- Boost - lists.boost.org

GSOC Proposal for BIGINT - Long Mail
by swagat konchada 25 Mar '10

25 Mar '10

I had a look around in the Boost Sandbox and I found a BigInt implementation done by zeux during GSoC 07. I had a look at the code and I wanted to know its deficiencies. I'm interested in proposing some improvements in bigint and new bigdecimal implementation that is building on this code. There are 2 questions that I wanted to ask the list before moving further on this track: 1. Is there sufficient interest for a BigDecimal proposal on the lines of the Java implementation of BigDecimal (or has there been some prior work that can be used as a foundation?) 2. What are the deficiencies of bigint that have prevented its inclusion into the Boost distribution so as to address the same as part of my GSoC proposal. I did some homework and contrasted the bigint api in Boost and Java and found some interesting things that could be done. Boost.bigint was written by zeux and he used GMP bigint's features writing the Boost.bigint code around it. I think I could learn and analyse more after having a look at GMP bigint. After all, I think Boost.bigint should be more exhaustive. -------------------------------------------------------------------------------------------------------------------------------------------------------------- Feature Boost bigint Java BigInt -------------------------------------------------------------------------------------------------------------------------------------------------------------- Constructors from | int | byte[] #byte array with 2's complement representation of the BigInt. | unsigned int | int bitlen, int certainity #random BigInt with given bit length | int64_t | | uint64_t | | char *, int base | String | string& str, int base | String, radix -------------------------------------------------------------------------------------------------------------------------------------------------------------- *Conversion from string of digits to bigint is present. However conversion from/to byte-array(2's complement notation) to/from bigint isn't. Also the generation of a random arbitrary precision bigint should be done. int bitCount() #return the no of bits in 2's complement of the bigint; int bitLength() #no of bits excluding the sign bit double doubleValue() #casting bigint to double must be done. #even casts to int, uint and int64_t, char * must be done.* -------------------------------------------------------------------------------------------------------------------------------------------------------------- Arithmetic Operators += In Java, add, subtract, multiply, etc., | -= | *= | /= | %= | |= | &= | ^= #xor | <<= | >>= | ++ #postfix & prefix | -- #postfix & prefix -------------------------------------------------------------------------------------------------------------------------------------------------------------- Unary Operators + - BigInteger[] negate() ~ BigInteger[] not() ! #is zero < int compareTo(BigInteger val) <= BigInteger max,min(BigInteger val) > >= == != -------------------------------------------------------------------------------------------------------------------------------------------------------------- *These are some bit manipulators BigInteger clearBit(int n) #not in Boost.bigint #even setBit & toggleBit could be done, with a feature of changing the precision of the bigint. BigInteger[] testBit(int n) #also bit testers l* -------------------------------------------------------------------------------------------------------------------------------------------------------------- Binary Operators + BigInteger add(BigInteger val) - * BigInteger multiply(BigInteger val) {bigint lhs, int rhs} {int lhs, bigint rhs} / BigInteger divide(BigInteger val) % BigInteger mod(BigInteger val) , BigInteger modPower(BigInteger val) | BigInteger or(BigInteger val) & BigInteger and(BigInteger val) ^ BigInteger andNot(BigInteger val) << BigInteger shiftLeft(int n) >> BigInteger shiftRight(int n) -------------------------------------------------------------------------------------------------------------------------------------------------------------- Math function overloads abs(bigint_base) BigInteger abs(int exponent) pow(bigint_base, uint64_t) BigInteger pow(int exponent) div(bigint_base lhs, bigint_base rhs, bigint_base& remainder) BigInteger[] divideAndRemainder(BigInteger val) sqrt(bigint_base) ----------No Java version for BigInterger------------ -------------------------------------------------------------------------------------------------------------------------------------------------------------- *BigInteger gcd(BigInteger val) #GCD would be one of the most important requirements for bigint #Also the Java function of bool ProbablyPrime() is very interesting and would be challenging to implement.* -------------------------------------------------------------------------------------------------------------------------------------------------------------- -- Problem is a phenomenon indicating lack of thought. Swagat Konchada

3 2

Debugging Visualizers
by Alex Miller 25 Mar '10

25 Mar '10

quoted from the ongoing "5 Observations" thread: On Tue, Mar 23, 2010 at 4:36 PM, Steven Watanabe <watanabesj(a)gmail.com>wrote: > > Perhaps you would find this useful: > https://svn.boost.org/trac/boost/wiki/DebuggerVisualizers > > It is mentioned there that the same thing could be done for GDB,etc. However, I can't seem to find anything of the sort by googling (although it's very possible I'm just searching poorly). Could anyone point me towards where docs on how to do this could be found at?

2 2

filesystem::create_directories error
by Zachary Turner 24 Mar '10

24 Mar '10

I'm trying to use boost::filesystem::create_directories() with a path of the form \\?\Volume{Guid}\foo on windows. This seems to break the filesystem library because it doesn't understand the path syntax. when recursing up the parents it removes the slash at each level, so it ends up calling boost::filesystem::exists(L"\\?\Volume{Guid}") which always returns false, I think because GetFileAttributes() requires that paths in this syntax end with a trailing \, whereas the trailing \ has been removed. I'm tempted to fix this, but I'm sure this algorithm is fragile and any change I apply is likely to break another case. Does anyone have any suggestions on how to fix this? Thanks

3 4

[log] review part 1
by Steven Watanabe 24 Mar '10

24 Mar '10

AMDG I mostly have documentation notes at this point. Andrey, since I have so many, and they're mostly fairly minor, would it be better if I made a patch for the documentation source? I think newbies should probably be replace globally with new users, since I kind of feel that the term newbie is somewhat derogatory. http://boost-log.sourceforge.net/libs/log/doc/html/log/moti.html: ================================================================================ Mismatched number: Most of the time ***applications*** ... to monitor ***its*** execution... The beginning of the second paragraph could be improved: "This is where logging may help. The application stores all essential information of its run time to log, and once something goes wrong this information can be used to analyze program behavior and make necessary corrections." Here's my suggestion: "This is where logging can help. The application stores all the essential information about its execution to a log, and when something goes wrong this information can be used to analyze the program's behavior and make the necessary corrections." "out-of-box" should be "out-of-the-box" This phrase sounds rather clunky to me: "...along with public interfaces ready to be used to extend the library." This too: "A user should be able to extend functionality of the library with regard to collecting and storing information into logs." "The library should make as least performance impact..." should be "The library should have as little performance impact..." http://boost-log.sourceforge.net/libs/log/doc/html/log/how_to_read.html: ================================================================================ In the beginning of the second paragraph, I don't like the phrase "to have the first glance on" How about something like "...to get a feel for the library's capabilities..." "The tutorial gives the overview of the library features..." Suggestion: "The tutorial gives an overview of the library's features..." "The simple form typically describes the most common and easy way to do the task and it is being recommended to be read by newbies." Suggestion: "The simple form typically describes the most common and easy way to do the task and is recommended for new users." "The advanced form usually gives an expanded way to do the same thing but with ability to do some extra customization and in-depth explanation." I would like a definite article before "ability" and I would also like "and in-depth explanation" to be placed differently so that it doesn't sound like you're saying that the user has the ability to do in-depth explanation. "This form may come handy for more experienced users and should generally be read if your needs are not satisfied with the easy way." Suggestion: "This form may come in handy for more experienced users and should generally be read if the easy way does not satisfy your needs." "This part gives description of other tools..." should be "This part gives descriptions of other tools..." "This chapter is better to be read on per-case basis." Suggestion: "This chapter is best read on a case by case basis." "there is a reference which gives the formal description of library components. " The work "description" should be plural, "descriptions". "For simplicity in the code snippets in this documentation it shall be assumed that the following namespace aliases were defined" Suggestion: "To keep the code snippets in this documentation simple, the following namespace aliases are assumed to be defined" "Aside from these namespaces the library also provides boost::log::experimental namespace." Suggestion: "In addition to these namespaces the library also provides the boost::log::experimental namespace." "and thus are not advertized for productional usage" The word "productional" should be "production" http://boost-log.sourceforge.net/libs/log/doc/html/log/supported_compilers.…: ================================================================================ "The library should build and work with a reasonably well-compliant compiler" delete "well-" http://boost-log.sourceforge.net/libs/log/doc/html/log/installation.html ================================================================================ "The library has the compiled part which should be built as described in the Getting Started article." Suggestion: "The library has a separately compiled part which should be built as described in the Getting Started guide." For BOOST_LOG_NO_THREADS "Has effect on both the library compilation and user's code compilation." Suggestion: "Affects compilation of both the library and users' code." "The macro is automatically defined if no threading support detected." add "is" before "detected" For: BOOST_LOG_USE_CHAR Same as the first comment for BOOST_LOG_NO_THREADS. (This appear many times, I won't repeat it again) For: BOOST_LOG_USE_WCHAR_T "it is assumed the both character types support is enabled" Suggestion: "support for both character types will be enabled" "Defining only one of them allows to significantly reduce binary size of the library." Suggestion: "Defining only one of them significantly reduces the binary size of the library." For BOOST_LOG_NO_QUERY_PERFORMANCE_COUNTER: "If defined, disables support for QueryPerformanceCounter API" Suggestion: "If defined, disables support for the QueryPerformanceCounter API" For BOOST_LOG_USE_NATIVE_SYSLOG: "Has effect on the library compilation stage." Suggestion: "Affects only the compilation of the library." (repeated later as well) "If for some reason the support for native SysLog API is not detected, define this macro to forcibly enable support for native SysLog." Suggestion: "If for some reason support for the native SysLog API is not detected automatically, define this macro to forcibly enable it." For BOOST_LOG_NO_SETTINGS_PARSERS_SUPPORT: "If defined, all facilities related to settings parsing are not built." Suggestion: "If defined, none of the facilities related to the parsers for settings will be built." "You may define configuration macros..." should be "You can define configuration macros..." http://boost-log.sourceforge.net/libs/log/doc/html/log/defs.html ================================================================================ For Log record: I don't like the expression "A single pack of information," although I can't think of anything better off hand. "that is candidate to be put into log" Suggestion: "that is a candidate to be put into the log" For Attribute "functional objects" should be "function objects" "which return the actual attribute value upon invoking" Suggestion: "which return the actual attribute value when invoked" For Log sink: "collected from user's application" should be "collected from the user's application" "It is sink's nature that defines where..." Suggestion: "The sink defines where..." For Log source "An entry point for user's application..." should be "An entry point for the user's application..." "upon user's request" should be "upon the user's request" For Log formatter "functional object" should be "function object" "...that forms up the final shape of the textual output" is awkward. "Some sinks, like binary logging sink" should be "Some sinks, like the binary logging sink" For Logging core "The global entity that maintains connection between" should be "The global entity that maintains the connection between" "It is mainly used on the logging library initialization stage." replace "on" with "in." http://boost-log.sourceforge.net/libs/log/doc/html/log/design.html ================================================================================ I htink "scheme" with "figure" "The library provides configuration facilities to leave only one needed version of the library to be compiled." Suggestion: "The library provides configuration facilities to compile only the version of the library needed." "For example, if your application is in critical state" should be "For example, if your application is in a critical state" "so that user sees an error message..." should be "so that the user sees an error message..." "This allows to use the library" Suggestion: "This allows use of the library" "...require users to use loggers to write logs" "use" and "log" are used too many times here. "The more generic term "log source" is used to designate the entity" Suggestion: "The more generic term "log source" designates the entity" "You may see on the scheme that the" suggestion: "You can see on the scheme that the" "and allows to quickly wipe away unneeded log records" suggestion: "and allows quickly wiping away unneeded log records" anyway, don't split infinitives. "The sink-specific filtering allows to direct log records to particular sinks." Suggestion: "The sink-specific filtering allows log records to be directed to particular sinks." "Backends, on the opposite" suggestion: "Backends, on the other hand" "most probable places of the library extension" suggestion "most likely places for extending the library" "Along with the described above primary facilities" should be "Along with the primary facilities described above" ================================================================================ Issues: Instead of compiling a narrow or wide library, why don't you use libboost_log and libboost_wlog? I agree with other comments that I saw about Attribute being a poor choice for what it does. I would prefer that the term "Attribute" should be used for what you call an "Attribute Value," and that there should be a different term, like "Attribute Extractor" for what you call an "Attribute." In Christ, Steven Watanabe

3 10

[ASIO] Create SSL stream from native socket?
by Marcel Fransen 24 Mar '10

24 Mar '10

Hi, I would like to create an ssl stream for an already accepted native socket, just like is possible with basic_stream_socket< tcp >. How should this be done? Currently there is only one ssl::stream constructor which takes one parameter that is passed to the basic_stream_socket: template <typename Arg, typename Context_Service> explicit stream(Arg& arg, basic_context<Context_Service>& context) : next_layer_(arg), service_(boost::asio::use_service<Service>(next_layer_.get_io_service())), impl_(service_.null()) { service_.create(impl_, next_layer_, context); } To be able to create an ssl stream from an existing socket we would need a constructor which passes three parameters to the basic_stream_socket like this: template <typename Arg1, typename Arg2, typename Arg3, typename Context_Service> explicit stream(Arg1& arg1, Arg2& arg2, Arg3& arg3, basic_context<Context_Service>& context) : next_layer_(arg1, arg2, arg3), service_(boost::asio::use_service<Service>(next_layer_.get_io_service())), impl_(service_.null()) { service_.create(impl_, next_layer_, context); } Would it be a problem to add constructors that passes two and three parameters to the basic_stream_socket or can this be done in another way? Kind regards, M. Fransen _________________________________________________________________ New Windows 7: Simplify what you do everyday. Find the right PC for you. http://windows.microsoft.com/shop

2 1

Re: [boost] GSOC 2010 Boost.Graph
by Chinmaya Poswalia 24 Mar '10

24 Mar '10

Topology Generators : 1) along with path, cycle, star how about also adding bus ,mesh and hybrid topologies. 2) we can also include dual, complement , line and power of a graph.

1 0

Re: [boost] GSOC 2010 Boost.Graph
by Chinmaya Poswalia 24 Mar '10

24 Mar '10

Bindings to Python : there are tools as SWIG as language bindings. is it good employ them in making the bindings. Could anyone please suggest some tutorial for binding creation. I was not not able to find on google

1 0

Willing to Join GSOC 2010
by Xin Shuai 24 Mar '10

24 Mar '10

Hi: I'm interested in your Boost.Graph project. I think my current graph theory related research as well as C++ and python programming experiences makes me fit for the project. I'm wondering who's the mentor for the Boost.Graph project and I hope he/she can give me useful suggestions about the project. -- Xin(David) Shuai PhD of Complex System in School of Informatics & Computing Indiana University Bloomington 812-606-8019 The way to success is to do as much as important things, and as less as unimportant things, as you can...

2 1

[GSoC]I am interested in Boost.Graph
by doraemonok 24 Mar '10

24 Mar '10

My name is Rui Diao. I am a student of Graduate University of Chinese Academy of Sciences. My major is Computational Mathematics. I have never contributed to any open software project, but only spread the open source spirit to the people around me and join some open source groups. The only contribution of my source file is developing a Client of IP Gateway of My University by Qt open source version. Contributing to Boost C++ Libraries is an interesting and challenging job, and also, of course, fantastic. I am interested in one project of Boost.Graph : Graph Connectives. I like Graph Theory very much. I have experiences on coding (and teaching the friends in open software association of my university) some algorithms about Graph Theory: Dijkstra, Floyd, Bellman Ford, Ford Fulkerson, SAP, Kuhn-Munkres, etc. I have learned a lot in "How to finish an excellent project" during the internship of Microsoft Research Asia. Moreover, I am familiar with Mathematica which has similar function with Boost.Graph. I am very familiar with STL, but seldom use Boost. I just have experiences on using Boost.Regex before. I have checked out the source code by Subversion and began to read the part of Boost.Graph recently. The Boost.Graph library is missing connectives which contains Graph Join, Graph Union, Graph Product, Graph Intersection, Graph Difference, Graph Sum, Graph Product, Line Graph, etc. I think the algorithms may be not too difficult to me, but how to code with "Boost Style" (of course not only the code style) is a problem. I will try my best to finish the challenging work. I think may be there is some mentor can give me some advices or suggestions. Thank you all. Rui Diao

2 1

[log] Boost.Log formal review
by barend 24 Mar '10

24 Mar '10

Hi, This is my review of Boost.Log submitted by Andrey Semashev. > Please explicitly state in your review whether > the library should be accepted. I've been doubting about this for a few days. When I started my review, by trying to use the library, I discovered that the library is not header-only. I had thought that a logging library should be lightweight and had therefore expected only headerfiles. I then tried the library from John Torjo from 2007 but it is not header- only neither, so no reason to favour that one (I didn't dive into it further). Because of some messages on the list, a.o.: > I think that any reasonable log library must be compiled > to it's own lib due to the intermodule singleton requirement I realized that it is indeed useful to have a non-header-only option as well. The answers of Andrey were satisfactory, a.o.: > If we have a list of places in Boost where logging may be applied, > perhaps we could compile a set of requirements for such a lightweight > wrapper. When we have it, it would be easier to tell, whether Boost.Log > fits the role or not, or whether the wrapper should be part of it or a > separate submission. I'd be interested to hear from the library authors > on this topic. And I found in the Boost.Log documentation, TODO-list: > Think over a header-only configuration. Perhaps, with a reduced > functionality. So, good. Then I started the real review and I started to like the library, it is definitely useful, is configurable, trivial logging is quite simple. It also contains very useful non-trivial features such as writing to the Windows Event Log and alarm critical states as balloon tips. So I finally decided to give my YES vote, contingent on the condition that a lightweight header-only configuration, e.g. single-threaded, single-module, will be available for library writers. Some more about this below. --------------------------------------------------------------------- > What is your evaluation of the design? I like the options, the trivial logging, multiple back-ends, modularity and extensibility. I didn't have a detailed look to the design but in general it looks good to me. > What is your evaluation of the implementation? I didn't have a detailed look either. I glanced through the sources and the implementation is looking neat. I wonder why the class "basic_slim_string" is necessary, to me it sounds very strange why a task as logging needs an own string implementation. However, I didn't have the time to dive into it and didn't ask questions on this on the list. > What is your evaluation of the documentation? The documentation is good. Pages like design overview, tutorial, installation are easy to follow. The tutorial pages however contain some omissions and you have to refer to the examples to see what is missing. I mean here the chapter "Trivial logging with filters" which seems to give a complete example, but misses the namespace aliases, two of them would have been enough. > What is your evaluation of the potential usefulness of the library? Very useful, definitely, for applications and for libraries. > Did you try to use the library? With what compiler? Did you have any > problems? Yes, I used it using Visual C++ 2008 Express Edition. Because Boost.Log is not header-only, and I don't have the Boost libs compiled on my machine (don't ask me why, I never have them), I had several issues which other reviewers probably didn't run into. The defines _CRT_SECURE_NO_WARNINGS and _SCL_SECURE_NO_WARNINGS were necessary to surpress M$ warnings. The define BOOST_ALL_NO_LIB is of course necessary to build it like I did, and it works for Boost.Log. The compiler complained about a missing simple_event_log.h by event_log_backend.cpp, so I drafted it by hand, but I later saw that it is documented and you should neatly use the message compiler, so OK. After adding some Boost source files from thread, filesystem, regex and system it worked, datetime was not necessary. A complete build (including all mentioned sources) costs about 2 minutes. The tutorial did work (see message above on filtering, adding namespace aliasses was necessary) and I experimented with some things, many options seems very useful to me such as the log format, automatic date time, log file rotation, log file max size, etc. etc. I like the streaming feature (so for me printf is not necessary). I have some reservations about the compiler support: GCC 3.4.5, the default on Linux and MinGW, is not supported and for a task as logging this seems to me as a disadvantage. > How much effort did you put into your evaluation? A glance? A quick > reading? In-depth study? About five to six hours spread over some days. > Are you knowledgeable about the problem domain? Yes, I always need logging. In our library at Geodan we built logging as a singleton in 1995. We rewrote it to a logging-over-DLL's in 2000, but these two were not 10% as sophisticated and complete as this logging library is. So yes, these things were not header-only neither. --------------------------------------------------------------------- So back on the logging for libraries. If Boost libraries which are header-only will start to use this library, without protection, thousands and thousands of project files and makefiles worldwide will be broken. This answer: > If an existing header-only library starts using the Boost.Log library > we'll need to update the makefiles. This is a reasonable cost (not from Andrey)is therefore not satisfactory to me. This answer: > Which probably means that the decision is with the author(s) of each > individual library. (from the review manager) surprised me a bit, I think that in a library under review such an important thing may certainly be discussed, per library. However, as said, it is not a reason to vote no, because the rest of the library is quite good. A small macro (sorry) wrapper would already do most of the trick, and avoid breaking project files: #if defined(BOOST_ALL_LOG) || defined(BOOST_GEOMETRY_LOG) #define BOOST_LOG_GEOMETRY(o, s) BOOST_LOG_TRIVIAL(o) << s #else #define BOOST_LOG_GEOMETRY(o, s) #endif Where it can be called in libraries as: BOOST_LOG_GEOMETRY(info, "Log line with a value " << 3); This is inspired on the general define "BOOST_ALL_NO_LIB" and library- specific define "BOOST_LOG_NO_LIB". This small wrapper would really be useful, at least for us, and replace the #ifdef BOOST_GEOMETRY_DEBUG lines which are currently in our Boost.Geometry library. This is zero cost when one of the macro's BOOST_ALL_LOG or BOOST_GEOMETRY_LOG is not defined (by default it is not), and contains much functionality of Boost.Log as filtering, rotation, etc, when one of these macros is defined. Of course it can be enhanced or there are better solutions. Anyway, a lightweight solution will not be difficult. --------------------------------------------------------------------- So thanks Andrey for submitting this library, thanks Volodya for managing the review, and I hope it will be accepted with a lightweight header-only configuration included. Regards, Barend

2 1