August 2010 - Boost-users - lists.preview.boost.org

[log] cost of record instantiation
by Matthias Vallentin 31 Aug '10

31 Aug '10

Hello, I am trying to understand the cost of log record instantiation. Say I have a global severity logger with several levels, of which the debug level produces a lot of output and record creation is expensive. Because the overhead should only incur when actually logging in debug mode, I would like to know how costly the level check is. That is, in the following code // The custom logger class here accommodates multiple severity // loggers for different facilities. #define LOG(level, facility) \ BOOST_LOG_SEV(global::logger->get(facility), level) LOG(debug, core) << ... some expensive construction would the stream expression be instantiated regardless of the logging level, yet simple be discarded when being in level lower than debug? (I am asking because I recall they are not streamed in a lazy fashion.) If this is the case, preventing the construction of the stream expression would need a different logging mechanism, e.g., /// Expensive logging when in Debug mode. #ifdef DEBUG #define LOGD(facility) \ LOG(log::debug, facility) #endif // Not being compiled. LOGD(core) << ... some expensive constrution However, this design does not have a uniform approach to logging, which makes it less attractive. Any thoughts? Matthias

2 1

[Boost.DateTime][Boost.Python] Year out of valid range error when streaming via Python module
by SamuelG 31 Aug '10

31 Aug '10

This problem occurs at a higher level but I've tried to strip it back. I've reproduced this problem GCC 4.3.4, Python 2.6.5 and Boost 1.41.0 32bit and 64bit Debians/Ubuntu. The problem does not seem to happen on Windows XP. I have embedded the Python interpreter in the attached http://old.nabble.com/file/p29579563/run_python.cpp run_python.cpp . When running this with the attached script http://old.nabble.com/file/p29579563/test.py test.py and which uses a test module http://old.nabble.com/file/p29579563/test_date_time.cpp test_date_time.cpp , the following Boost.DateTime error occurs: $ run_python test.py out() - entry out() - to_simple_string(): 2010-Aug-31 14:30:33.439172000 out() - manual format: 2010-Aug-31 14509:17:19.172000 out() - oss.str(): 2010-Aug-31 14509:17:19.172000 out() - now: Traceback (most recent call last): File "test.py", line 6, in <module> test_date_time.output() IndexError: Year is out of valid range: 1400..10000 However, when running the same script via python it is fine: $ python test.py out() - entry out() - to_simple_string(): 2010-Aug-31 14:30:31.208160000 out() - manual format: 2010-Aug-31 14:30:31.208160000 out() - oss.str(): 2010-Aug-31 14:30:31.208160000 out() - now: 2010-Aug-31 14:30:31.208160000 out() - exit Note that the time stamp in the first example is corrupt. Note also that the valid range error does not occur when using the date and time facets manually but when streaming the ptime object to std::cerr. The problem first arose for me when using Boost.Log with separate Python modules via *both* run_python and /usr/bin/python. See http://sourceforge.net/projects/boost-log/forums/forum/710022/topic/3748054 http://sourceforge.net/projects/boost-log/forums/forum/710022/topic/3748054 For this simple case, /usr/bin/python seems okay but the run_python error is identical to what happens in our proper application (where Boost.Log is used and multiple Python modules loaded). Has anyone witnessed similar behaviour? Or can they please test my example out? Thanks for any help, Sam -- View this message in context: http://old.nabble.com/-Boost.DateTime--Boost.Python--Year-out-of-valid-rang… Sent from the Boost - Users mailing list archive at Nabble.com.

1 0

Re: [Boost-users] boost::asio SSL connection thru proxy server
by Kumar, Akhilesh 31 Aug '10

31 Aug '10

Hi All, Considering that nobody replied my question I am wondering if anybody who has knowledge of this kind of problem has seen my previous message/question. I really need help on this issue if anybody has any thought on this pls. reply. Thanks, Akhilesh Kumar From: Kumar, Akhilesh Sent: Monday, August 23, 2010 5:10 PM To: 'boost-users(a)lists.boost.org' Subject: boost::asio SSL connection thru proxy server Hi, I am writing a boost socket client using boost::asio to do secure HTTP (HTTPS) communication. I Implemented my client based on the code sample based on (http://www.boost.org/doc/libs/1_39_0/doc/html/boost_asio/example/ssl/client…) with little changes as I am using sync socket. My client woks fine, when I am connecting to server directly, but I am not sure what changes I need to make when my communication going thru proxy. The unsecure connection(HTTP) work fine thru proxy, and here is what I do for unsecure connection, 1. Initialize boost socket(IO service and other stuff). 2. Resolve and Connect to Proxy server. 3. Start writing to actual server(HTTP send request). 4. Start uploading/downloading. 5. Close Socket. But I can't use this above procedure for (SSL) secure connection as I need to do handshaking with the server before I can write anything if my understanding is correct. Can somebody give me some pointer or code sample or step I need to do when I am doing secure connection thru proxy server. Just to clarify Microsoft WinInet does do (secure) HTTPS connection thru same proxy fine. If providing my code sample I am using for SSL might help understand the problem I can provide that. Thanks in advance for any help. Akhilesh Kumar

3 5

Re: [Boost-users] Boost.Serialization: Restoring object state (preventcreation of a new object)?
by ray＠maya-studio.com 30 Aug '10

30 Aug '10

>> Just avoid serialization through a pointer, serialize either the >> object or a reference to the object. Thanks, Robert. Serializing through a reference seems to be the only viable solution in my case. However, I'm not able to make it work (I didn't find any specific information about references in the library documentation - except for member variables) . For example: A a; ... A& aRef = a; ia & BOOST_SERIALIZATION_NVP(aRef);; // Throws an exception in load_override

2 1

[parameter] Using named parameter types in return type
by Jeremiah Willcock 30 Aug '10

30 Aug '10

In BOOST_PARAMETER_FUNCTION, one argument to specify is the return type of the function. How do I refer to the types deduced for the named parameters in the return type? I am creating a factory function for a class template (similar to std::make_pair) and would like to use named parameters for it. I did not see any obvious way to do this in the documentation. Is there a way? A documented name for the ArgPack? I am using C++0x so it would be OK to have a solution that requires that. -- Jeremiah Willcock

1 0

Re: [Boost-users] [filesystem] Using filesystem library from Boost version 1.44 on Cygwin 1.7
by Clarke, Carus V 30 Aug '10

30 Aug '10

> Date: Mon, 30 Aug 2010 08:36:49 -0700 > From: Steven Watanabe <watanabesj(a)gmail.com> > To: boost-users(a)lists.boost.org > Subject: Re: [Boost-users] [filesystem] Using filesystem library from > Boost version 1.44 on Cygwin 1.7 > Message-ID: <4C7BD011.6010707(a)providere-consulting.com> > Content-Type: text/plain; charset=ISO-8859-1; format=flowed > > AMDG > > Clarke, Carus V wrote: > > Steven Watanabe wrote: > >> Your code doesn't appear to print this last line. > > > > You are right. I added the last line after pasting in the > code. Here it is: > > > > boost_fs::file_status st = boost_fs::status(path, ec); > > <snip> > > std::cout << "Does file exist? " << boost_fs::exists(st) > << std::endl; > > return 0; > > } > > > > Is there anything I can do to make this work? > > > > It looks like stat is setting errno to ENOENT. > Do you get the same behavior if the file exists before > starting the program? > > In Christ, > Steven Watanabe > Yes, the behavior is the same whether the program writes the file or if the file had been previously written. Thanks, Carus V. (Bud) Clarke carus.v.clarke(a)boeing.com

2 1

Boost.Serialization: Restoring object state (prevent creation of a new object)?
by ray＠maya-studio.com 30 Aug '10

30 Aug '10

If an object is serialized (saved) from a pointer, the loading part automatically creates (instantiates) a new object. However, there are situations when the object is already created and all we need is to restore its state. The following code demonstrates one possible - although not elegant - solution using a temporary object:#include <fstream> #include <iostream> #include <boost/archive/xml_iarchive.hpp> #include <boost/archive/xml_oarchive.hpp> #include <boost/serialization/string.hpp> #include <boost/serialization/export.hpp> class A { public: A() : m_val(100) { } int m_val; private: friend class boost::serialization::access; template <typename Archive> void serialize(Archive &ar, const unsigned int version) { ar & BOOST_SERIALIZATION_NVP(m_val); } }; int main(int argc, char* argv[]) { A *a = new A(); // Save the object { std::ofstream os("d:/test.xml"); boost::archive::xml_oarchive oa(os); // Save from pointer oa & BOOST_SERIALIZATION_NVP(a); } // Change the object state a->m_val = 200; // Restore the state { std::ifstream is("d:/test.xml"); boost::archive::xml_iarchive ia(is); A* aTemp; ia & BOOST_SERIALIZATION_NVP(aTemp); a->m_val = aTemp->m_val; // Copy object state assert(a->m_val == 100); delete aTemp; } } Here is my question: Is there some more convenient way to restore the object state? In other words, is it possible to prevent creation of a new object and just 'populate' an existing object?

2 1

[filesystem] Using filesystem library from Boost version 1.44 on Cygwin 1.7
by Clarke, Carus V 30 Aug '10

30 Aug '10

Date: Fri, 27 Aug 2010 15:39:29 -0700 From: Steven Watanabe <watanabesj(a)gmail.com> To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] [filesystem] Using filesystem library from Boost version 1.44 on Cygwin 1.7 Message-ID: <4C783EA1.4070400(a)providere-consulting.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed AMDG Clarke, Carus V wrote: > I have been using the filesystem library from Boost version 1.43 on a Cygwin 1.7.6-1 platform without any problems. The library was built using the Cygwin version of GCC 4.3.4. The library from the 1.44 version of Boost compiles compiles o.k., but the functions that access files no longer work. This appears to be true for both versions 2 and 3 of the library. Below is a simple program that runs o.k. with the Boost version 1.43 version of the library. The program runs but can not read the file status using Boost version 1.44. This program also works fine when built using MSVC 9.0 and Boost version 1.44. > > //-------------------- > > <snip> > > int main() > { > <snip> > std::cout << "Error code message: " << ec.message() << std::endl; > return 0; > } > > //-------------------- > > This program's output on my system is shown below. The file 'foo' is written by the program and does actually exist. > > For version 2: > > boost::filesystem library version is v2 > Checking status of file foo... > Error code value: 0 > Error code message: The operation completed successfully > Does file exist? false > Your code doesn't appear to print this last line. In Christ, Steven Watanabe ------------------------------ You are right. I added the last line after pasting in the code. Here it is: boost_fs::file_status st = boost_fs::status(path, ec); <snip> std::cout << "Does file exist? " << boost_fs::exists(st) << std::endl; return 0; } Is there anything I can do to make this work? Thanks, Carus V. (Bud) Clarke carus.v.clarke(a)boeing.com ------------------------------

2 1

[parameter] Weird behavior with bools
by Ryan McConnehey 30 Aug '10

30 Aug '10

I'm new to the parameter library and started with a simple example to get my feet wet. The first problem I'm having is with type restriction. The "name" variable is being restricted to a boolean type. This doesn't seem to be enforced since I can pass this variable a value of 15 and get this same value out. Currently my output has "name = 15" and I expected the output to be "name = 1". Did I not correctly enforce the type? The second problem is with the default value of my variable. I thought if the variable wasn't used then the third parameter, in the declaration, is used for instantiation of that variable. I thought that meant I could do this "m_name(args[_name])". The compiler gives an error saying the operator[] can't deduce the arguments. What did I not do that would allow this behavior. Thank you for your time in helping me out. Ryan BOOST_PARAMETER_NAME(name) struct myclass_impl { template <class ArgumentPack> myclass_impl(ArgumentPack const& args) : m_name(args[_name | false]) //: m_name(args[_name]) //gives a compile error { std::cout << "name = " << m_name << std::endl; } private: int m_name; }; struct myclass : public myclass_impl { BOOST_PARAMETER_CONSTRUCTOR( myclass, (myclass_impl), tag , (optional (name, (bool), bool(false)))) // no semicolon }; int main(int argc, char* argv[]) { myclass x(_name = 15); return 0; }

4 7

Can I fstat or get the file date/time for a file stream ?
by Timothy Madden 30 Aug '10

30 Aug '10

Hello Is there a way to stat or to get file last accessed time from an opened std::fstream or boost::filesystem::fstream or a boost::iostreams::stream ? Or is there a way to get the underlaying file descriptor (or a FILE *) from a fstream so I can call POSIX fstat ? Thank you, Timothy Madden

1 0