March 2010 - Boost - lists.boost.org

[log] Boost.Log formal review result
by Vladimir Prus 25 Mar '10

25 Mar '10

First of all, I would like to thank Andrey for this submission and the active discussion during the review. I also would like to thank everybody who participated in the review. The comments on the library were generally positive. Several reviewers have tried it in practical projects already, and found it superior to other libraries. It clearly will benefit the users of Boost right away. Therefore, the library is accepted. Congratulations, Andrey! On the other hand, some concerns about the ease of use and performance and implementation details were raised. Of those, some can be easily addressed as part of normal maintenance, while some seem to require significant rework and may change the library quite consirably. Therefore, the library is accepted subject to several conditions -- listed below under "Critical issues". Because the changes can be large, I think a mini-require is required when those issues are addressed. To avoid delays, it's probably not necessary to formally schedule and conduct a mini-review -- just posting a email asking for discussion will be sufficient. Statistics ========== There were roughly 300 emails during formal review (I have a mbox archive if anybody needs it). The following persons submitted a review with a vote. Some of votes are "provisional", this is not reflected below, but affects the list of critical issues below. - Boris Schaeling (yes) - Roland Bock (yes) - Ralf Globisch (yes) - Sean Chittenden (yes) - Darryl Green (no) - Tom Brinkman (no) - Barend (yes) - Sam Gundry (yes) - Vicente Botet (yes) - Robert Stewart (yes) - Alexander Arhipenko (yes) - "M-Square" (yes) Critical issues =============== These issues must be addressed, and be passed through informal mini-review, before the library can be added to SVN. - It appears that the 'trivial' logging is not sufficiently good to just take and use. The library should provide an out-of-the-box mechanism that: - supports channel/severity attributes, without creating new logger for each severity. It also should be possible to filter out specific channel by essentially poking at some "map", and without manually composing a new filter that checks for a set of allowed channels sequentally. I believe that at least Christian, Roland, Robert and myself requested this. - has no observable effect on compile time. One second (or more) that trivial.hpp imposes is not acceptable. - outputs to console by default, but permits changing that - outputs only component/severity/message by default, but probably permits changing that. - By default, the library should try hard to continue working no matter what errors happen during processign a log record -- including things like invalid attribute names or types, or other exceptions. - The library should not reinvent wheels. In particular, custom implementations of TSS and RW mutex seems like a very bad idea. Use of a custom lambda implementation was pointed by many as not great -- we already have a couple, so let's not add more to the mix. - Library interface is too geared towards lambda. In particular, implementing formatters/filters as free-standing functions appears to require considerably more code. In particular, it should be possible to access an attribute using a single function call. Also, logging::extract taking a lambda for callback complicates matters, it's desirable to have a more direct variant. Important issues ================ These issues are important enough to be mentioned here, but can be addressed as convenient. - The reported run-time performance can be improved. In particular the performance with date attributes seems scary and 2x slowdown relative to some other library in the 'most records are filtered out' case is also concerning. - I don't think it was satisfactory explained why 'attr' exists in two namespaces, as opposed to being a single type that is used for both formatting and filtering. Other notes =========== These issues seem important and should be considered, but might not even be fixable within the current design. - Several reviewers said the library is "too flexible" and this makes it harder to understand. Probably nothing can be done about that at this point, but interesting data point notwithstanding. - It was suggested that the library include an API that another Boost log may use in a way that permits plugging another logging library should end user of a Boost so desire. This seems a good idea, but I think it's up to Andrey to implement such API or not. - Alternative design approaches were proposed, in particular (i) permitting user log records and templating/basing other components on a user-defined record type and (ii) using compile-time indexing to access attributes. Also, it was proposed that named parameters not be necessary to use the library, and a more conventional interface be available. It does not seem that implementing those suggestions is fully possible without writing a different library from scratch, so it's up to Andrey to consider those proposals. - It was suggested to use newer Spirit to address poor compile times for the library itself. Thanks, Volodya

4 4

[log] How to derive from a logger
by Andreas Huber 25 Mar '10

25 Mar '10

Hi Andrey As discussed earlier, we're kind of desperate to add functionality to a logger class provided by Boost.Log. Instead of duplicating severity_channel_logger_mt, I'm wondering whether it's a good idea to derive from it, as follows: class Logger : public ::boost::log::sources::severity_channel_logger_mt< Severity > { public: typedef ::boost::log::sources::severity_channel_logger_mt< Severity > logger_base; BOOST_LOG_FORWARD_LOGGER_CONSTRUCTORS(Tracer) void SetMinimumSeverity(Severity severity); Severity GetMinimumSeverity(); private: Severity minimumSeverity; }; Elsewhere we'd use this class as follows: BOOST_LOG_DECLARE_GLOBAL_LOGGER_INIT(Whatever, Logger) { // ... } It seems to work, but is this a good idea? Thanks & Regards, -- Andreas Huber When replying by private email, please remove the words spam and trap from the address shown in the header.

2 4

Re: [boost] GSOC 2010 Boost-Graph
by Chinmaya Poswalia 25 Mar '10

25 Mar '10

On Thu, Mar 25, 2010 at 3:21 AM, Chinmaya Poswalia < chinmaya.poswalia(a)gmail.com> wrote: > Hi Andrew, > I am finding all the graph topic(my favorite topic to work on) very > interesting that's why i asked a lot of questions regarding all in one mail. > > > - i want to work on semantic graph > > > - how about maintaining an adjacency list, denote V as nodes and E as > edges, so the space complexity will be O(V+E). > > > - Also we can employ pointers of type void* to point to edges/vertices > of different types so as to maintain the graph of different edge and vertex > types. > > > - rather than employing a vector in each node and edge either we should > allocate space during insertion or we maintain the size of vector array by > ourself. > > > - Through generic programming, all the operations being implemented on > graph up till now will be made applicable on these new members of the > library. > - various functions that can be induced on these graphs are > > > 1. Distribution of node degree > 2. finding clustering co-efficient of semantic graph > 3. relevance of a link and node > 4. scale on semantic graph > 5. disparity of connected types/dispersion co-efficient. > > > - Semantic graph are generating huge level of importance in networking > (network semantics on web), video screening, statistical analysis of > data,etc so it will be a great thing to work on it. > > Can you suggest some more links to read on for these > > > > I am working on my proposal about the graph and your help is crucial for > me. > With regards > Chinmaya >

1 0

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