk 12 Apr '10

12 Apr '10

Hi, Boost Community! I am interested in participation in GSoC 2010. The project I liked the most is the generic implementation of Sweepline Algorithm. Basically fast and robust Sweepline Algorithm requires implementation of next data structure: 1) Event Queue (priority queue). 2) Sweep Line (self-balancing binary search tree). 3) Output datastructures (voronoi vertices, triangulation edges, etc.). Based on that I got next questions: - Which kind of self-balancing binary trees is better for that kind of approach? I am thinking about Red-Black Trees in prefer to AVL trees, as sweepline algorithm requires more adding and deletion operations. And AVL trees are better for look up operations. - STL library includes two generic implementations of self-balancing binary trees (red-black): map, set. I am quite new with Boost library, but probably there are some more specific self-balancing binary tree implementations there? - For Voronoi diagram and delaunay triangulation input data is set of points. What about input of the Medial Axis problem? -- Best, Andrii Sydorchuk

4 33

[BGL] Revisiting Parallel BGL
by Sandeep Gupta 12 Apr '10

12 Apr '10

Hi, I raised the issue of parallel BFS not working correctly here http://lists.boost.org/boost-build/2009/10/22631.php. The issue remained unresolved. I revisited it for last two days and the incorrectness in likelihood is probably due to initialization of distance property map for all non-start vertex to be infinity. Hope this helps. Thanks Sandeep

3 17

How to use mutable_queue with non basic types
by Emili Hernàndez 12 Apr '10

12 Apr '10

Hi all, I've used mutable_queue with basic types (double and int) with no problem. Now I trying to do it with my own type: typedef MyObject entry_t; typedef vector<entry_t> storage_t; typedef MySortingCriterion comp_t; typedef boost::identity_property_map prop_map_t; typedef boost::mutable_queue<entry_t, storage_t, comp_t, prop_map_t> queue_t; queue_t Q(100, comp_t(), prop_map_t()); The definition compiles ok but when I perform an operation such Q.push(obj) //MyObject obj The following errors appear: C:\boost\boost_1_42_0/boost/property_map/property_map.hpp: In function 'Reference boost::get(const boost::put_get_helper<Reference, PropertyMap>&, const K&) [with PropertyMap = boost::typed_identity_property_map<unsigned int>, Reference = unsigned int, K = boost::shared_ptr<CAStarNode>]': C:\boost\boost_1_42_0/boost/pending/mutable_queue.hpp:94: instantiated from 'void boost::mutable_queue<IndexedType, RandomAccessContainer, Comp, ID>::push(const IndexedType&) [with IndexedType = boost::shared_ptr<CAStarNode>, RandomAccessContainer = std::vector<boost::shared_ptr<CAStarNode>, std::allocator<boost::shared_ptr<CAStarNode> > >, Comp = astarnode::SGreater, ID = boost::typed_identity_property_map<unsigned int>]' ..\CAStar.cpp:62: instantiated from here C:\boost\boost_1_42_0/boost/property_map/property_map.hpp:318: error: no match for 'operator[]' in '(const boost::typed_identity_property_map<unsigned int>&)((const boost::typed_identity_property_map<unsigned int>*)(& pa))[k]' C:\boost\boost_1_42_0/boost/property_map/property_map.hpp:553: note: candidates are: T boost::typed_identity_property_map<T>::operator[](const T&) const [with T = unsigned int] It seems the problem is with property_map. Anyone knows how to solve it?? Thank you very much, Cheers, Emili

2 4

[GSOC] summary of .BigInt library details.
by Ankit Malpani 12 Apr '10

12 Apr '10

hi all, thank you Scott McMurray, Vicente and Bruno for the prompt feedback regarding my proposed BigInt library ideas .(Sorry for the delayed reply as was busy with mid-semester exams.) In general this is what i have gathered from your responses regarding what exactly is being expected by the community from a BigInt library for Boost. 1. frontend based off Boost.Proto 2. backend implemented with the help of Boost libraries like Boost.Math. 3. probably provide two levels of implementations. ( although it would really help if you could weigh the pros and cons of both and eliminate one the the two) 3a. looking at bit-level implementation(less space and probably more execution time) 3b.) digit level manipulation (more space - and probably less execution time). i.e. for an n-digit number is it ok to use up [n/2 to n] bytes of space or necessary to stick to log n amount of space for a BigInt object. my questions are :- 1. Please give your feedback/concerns regarding the above design. 2. Any other libraries like Boost.Math which would help with the backend. -- Ankit Malpani 3rd Year Undergraduate | BE. (Hons) Computer Science BITS Pilani Goa Campus

5 5

[Bind] Function arithmetic (sample impl inside)
by ivan.lelann＠free.fr 12 Apr '10

12 Apr '10

Hi, It's my first message here, so please excuse possible mistakes. I found that Boost.Bind offers logical operators but no arithmetical operators. Looking into boost/bind/bind.hpp, and using "logical_not" impl, I wrote the following naive patch against version 1.40. It is certainly not as robust as Boost exspects, it is quickly tested, but it seems to do the job, as you can see in the test. Please use attached documents as you see fit. Regards, Ivan. PS: attached main.cpp is compiled with g++ 4.4.1, -std=c++0x, "my" bind.hpp in same directory, and boost 1.40 in include path.

2 1

Re: [boost] Crypto Proposal
by Chad Seibert 12 Apr '10

12 Apr '10

> Hi, > > Perhaps this link will work: > > http://socghop.appspot.com/gsoc/student_proposal/show/google/gsoc2010/chads… > Yep, it worked, thanks. > Thanks to all who gave advice, please ask if you have any questions! > It all sounds very good. Can you give an example of what porting some > Botan functionality would look like? Would it be as simple as replacing > the namespace Botan with boost::crypto/whatever in the source code? > An example of what one of the Botan tutorial examples would look like in > the new world would be nice. > Honestly, I think there is no point if you can't commit to part 2. > However, splitting it up into GSoC/non-GSoC is smart on your part. How > do you plan to commit to part 2? > Thanks and good luck :) Here's an example of boostified code, from one of the examples on the site. It's not too different from the original code (it's just commented out); this is because I don't want it to be a chore to port from Botan to Boost.Botan. Also, it is intuitively designed and I see no reason to change it. Of course, if it needs redesigning, we can discuss that afterwards. Also, there are at least 600 files to move that total 2 mb's of source, so it'll take a while, even if it is just changing namespaces and the like. But I'll get started on part 2 if part 1 finishes early. As for committing to part 2, I have some (rough) guidelines for the work to be done afterwards. I'm also taking a lighter courseload so I can spend more time working. I don't know how else to convince you. Anyways, thank you for considering my proposal! Chad Seibert #include <boost/botan.h> #include <iostream> #include <memory> int main(int argc, char* argv[]) { if(argc != 5) { std::cout << "Usage: " << argv[0] << " <passphrase> " << "<ca cert> <ca key> <pkcs10>" << std::endl; return 1; } try { const std::string arg_passphrase = argv[1]; const std::string arg_ca_cert = argv[2]; const std::string arg_ca_key = argv[3]; const std::string arg_req_file = argv[4]; //AutoSeeded_RNG rng; boost::botan::rng::auto_seeded rng; //X509_Certificate ca_cert(arg_ca_cert); boost::botan::x509::certificate ca_cert(arg_ca_cert); //std::auto_ptr<PKCS8_PrivateKey> privkey(PKCS8::load_key(arg_ca_key, rng, arg_passphrase)); std::auto_ptr<boost::botan::pkcs8::private_key> privkey(pkcs8::load_key(arg_ca_key, rng, arg_passphrase)); //X509_CA ca(ca_cert, *privkey); boost::botan::x509::ca ca(ca_cert, *privkey); // got a request //PKCS10_Request req(arg_req_file); boost::botan::pkcs10::request req(arg_req_file); // you would insert checks here, and perhaps modify the request // (this example should be extended to show how) // now sign the request //X509_Time start_time(system_time()); //X509_Time end_time(system_time() + 365 * 60 * 60 * 24); //X509_Certificate new_cert = ca.sign_request(req, rng,start_time, end_time); boost::botan::x509::time start_time(system_time()); boost::botan::x509::time end_time(system_time() + 365*60*60*24); boost::botan::x509::certificate new_cert = ca.sign_request(req, rng, start_time, end_time); // send the new cert back to the requestor //std::cout << new_cert.PEM_encode(); std::cout << new_cert.pem_encode(); } catch(std::exception& e) { std::cout << e.what() << std::endl; return 1; } return 0; } _________________________________________________________________ The New Busy is not the too busy. Combine all your e-mail accounts with Hotmail. http://www.windowslive.com/campaign/thenewbusy?tile=multiaccount&ocid=PID28…

4 3

[preprocessor] BOOST_PP_IS_UNARY()
by Lorenzo Caminiti 12 Apr '10

12 Apr '10

Hello all, Do you know what are the drawbacks of using BOOST_PP_IS_UNARY()? If you have used it, what is your experience with its portability? BOOST_PP_IS_UNARY() is *not* part of the Boost.Preprocessor public API as it is #defined in boost/preprocessor/detail/is_unary.hpp. However, Boost.Preprocessor, Boost.Typeof, and Boost.Spirit seem to use BOOST_PP_IS_UNARY() at least internally: $ find /usr/include/boost/ -name "*.[hc]pp" -exec grep -Hn "BOOST_PP_IS_UNARY" {} \; /usr/include/boost/preprocessor/facilities/apply.hpp:24:# define BOOST_PP_APPLY_I(x) BOOST_PP_EXPR_IIF(BOOST_PP_IS_UNARY(x), BOOST_PP_TUPLE_REM_1 x) /usr/include/boost/typeof/template_encoding.hpp:61: BOOST_PP_IS_UNARY(x),\ /usr/include/boost/spirit/utility/rule_parser.hpp:826: BOOST_PP_EXPR_IIF(BOOST_PP_COMPL(BOOST_PP_IS_UNARY(elem)),elem) In addition, Paul Mensonides mentioned the following in http://lists.boost.org/Archives/boost/2004/08/70238.php > Paul Mensonides wrote: > I know. They (and other types of detection macros) are very useful. Right now, > the Borland configuration is used by the IBM preprocessor and the Sun > preprocessor also. I'd like to know whether or not those macro definitions work > correctly on those preprocessors before I add it. Though those macros aren't > part of the public interface, they are part of the internal interface. I.e. > they are used internally as if they were a public interface (though they are > used carefully on some preprocessors). They are stable, and the worst thing > that might happen in the future is that they will be moved to a different > directory. Thus, you can use them, but beware of VC and versions of MWCW prior > to v9. They in particular have severe problems with expansion order. E.g. VC > will even fail this test: My Boost.Contract code internally uses BOOST_PP_IS_UNARY() and it works fine on GCC 4.2.4 (Ubuntu Linux), MSVC 8.0 (Windows XP), and GCC 3.4.4 (Cygwin) but I have not tried other compilers (including Borland's). Thank you. Lorenzo

1 1