- Boost - lists.boost.org

[soc2009] Proposal for serialization library
by Hrvoje Ban 19 Mar '09

19 Mar '09

Dear Boost developers, Congratulations on being accepted as a mentoring organization. I'd like to work on Boost serialization library for this year Google Summer of Code. The feature I'd like, JSON archives, was suggested last year, but to my knowledge wasn't implemented. As part of my college project I have written LR(1) parsers for BBCode and Wiki Creole markup using JavaCup/JFlex, and also LL parser for BBCode using ANTLR. I'm not (jet) comfortable with Boost.Spirit, but since I'm quite familiar with (E)BNF notation, I can learn it quickly. I'd like to know are you still interested in this proposal and will there be mentor for me, if I get accepted? Sincerely, Hrvoje Ban

4 4

Exposing command line options within Jamroot
by brook＠biology.nmsu.edu 19 Mar '09

19 Mar '09

I am helping to maintain the pkgsrc packages for Boost. When compiling the shared libraries it is important to inject rpath information into them so that the linker will correctly find dependencies. Thus, I would like to add a <dll-path> clause to the following stanza in Jamroot. # Install just library. install stage-proper : libs/$(libraries)/build : <location>$(stage-locate)/lib <install-dependencies>on <install-type>LIB <install-no-version-symlinks>on ; explicit stage-proper ; The obvious thing is to use something like <dll-path>$(libdir) or <dll-path>$(lib-locate) since those variables should be derived from the various bjam command line options or defaults. However, this variable (as well as the other obvious candidates like $(prefix), $(exec_prefix), etc.) do not seem to be defined within the context of Jamroot. The only place I am aware of these command line arguments being handled is deep within boost/tools/build/v2/tools/package.jam, in which case they are part of the install rule. I am guessing that burying the logic like that prevents the variables from being useful anywhere else, but I know very little about how bjam actually works. Thus, I believe there are two options: - Modify Jamroot itself or some other *.jam file to expose these variables (prefix, exec_prefix, etc. or some appropriate subset of them) so that command line options can better control what bjam does. - Modify Jamroot to introduce a new command line argument (e.g., --dll-path) and variable ($(dll-path)) to use for this specific purpose. I would like the solution to be one that will be acceptable into the Boost code base. Please advise me on how to proceed. Thank you very much. Cheers, Brook

2 1

Re: [boost] [gsoc][student] Proposal
by Joseph Gauterin 19 Mar '09

19 Mar '09

You're proposal has significant overlap with Crypto++ (public domain). You might want to investigate some sort of code sharing, or at least have a look at how they've done things.

2 1

[gsoc][student] Proposal
by Kasra 19 Mar '09

19 Mar '09

Hi, My proposal is to have a crypto library that provides high-level facilities. The library has different encapsulation layers. Within the low-level (internal) there are algorithms that share the same concept i.e. class block_cipher_type: + setkey(const void* key, size_type key_size) -> void; + encrypt(void* output, const void* input) -> const void; + decrypt(void* output, const void* input) -> const void; class stream_cipher_type: + create() -> void; + update(const void* input, size_type output) -> void; + update_final_block(const void* input, size_type output) -> void; and etc. Also entropy sources that use kernel specific entropies with addition of asynchronous generators. Finally a std:: iostream interface for cryptographically transformed streams. That use a transformer for the cryptographical transformation. Any interest??? -- Kasra

4 3

range transformed -adaptor incompatible with lambda functions
by Sandeep Gupta 19 Mar '09

19 Mar '09

Hi, I would be nice if range-transformed worked with lambda functions. Is this an intended behaviour. To illustrate the following two statements fail for_each(myvec|boost::adaptors::transformed(arg1+mask), std::cout<<arg1<<' '); for_each(myvec|boost::adaptors::transformed(boost::lambda::_1 + mask), std::cout<<boost::lambda::_1<<' '); But this is fine: for_each(myvec|boost::adaptors::transformed(double_int()), std::cout<<boost::lambda::_1<<' '); where double_int hand created function object with result_type declaration. Thanks sandeep

2 2

Boost.Asio and Steinberg ASIO
by Alexander Churanov 19 Mar '09

19 Mar '09

Hi folks! I wonder how Boost.Asio is related to Steinberg ASIO? It seems that ASIO is an older technology and it's name is a registered trademark. http://www.steinberg.net/en/company/steinberg_technology.html It's a source for confusion. For the first time I've noticed the Boost.Asio library, I was thinking it's an STL-compliant facade for Steinberg ASIO :-) Alexander Churanov

3 2

[gsoc][student] Proposal
by Kasra 19 Mar '09

19 Mar '09

Hi All, I was wondering how the community is feeling about a new library proposal? Since few month ago I started talking about boost.crypto on the mailing list and considered its implications with few (namely Kevin Sopp). At present the library has more than 60K LOC. Any thought? :D -- Kasra Nassiri

2 1

[asio] global timer cancellation?
by Brian and Victoria 19 Mar '09

19 Mar '09

Hi, I've got various objects that make use of the asio deadline timer to schedule events. The problem is that there isn't a central repository of timer references that I can walk to cancel them when shutting down asio. I'd like to be able to just tell asio to cancel any pending work and have the thread(s) return. Is this possible? I came across the io_service::work class but haven't yet found an example of how it works. Thanks, Brian

2 1

dynamic_bitset count
by Andreas Klein 19 Mar '09

19 Mar '09

Hello. Sorry for multiple posts. Hopplfully I got I now right. ----------------------------------------------------- Dear boost developers, I checked the implementation of boost::dynamic_bitset::count and find that it uses a table look up. This is not optimal. First on 64-bit machines divide-and-conquer algorithms are faster than table look up and save even memory. Second if the bitset contains several words, there are for both 32 and 64 bit machines population count algorithms that are faster than the simple loop over all words used by boost. Together with Y. Edel I developed a new version,which saves 60% against the simple algorithm. We put the code online (http://cage.ugent.be/~klein/popc.html) A disadvantage of our implementation is that we relay on aggressive optimization (-O3) of the compiler. If you are searching a compromise between speed and code complexity. The first iteration of the Harley-Seal-method (25% faster than the simple loop) would be a candidate. You find this and other population count algorithms on our homepage, too. If there is interest I can add one of the advance algorithms in boost. Best wishes Andreas Klein

2 2

[shared_ptr] calling shared_from_this from constructor
by Пюрецкий 18 Mar '09

18 Mar '09

Hi, A have just get latest version from SVN. And now my code does not work. I found problem in a changed version of shared_ptr and shared_from_this (revision 51485). Now calling shared_from_this in constructor throw's boost::bad_weak_ptr exception. -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.

2 1