- Boost - lists.preview.boost.org

Google Summer of Code Proposal Feedback
by Paul Kirth 25 Apr '13

25 Apr '13

Hello Boost community, I am a student interested in the Google Summer of Code (GSOC), and have a few proposals in mind. If I have posted this in the wrong place, my apologies, but applicants are encouraged to post to the mailing list, and after some research this seemed like the best list to post to. I have an original proposal for Boost, and strong interest in the Boost.ASIO idea. I've decided to seek some feedback from the community on their feelings about my idea, in the hopes that I can better define them, and increase my chances of success in being selected for the GSOC. I realize it is a bit late to begin discussing proposals, but I wasn't really aware of GSOC until recently. I had strongly considered writing a proposal for Boos.ASIO, but I fear my lack of experience with Linux, POSIX, and concurrency in general might prove to be too great of an obstacle. If the subject matter can be learned relatively quickly, I would love to discuss what such a project would require with the mentor, Niall Douglas, or a community member who has is familiar with the project. *About Me* First, let me begin with a bit about myself. My name is Paul Kirth, and I am currently pursuing a masters in Computer Science. This statement may be a bit misleading, however, as I received my BS in Civil Engineering from UCLA in 2005, and have only recently decided to pursue a course of study in Computer Science. I have for the last two years taken CS classes to prepare for Masters level courses, but am still a bit of a novice in CS. I would say that I am at about a 3rd year level in CS, proficient in C++ and python, and competent in Java, with a limited amount of assembly, and very little work in concurrency. I have not spent much time doing engineering work in the last few years, as my focus has been on my education, but I was fairly good at solving linear systems, and had a good amount of experience with design and collaboration. As an undergraduate I made the mistake of not pursuing summer internships or jobs, and as a result I found that many of my peers had a much easier time with our courses as a result of their work experience. After all, 20-30 hours a week solving engineering problems will generally make you more proficient at solving them than someone who doesn't. I do not wish to make the same mistake with my new career path, and desperately want to gain experience in professional software design, and in working on a collaborative project. I am keenly aware of my lack of experience, and really wish to use this opportunity to push myself, and gain as much expertise as possible during the few months of the summer. *Original Proposal* My original proposal probably isn't unique, and much to my chagrin as I read the hints for successful proposals it seems a popular topic. I am of course suggesting a project dealing with Trees. Rather than suggest building a generic tree, I am suggesting the start of a tree library, or the addition of several types of trees to existing libraries, as I am not sure which is more appropriate. Because trees are such a fundamental data type in computer science, I have found it difficult to understand why there are no explicit tree classes in the c++ standard. I have read the arguments that there is no standard use for a tree, and as such it cannot be made to satisfy enough users. I have read more arguments stating that std::sets and std::map use tree structures underneath their interface. Furthermore I have read in several places that it is common to use Boost.graph to make trees. Trees have myriad uses, and a library implementation should consider this heavily when designing the data structure, but trees are common enough to warrant a standard implementation in the same way as lists, queues, and stacks. I will concede that trees can be considered, and in fact are, graphs, and that many graph algorithms are very powerful tools. But trees don't always need those algorithms, and graphs don't always maintain the tree in the way we wish. Using an std::set or std::map is not the same thing as using a tree, regardless of how they are implemented. Conservatively, a linked list can be fully implemented and tested in an afternoon, while implementing a Red Black Tree would take considerably longer, yet the C++ standard has linked lists and not Red Black Trees. This seems a bit counter intuitive. We have a relativity simple data structure in the standard, and a more complex one is not. I'm sure that linked lists are more common than red black trees, but by how much? If self balancing binary trees are considered a fundamental data structure by every computer science curriculum, why aren't comprehensive, reliable implementations of them part of the standard, or widely available in 3rd party libraries like Boost? The library I am proposing should eventually be able to satisfy almost anyone requiring the use of a tree. It should have many different types of trees, Binary trees, AVL Trees, Red Black Trees, Treeps, B-Trees, Skip-Lists, etc. It should support common methods of traversal, inorder, preorder, postorder, and user defined traversal methods. Eventually, the library should support an n-ary trees and their operations. Trees in the library should also be performant, fail gracefully, be well documented, unit and integration tested, as exception safe as possible, and, someday, concurrency safe/compatible. I'm unsure of the Boost stance on inheritance is, but my initial feeling is that users should be able to use tree classes as base classes if they desire. Proposing a large library as a summer project is ridiculous, hence I am proposing a start to this library, containing a few tree classes that can be implemented, tested, and documented within a few months, and a set of standards for new entries in the library. I don't expect the whole library to immediately be accepted into Boost, but the project will aim to meet those standards, with a long term goal of eventually being included into Boost. My research has found several projects that implement single tree classes, or groups of tree classes, but no tree library. A good c++ example is the rbrees project at https://code.google.com/p/rbtrees/ , or python has a few implementations of trees, such as bintrees 0.3.0, https://pypi.python.org/pypi/bintrees/0.3.0, which is a binary tree library, containing BinaryTree, AVLTree, and RBTree classes. I feel that bintrees 0.3.0 is a good project to mimic in the early stages, as binary trees are fairly easy to implement, test, and maintain with AVL and RB Trees. I am comfortable with binary, AVL and Red Black Trees, and feel very confident in my ability to implement them. I would like to do more, and push myself this summer, but I am not sure how much work unit and integration testing are, how long they take, and how much extra time is required for classes/libraries to be refined to Boost standards. I am sure that this time is non trivial, but as a student I'm very unsure about how much time things take, or how much time they should take. In general I would like to work on a project that would expose me to more new ideas, but I really think the world deserves access to a C++ tree library that can have some guarantees for performance, and is part of a larger, well documented, highly respected library. I understand that this project is virtually guaranteed to take more that one summer's worth of effort, and I am comfortable with staying involved in the project beyond the length of GSOC, or having the work done absorbed into another library/project. Whether or not this project is accepted into GSOC, if there is support for this effort in the community I would love to help as I am able. Obviously if I am working this summer my commitment will not be able to be as substantial, but I do feel strongly about the need for standard, performant, reliable tree classes in c++. Thank you for reading my lengthy first post to the mailing list. I hope to have more discussions with this community in the future, and I am eager to hear your thoughts about how this proposal could be improved. -- Paul Kirth (310) 709-2516

1 0

proposal for date_time implementation
by Shamsher Siddiqui 25 Apr '13

25 Apr '13

hi mentors plz review my project and give me suggestion to improve to my proposal Name :Shamsher Ahmed College/University:University Insititute of Engineering &Technology,Chandigarh,India Course/Major:Computer Science & Engineering Degree Program:Bachelor of Engineering Email:mail.shamshersiddiqui@gmail.com Homepage:Not yet Availability: How much time do you plan to spend on your GSoC? I can work full time to complete this project. What are your intended start and end dates? I will start the project according to GSOC 2013 schedule but i will try to finish the project 2 weeks before so that I can have the necessary time for debugging and for unforeseen situation it occurs. What other factors affect your availability (exams, courses, moving, work, etc.) ? My semester exam will get start from 18 may and it will end on 28 may .So I will have complete time to dedicate. I don’t have any course or work during this summer . Background Information: • Please summarize your educational background (degrees earned, courses ,taken, etc.) I passed out 10+2 from Indian School Certificate in 2010 and I am currently pursing bachelor of engineering from university institute of Engineering & Technology, Chandigarh Till now I mainly focused in learning programming language (C,C++,Java). I have mastered over Java and c++ but scratchy in C.I have also taken course on Algebra, Calculus, Data Structure & Algorithm ,database etc etc. • Please summarize your programming background (OSS projects, internships,jobs, etc.) This will be my first OSS project at such large level. I have developed many things in c++ & Java. During my last summer I have done my internship in slr infotech ,a small software firm . • Please tell us a little about your programming interests : I am programming from 8th grade .In 10 and +2 board exam I obtained 96 and 97 respectively in my computer science subject. I love to code for hours .And most important thing about open-source I like about is sharing, helping and hacking code . • Please tell us why you are interested in contributing to the Boost C++ Libraries? Boost c++ Libraries are one most powerful libraries of c++ used by many well known organizations .It is cross platform library which has high efficiency And it goes together with standard template library in c++. I will get support from boost community and it is documented very well which makes it very easy to understand and if any bugs comes than it is easily recognized and eliminated , lastly while working with bright minds I will be able to learn from them and impart my experience , knowledge to other developers • What is your interest in the project you are proposing? Libraries were mystery for me when I started learning my first programming language c but after some time I was able to perform the same task without using library but there was increase in time and length of code for doing the same program.I want to develop this project because it will increase the reusability of code and in Boost community best minds works together to give the one of the best libraries of c++. Date library will be always used in every development project whether big or small and When I read this project https://svn.boost.org/svn/boost/sandbox/chr ono_date/libs/date/doc/date.html .This is small library but will be most efficient and will be able to perform all task and save time ,reusability of code is very high. • Have you done any previous work in this area before or on similar projects? No ,I haven’t done any project regarding libraries but I am using libraries in my other project. But me and my friend are developing a tagging software which creates a folder on the basis of album,artist,track number etc. We have implemented internal structure but GUI part is left which my friend is working on. and developed small games in c++ as my project assignment. • What are your plans beyond this Summer of Code time frame for your proposed work? After Summer of code ,I will remain involve with the boost c++ community and I would like to add more feature to current project.Secondly I will work with the boost community in building in their library because I love coding for long hour. I enjoy this work. • Please rate, from 0 to 5 (0 being no experience, 5 being expert), your knowledge of the following languages, technologies, or tools:  C++: 4  C++ Standard Library: 3  Boost C++ Libraries: 2  Subversion :2 • What software development environments are you most familiar with (Visual Studio, Eclipse, KDevelop, etc.) I generally use Qt creater or Visual studio depending on the operating system I am using . • What software documentation tool are you most familiar with (Doxygen, DocBook, Quickbook, etc.)? I am familiar with Doxygen Project Proposal Goals & Implemenatations 1) Construction of Date: • Date can be represented in two ways field representation or as an serial representation . In field based representation(YMD) date value is constructed from individual year,month,and day value in three integral value. In serial reprensentation(SERIAL) data values are stroed in intergral number of days from an implementation defiend epoch date, there is third hybrid version to represent the date which implemantd the both representation and it is named HH_YMD_SERIAL_X where x represents the size of types in byte. • Providing support for entering the date in form YDM(year,day,month) which is lacking currently in chrono:date library • No invalid dates can be entered considering range cheaking ,non existent date if entered bad_date exception is thrown . • Constructing constant object for day,week_day. Eg. last/may/2000 represents 31 may 2000 • Construction of meta data for date 1) Leap year flag 2) Nth day of month 3) Nth weekday month 2) Arithematic Operation On date Date airthematic can be done on three units i.e day,month,year.In chrono date ::calendar(Gregorian calendar) length of month and year is not fixed : • day arithematic 1) Adding number of days to date until range doesn’t exceed 2) Implementing last day arithmetic to avoid the bad date excepti Eg. aug/last/2011 +months(1)= sep/30/2011 aug/31/2000 +months(1)=sep/31/2000 //invalid date bad date excep thrown • month arithmetic Month arithmetic is quite complex compared to day arithmetic since number of are not fixed . 1) (month day <=28 || month day =last day ) always add month 2) month day =29 cheak for leap year and than add month 3) Month day>=30 add but don’t add if resulting month is feb 4) Using constant objects for month arithmetic. • year arithematic 1) Always add year to date considering the condition of leap year 2) Reversal of date back to original date 3) Usage of last day during year arithematic During any of the airthematic operation result should not be an invalid date if it occurs than bad_date exception is thrown. 3) InterConversion field representation & serial representation Some operation are quite fast on field but airthematic operation are much faster on serial campared field representation .So it is necessary to provide the inter-coversion so that it can be easily converted from one representation to another. 1) Constructing bi-directional conversion of serial and field by making use of static cache to optimize conversion 4) Building function for week days & date formatting 1) weekday of a date used. 2) nextweek day of date used. 3) previous weekday of date used. 4) finding number of weeks in month 5) formatting of date according of all types 5)Interoperability with other calendar Current chrono::calendar is Gregorian calendar which exist everywhere in date & time library of every programming language.(c,c++,java) and ISO 8601 and POSIX. This section provides conversion of chrono calendar to client calendar. 1) Conversion of chrono ::calendar to Julian Calendar 2) Conversion of chrono :: calendar to Hebrew calendar 3) Conversion of chrono :: calendar to hindu calendar 4) Conversion of chrono :: calendar to Chinese calendar 5) Conversion of chrono :: calendar to Buddhist calendar 6) Conversion of chrono :: calendar to Islamic calendar Hindu ,Chinese ,Buddhist ,Hebrew(Jewish ) calendars are lunisolar calendar with a very little difference. Julian, Gregorian are very much similar except the method of finding the leap year. Islamic calendar is lunar calendar(totally dependent on moon ) which is quite unpredicatable. Hence accurate conversion is rare. TimeLine Until 17 June :Community bonding , understanding boost c++ libraries ,discussing with mentor about the details of proposal . 17 June to 27 June: Construction of date 28 June to 12 july : testing &debugging of construction of date and implementing date arithemetic 13 july to 29 july (Mid term evaluation) : implementation of interconversion of serial and field representation and testing ,debugging and making more efficient the previous code develop. 30 july to 15 August :Implementation of date formatting and developing function on weekday. 16 August to 30 August: testing ,debugging and aadding more feature to weekday and implementation of interconversion of calendar. 31 August to 23 Sept:testing ,debugging ,making efficient the whole code .

2 2

Fwd: [asio]Extension for audio device service
by adrien courdavault 24 Apr '13

24 Apr '13

Hello. I make this new thread to be clearer. There is currently no way to manage audio endpoints cconnection easily. It looks like some people might find this usefull (as I do), and I've been suggested on the boost dev list to try to detail this, as an extension to Boost.ASIO. For this reason I create this thread. I'm trying to make a very basic first draft of the concepts and see if this may be a good idea. I attached here the first things I've written. This is very short and general. I would like to know: * do you think I'm going in the right direction by seing this as Boost.ASIO extension. * do you have suggestions * would someone like to participate ? Thank you

2 3

[gsoc-2013]Boost.multiprecision project
by Sagar Massand 24 Apr '13

24 Apr '13

Hi, I wanted to submit a proposal for the boost.multiprecision project. Please find attached the files:1)Proposal for the project 2)Naive code for karatsuba multiplcation in C. It would be very kind of you if you could give me suggestions, point out flaws etc. in this. Regards, Sagar Massand

1 0

Going forward with Boost.SIMD
by Mathias Gaunard 24 Apr '13

24 Apr '13

Hi, This week I presented a proposal to the C++ standards committee to provide a standard library component for SIMD computation based on the library in development Boost.SIMD (not yet a Boost library). I was hoping to get feedback on the interface and establish an API that would satisfy hardware and compiler vendors. Unfortunately the concurrency/parallelism group has decided that they do not want C++ to provide types representing SIMD registers. I'm afraid I don't quite understand the rationale for such a refusal; proposing more high-level constructs similar to valarray (or to our own library NT2) was suggested, but that's obviously a more complex and limited API, not a basic building block to program portably a specific processor unit. Development of Boost.SIMD will still proceed, aiming for integration in Boost, but standardization appears to be definitely out of the question. Any feedback of the API presented in the proposal is welcome. <http://open-std.org/JTC1/SC22/WG21/docs/papers/2013/n3571.pdf>

7 19

[Runtime parameters reference]Regarding Call Stack Trace
by Hemanth Choudary M V (RBEI/ETA1) 24 Apr '13

24 Apr '13

Hello, We are running a boost unit test using command line (cmd). The command is: C:\> /c "D:\MyExecutable.exe" --run_test=testcase1,testcase2,testcase3, --report_format=xml --report_level=detailed --log_format=xml --log_level=all >test_log.xml 2>test_report.xml The output format is in XML. I would like to have the call stack for the failed test cases. Is it possible to write the call stack trace in test_report.xml or test_log.xml ? Is there any macro or Boost API can help me achieve this? Mit freundlichen Grüßen / Best Regards, Hemanth Choudary MV RBEI/ETA1

1 0

'XPath error : Undefined namespace prefix' when building boost.test docs
by legalize+jeeves＠mail.xmission.com 24 Apr '13

24 Apr '13

[Please do not mail me a copy of your followup] I get this error when I try to build the boost.test documentation and I'm not sure where it's coming from or what to do about it. Since the HTML files are commmitted to the repository, I can't tell if the documentation was actually generated or not. D:\Code\boost\svn-boost-trunk\libs\test\doc>b2 standalone ...found 39 targets... ...updating 1 target... xslt-xsltproc-dir.windows html\standalone_HTML.manifest XPath error : Undefined namespace prefix set XML_CATALOG_FILES=../../../bin.v2/boostbook_catalog.xml "d:/code/boost/boost-tools/bin/xsltproc.exe" --stringparam boost.defaults "Boos t" --stringparam chunk.toc "/D:/Code/boost/svn-boost-trunk/libs/test/doc/src/btl -toc.xml" --stringparam example.dir "file:///D:/Code/boost/svn-boost-trunk/libs/ test/doc/src/examples" --stringparam manifest "standalone_HTML.manifest" --strin gparam manual.toc "/D:/Code/boost/svn-boost-trunk/libs/test/doc/src/btl-toc.xml" --stringparam snippet.dir "file:///D:/Code/boost/svn-boost-trunk/libs/test/doc/ src/snippet" --path "D:\Code\boost\svn-boost-trunk\libs\test\doc/src" --xinclude -o "html/" "D:\Code\boost\svn-boost-trunk\libs\test\doc\src\xsl\html.xsl" "..\. .\..\bin.v2\libs\test\doc\msvc-11.0\debug\utf-boostbook-on\src\btl.docbook" ...failed xslt-xsltproc-dir.windows html\standalone_HTML.manifest... ...failed updating 1 target... D:\Code\boost\svn-boost-trunk\libs\test\doc> Any suggestions? -- "The Direct3D Graphics Pipeline" free book <http://tinyurl.com/d3d-pipeline> The Computer Graphics Museum <http://computergraphicsmuseum.org> The Terminals Wiki <http://terminals.classiccmp.org> Legalize Adulthood! (my blog) <http://legalizeadulthood.wordpress.com>

2 7

Re: [boost] GSOC - Boost.SQL and Boost.Thread interest
by Guillaume Nuel 23 Apr '13

23 Apr '13

Hi Vicente ! > if you are interested on one of these projects, please send as soon as possible to this ML a concrete proposal.> If both subject interest you, you can make two proposals. My main choice for Thread is: Thread.Pool (Proposal n°2 on my homepage: http://immortal-pc.info/Boost.html )General idea: ( Syntax highlighting at: http://immortal-pc.info/Boost.html#threadPool_idea ) int f1( const char* i ) { std::cout << i << ':'; // ... return 'f'+1; } int f2( const char* i ) { //... return 'f'+2; } int f3( const char* i ) { //... return 'f'+3; } int f4() { //... return 20+20+2;// 42 Power ! ^^ } int f5() { //... return 'f'+5; } int main() { // I want 5 threads for my job. Boost::Pool bp(5); // Note: A great option would be to use the "main thread" for calculations. (OpenMP like) // With this idea, 5 threads needed => 4 new theads + the main thread // // Another interesting option would be to provide switch between thread and process (fork) // I love Python multiprocessing: // http://docs.python.org/2/library/multiprocessing.html#using-a-pool-of-worke… // Note: Using, C++11 Brace-Initialization => std::initializer_list // Map function will call f1("Data1"), f1("Data2"), ... // >> ONE CALL PER parameter bp.map(f1, {"Data1","Data2","Data3","Data4","Data5","Data6","Data7"} );//!< WAITING function // "apply_async" will call {f1} with the parameter "Test" five times. // >> 1 CALL for f1 ! // >> 1 CALL for f2 ! // >> 1 CALL for f3 ! // >> f1 // f2 // f3 // -> // = paralell bp.apply_async(f1, "Test"); bp.apply_async(f2, "Test"); bp.apply_async(f3, "Test"); bp.execute();//!< WAINTING function // Or maybe a function who provide a result list for each function // Like this: std::array<int> functionsReturns = bp.get_result<int>();//!< WAITING function // And another great things from https://svn.boost.org/svn/boost/sandbox/async/libs/async/doc/html/toward_bo… // Multi-function without parameter // >> 1 CALL for f4 // >> 1 CALL for f5 // >> f4 // f5 // -> // = paralell // NOTE: same as: // bp.apply_async(f4); // bp.apply_async(f5); // bp.execute();//!< WAITING function bp.call_all(f4, f5);//!< WAITING function // Maybe a function who provide a result list for each function // Like this: std::array<int> functionsReturns = bp.get_result<int>(); return 0; } > The link doesn't worksOups. Sorry for the typo. http://immortal-pc.info/Boost.html Do you keep my proposal for Boost.SQL? Best,Guillaume

2 1

GSoC: Student applications can now be submitted
by Boris Schaeling 23 Apr '13

23 Apr '13

Student applications can now be submitted to <https://google-melange.appspot.com/gsoc/homepage/google/gsoc2013>. The deadline is 3 May. Boris

3 3

[iterator] Code/documentation mismatch in iterator_facade
by Gabriel Redner 23 Apr '13

23 Apr '13

Hi folks, I've stumbled upon an issue with iterator_facade. When creating a random access facade, one must implement the distance_to member function in order for std::distance to work. However, it looks like the arguments passed to this method are reversed from what the documentation states. Specifically, the docs [1] say (paraphrased): a.distance_to(b) is equivalent to distance(a, b). However, when I call std::distance(a, b), I find that b.distance_to(a) is invoked in the facade - the mirror image of what the docs say! This is not a problem if just wrapping some other random access iterator, but in my case the calculation of distance is nontrivial and this issue led to a rather tiresome debugging session and the solution (negating before returning from distance_to, knowing that iterator_facade will re-negate my response) is annoying to explain. Should I file a bug? A test case is appended to this message. -Gabe [1] http://www.boost.org/doc/libs/1_53_0/libs/iterator/doc/iterator_facade.html… ==================== #include <boost/iterator/iterator_facade.hpp> #include <vector> #include <iterator> #include <iostream> class vector_iterator : public boost::iterator_facade< vector_iterator, int, boost::random_access_traversal_tag> { public: vector_iterator(char tag, std::vector<int>::iterator it) : _tag(tag), _it(it) {} private: friend class boost::iterator_core_access; int& dereference() const { return *_it; } bool equal(vector_iterator const& other) const { return _it == other._it; } void increment() { ++_it; } void decrement() { --_it; } void advance(int n) { std::advance(_it, n); } std::vector<int>::iterator::difference_type distance_to(const vector_iterator& other) const { std::cerr << _tag << ".distance_to(" << other._tag << ")\n"; std::vector<int>::iterator::difference_type result = std::distance(_it, other._it); std::cerr << "result = " << result << '\n'; return result; } private: const char _tag; std::vector<int>::iterator _it; }; int main() { std::vector<int> v; v.push_back(1); v.push_back(2); vector_iterator a('a', v.begin()); vector_iterator b('b', v.end()); std::cerr << std::distance(a, b) << '\n'; } ====================

2 2

[REVIEW] Boost.Predef
by Joel Falcou 23 Apr '13

23 Apr '13

Hi all, first, I want to apologize for the delay in getting this review results out. Looks like my last email get lost at some point. Anyway, after the initial review of Rene Riviera Boost.Predef, my decision was to Conditionnally Accept Boost.Predef under the request that the comments raised during the review period were addressed. Since then, Rene worked hard on actually making those changes so I now feel confident to announce that Boost.Predef is ACCEPTED as a Boost library. Regards

1 0

GSOC 2013
by Computer Vision 23 Apr '13

23 Apr '13

Hello, I was very busy with my studies, am I late? Are there any free projects left? Kind Regards

3 2

Going forward with Boost.SIMD
by Niall Douglas 23 Apr '13

23 Apr '13

> Unfortunately the concurrency/parallelism group has decided that they do > not want C++ to provide types representing SIMD registers. > I'm afraid I don't quite understand the rationale for such a refusal; > proposing more high-level constructs similar to valarray (or to our own > library NT2) was suggested, but that's obviously a more complex and > limited API, not a basic building block to program portably a specific > processor unit. As much as a refusal can be heartbreaking, trust me when I say that indecision or indifference is worse. The fact they gave you a definite refusal is actually one of the least worst outcomes from an ISO standards proposal, because you can now move on without uncertainty. > Development of Boost.SIMD will still proceed, aiming for integration in > Boost, but standardization appears to be definitely out of the question. > Any feedback of the API presented in the proposal is welcome. > <http://open-std.org/JTC1/SC22/WG21/docs/papers/2013/n3571.pdf> If I were still part of ISO standards, I'd observe the following: 1. GPU and CPU stream computation technologies are still merging. In other words, it's too soon to standardize this technology lest we accidentally break some novel form of new convergence. Happy to reconsider post-C++14. 2. It's hard to standardize current CPU SIMD implementations due to extremely irritating inconsistencies between vendors. For example, a generic straight port of SSE2 to NEON will have awful performance because SSE2 code does a lot of flipping between SIMD and non-SIMD, and because NEON is a coprocessor on ARM that generates poor performance. Another bug bear of mine on NEON is the lack of an equivalent to _mm_movemask_epi8(), which can be emulated in about eight NEON instructions, but in so doing you'll make code which was very speedy on SSE2 pretty slow in most cases on NEON. My point here is you cannot standardize such non-uniform behavior in a universally performant way, because you'll just get lowest common denominator performance across all SIMD implementations which kinda defeats its purpose. If all vendors were like NEON, or like SSE2, then we still have to see how CUDA and OpenCL pan out long run. 3. I am unsure if C++ is the appropriate language for SIMD standardization when perhaps a meta-form of JIT compiled C++ would be much superior (i.e. you supply LLVM bytecode, and it gets delivered to a GPU/CPU/whatever). We'll have those on the table with LLVM-type compilers. In other words, I would vote to wait and see what the market throws up. I appreciate that none of these three rationale are what you want to hear. Still, I hope my observations are useful to you. None of them suggests you shouldn't proceed with Boost.SIMD. Boost has a much wider remit than just as a testing ground for future C++ standard library features. But I suspect that if SIMD ever does get standardized, it won't look like your library or proposal because it will be based on technologies which don't exist yet. Hope that helps, Niall

3 2

[log] Patch: rotating file collector (Was: Boost.Log: Any Comments?)
by Alexander Arhipenko 23 Apr '13

23 Apr '13

Hi, Just to remind the context, I'm pasting content from previous mail thread. On Tuesday 09 April 2013 18:43:05 Alexander Arhipenko wrote: >> ... >> [snip] >> ... >> >> It would be very nice to have rotation feature work the same way as in e.g. >> python logging library or Apache web server: >> * latest file has the name filename.log >> * previous file has the name filename.log.1 or filename.log.YYYY-MM-DD >> AFAIK, it's impossible to achieve in current implementation. >> >> Andrey, what do you think about this? >The library does not provide this behavior out of the box. But the file sink >uses a file collector object to collect rotated files. This collector can be >implemented by the user (it should derive from sinks::file::collector >interface) and it should be possible to implement it the way you describe. >Would you like to submit such implementation? I've added implementation of rotating file collector - the patch against boost log trunk attached. This is not the final version, I'm going at least to finish most of the TODOs. However, it would be nice to receive feedback and questions from library author as well as from people interested in this feature. Andrey, could you please find several minutes to review the attached patch? Also, I would like to discuss following question with you : rotate_file invocation in ~text_file_backend - I would prefer to have this call during text_file_backend construction phase. Regards

1 0

std::optional<T>
by Fernando Cacciola 23 Apr '13

23 Apr '13

Hi People, I'm pleased to announce that thanks to the outstanding effort of Andrzej Krzemienski who wrote the standard papers, and Ville Voutilainen who champion it during the meetings, we now have std::optional<T> in C++14! Cheers -- Fernando Cacciola www.scisoft-consulting.com

11 15

Plugin/Extension : DLL/so loading?
by Mathias Gaunard 23 Apr '13

23 Apr '13

I remember there were a couple of projects (Boost.Plugin or Boost.Extension) to support dynamic loading of a DLL/so plugin in a portable fashion in Boost. What has happened to these, and is there some code I can use somewhere?

3 4

Boost Graph Library: why are `source()`/`target()` non-member functions?
by Max Moroz 22 Apr '13

22 Apr '13

Why is `source()` not a member function of the graph class? (See http://www.boost.org/doc/libs/1_53_0/libs/graph/doc/index.html for description, and boost/graph/adjacency_list.hpp and boost/graph/adjacency_matrix.hpp for implementation.) I thought the reason was that it was a generic algorithm. But it's not. There's no interface (beyond `source()` itself) exposed by an edge descriptor that would allow `source()` to avoid digging into the implementation. If I interpret the code correctly, `source()` reads directly from the internal data structure of the edge descriptor (`detail::edge_base.m_source`), which I am pretty sure is the implementation detail (`m_source`, despite being strangely exposed as a public member, is not the standard interface; otherwise it would have been defined as such, and `source()` would have been entirely redundant.) I think I'm missing something, but what?

2 3

GSOC
by Flórián Deé 22 Apr '13

22 Apr '13

Personal Details: Name: Flórián Deé University: Budapest University of Technology and Economics Major: Engineering Information Technology Degree Program: B.Sc. Email: deeflorian(a)gmail.com Availability: -starting after the exams, best case: June 10, worst case: June 21 -ending on at the 2nd week at the university: Sept. 15 -planned time spent on GSoC: 32 to 56 hours per week, as needed -other factors: two 3-day long trips - would catch up afterwards with extra work Background information: During my B.Sc. studies I'll already have finished all courses that are related to calculus, linear algebra, number theory, combinatorics and digital technology until summer, so in my opinion, all the educational background needed is present with the exception of the Signals and Systems course in next semester, which would cover digital Fourier transforms in detail. I'd gladly make up for this by reading the needed material. As for my programming background, I've fallen in love with C and C++, which now I consider to be my main languages. I'm currently working on a project as an outsider member of the dorm's Coder team (a poker framework for AIs), and I'm also a "consultant" for students with their C/C++ assignments. As a result, I've seen, written and debugged lots of different (admittedly not so enormous) projects with different coding styles. I'm interested in contributing to the Boost Libraries because two of my friends used a quite a lot of them while researching, and their feedback was always positive. The C++11 implementations also sounded nice, so adding to this codebase as a summer project would be an honor. I wanted to do a Solar System simulator during secondary school because I thought that I could calculate the acceleration (and approximate path) of a pen somehow shot near Earth. Then after the first console-based C++ implementation I realized that doubles aren't sufficiently precise to determine if an object is near or inside a planet, so I tried to write a "double double" class - which was too childish to work both code and mathematicalwise. Since then I know that I'd needed a precise mathematical library, and that radix-2 is my best friend when calculating on CPU-s, so I think that this project would be an ideal fit. Rating of knowledge: C++ - 4 C++ Standard Template Library - 4 Boost C++ Libraries - 2 Subversion - 3 Development environments: Eclipse, Visual Studio Documentation tools: Doxygen Project proposal, schedule and milestones: I'd like to work on the Boost.Multiprecision project described on the wiki page. As the description there is more than detailed, I wouldn't like to copy all that text here, but I'd follow the main steps described. I think that the I/O and the +, - operators should be finished, and the multiplication already started by the end of July. The *, / operators should be finished by the end of August at least with the Karatsuba multiplication, interoperable with Boost.Math, and the FFT multiplication should also be started by then. The final goal would be FFT multiplication, with passed correctness tests of wide variety from the beginning of the project.

1 0

shared_ptr ordering
by Matt Bergin 22 Apr '13

22 Apr '13

Back in 2008, changing shared_ptr operator< to compare by pointer value rather than owner was discussed: http://lists.boost.org/Archives/boost/2008/10/143175.php There seem to be convincing arguments for both implementations (aliasing, consistency with equality operator etc). Is boost::shared_ptr intended to be a compliant C++11 std::shared_ptr? If so, will shared_ptr be changed to comply with [util.smartptr.shared.cmp] or is the ensuing breakage too high a price to pay for 100% compliance? Thanks Matt

2 1

[ublas] GSoC 2013 Proposal - uBLAS Math Library and Performance Optimization
by Sathyam Vellal 22 Apr '13

22 Apr '13

Hello everyone. I've been in touch with Mr. David Bellot and Mr. Nasos Iliopoulos and with their guidance and support I've written this proposal for Boost.uBLAS. The proposal can be viewed here -http://goo.gl/Qo7Ry. (This is a link to a google doc; a pdf version of the proposal). Thank you -- Sathyam M Vellal P.E.S. Institute of Technology, Bangalore, india

1 0

[config][mpl][type_traits][integer] BOOST_STATIC_CONSTANT should use constexpr
by Vicente J. Botet Escriba 22 Apr '13

22 Apr '13

Hi, BOOST_STATIC_CONSTANT could be improved to make use of constexpr. This is needed for instance to make integral_constant compliant with C++11. The patch consists in changing const by BOOST_CONSTEXPR_OR_CONST in boost/config/suffix.hpp // BOOST_STATIC_CONSTANT workaround --------------------------------------- // // On compilers which don't allow in-class initialization of static integral // constant members, we must use enums as a workaround if we want the constants // to be available at compile-time. This macro gives us a convenient way to // declare such constants. # ifdef BOOST_NO_INCLASS_MEMBER_INITIALIZATION # define BOOST_STATIC_CONSTANT(type, assignment) enum { assignment } # else # define BOOST_STATIC_CONSTANT(type, assignment) static *BOOST_CONSTEXPR_OR_CONST* type assignment # endif and in each out of line static constant definition as in template <int b1, int b2> bool *BOOST_CONSTEXPR_OR_CONST* ice_eq<b1,b2>::value; The main problem is that this is a *breaking change*: the list of concerned libraries is quite big (mpl, type_traits/integer_traits, ...) and users of BOOST_STATIC_CONSTANT would need to make this modification also. An alternative is to define a BOOST_STATIC_CONSTANT_11 that includes this modification and update the Boost libraries that could make use of the new macro without introducing a breaking change. Library providing a template class containing a field with such a change that can be specialized by the user would need to either announce a breaking change or just don't use the new macro. While this mean more changes the backward compatibility can be mastered. What do you think, it is worth try to fixing this? Best, Vicente

4 17

GSOC - Boost.SQL and Boost.Thread interest
by Guillaume Nuel 22 Apr '13

22 Apr '13

Hi Boost C++ Team ! My name is Guillaume NUEL. I'm a French student in computer security (fourth year in computer science). I'm actually making a game witch uses Boost (Thread, Signal,...). I also use CouchBase for database. I really would like mixing my game development and Boost development. My project proposal Idea : I'm mainly interesting in adding a new module: Boost.SQL for managing database: MySQL, Couchbase, … If the project proposal is not suitable, I have others interests. I would also be interesting in adding new functions to Thread abilities. So I am interested in the following topics : - Boost.Thread/SyncQueue: Provide boost::sync_bounded/unbounded_queue implementations base on the C++ Concurrent Queues proposal. Trying an adaptation of the Boost.LockFree queues to follow the same interface. - Boost.Thread/ThreadPool:Provide a boost::thread_pool class scheduling arbitrary functions based on Toward Boost Async 0.2 and extend the boost::async function to manage with this new class. Projects mentored by Vicente J. Botet Escriba I'm an expert in C++ and I have good skills in OpenMP and Cuda ( GPU ) and so in parallelization of code. Homepage: http://immmortal-pc.info/Boost.html My CV is available at http://immortal-pc.info/cv.html Thank you for taking the time to read my e-mail, I look forward to receive a response from you. With regards, NUEL Guillaume

2 1

GSoC: More mentors?
by Boris Schaeling 22 Apr '13

22 Apr '13

From Monday on students can apply for GSoC projects. Mentors can already apply at <https://google-melange.appspot.com/gsoc/homepage/google/gsoc2013>. So far we have four possible mentors (for uBLAS, odeint, graph and benchmark). Are there others who would like to mentor? Then please pick a project from <https://svn.boost.org/trac/boost/wiki/SoC2013> or propose your own (like I did a few days ago: <http://article.gmane.org/gmane.comp.lib.boost.devel/240251>). Applying as a mentor doesn't mean that you must mentor a student. Among others it all depends on if and how many student applications we will receive. But having a few more mentors just in case would be nice. :) Thank you, Boris

4 6

media > audio project (gsoc or future boost)
by adrien courdavault 22 Apr '13

22 Apr '13

Hi everyone. I read the project list for the GSOC 2013 and in Ideas I've seen the mention of Audio. I would like to know who is interested in this and if someone could give some details about this line : 'Audio library (OpenAL/FMOD/etc)' see here :https://svn.boost.org/trac/boost/wiki/SoC2013#Media I would like to know if this might be related to this: http://comments.gmane.org/gmane.comp.lang.c%2B%2B.isocpp.proposals/310 Thank you in advance. Adrien

6 11

Writing timer output to file
by Mohamed Nemri 22 Apr '13

22 Apr '13

Hi, I want customize boost::timer::auto_cpu_timer to put results in a logging stream(file, console,...). What is the best way to do that? Thanks --

1 0