- Boost-users - lists.boost.org

[Multi_index] Performance like sequenced.cpp example
by Manuel Jung 30 Apr '07

30 Apr '07

Hi, I have to count a lot of words. Up to now i did it with MySQL, because it was easy. The result is safed there anyway. Now i thought i could speed up this a little if i would use internally a Multi_index list to store the words, so i have only to insert all different words. The words are stored in a UnicodeString from the ICU library. My code is really near to the one from the example "sequenced.cpp". Im using the following definition: typedef multi_index_container< UnicodeString, indexed_by< sequenced<>, ordered_non_unique<identity<UnicodeString> > > > text_container; typedef nth_index<text_container,1>::type ordered_text; text_container tc; Im inserting new words with "tc.push_back(UnicodeString(NewWord));" And count them exactly like in the example. I thought this should be fast, but it isnt. It eats up all my CPU, but isnt fast. It is a lot slower than my old solution. I have still hope i could speed this up, before i have to switch back MySQL. The profile of a run says that "boost::multi_index::safe_mode::check_same_owner<..." eats most CPU time. Some suggesting how to speed it up with Multi_index? Or some ideas which other way would be faster than MySQL inserts? Thanks Manu

3 6

An easy to use grid like random access/dynamic container?
by Jeff 30 Apr '07

30 Apr '07

I have been using STL containers to store structures of data but I would like to know if there is an easier solution such as a dynamically sized container that stores data in a spreadsheet like fashion. For example, I would like to store rows of columns of different data types. Here is a sample of data I would like to store, also, I would like fast random access to the elements to read and modify existing data: Name, Age, HomeOwner John 42 true Mark 24 false Mary 29 true If anyone can suggest such a container please let me know. Thank you

3 3

[unit test framework] undef ref to main() when dyn linking unit_test_framework lib, boost 1.34, linux
by Chris Fairles 30 Apr '07

30 Apr '07

Testing out boost beta 1.34. I built the whole she-bang with the gcc toolset. I tried compling my unit tests and they all get: /usr/lib/gcc/x86_64-redhat-linux/4.1.1/../../../../lib64/crt1.o: In function `_start': (.text+0x20): undefined reference to `main' collect2: ld returned 1 exit status I don't use the auto test case mechanism. My tests suites look like: #include "boost/test/unit_test.hpp" using boost::unit_test::test_suite; void free_test_function() { BOOST_CHECK(1 == 1); } test_suite* init_unit_test_suite( int, char* [] ) { test_suite* test= BOOST_TEST_SUITE( "Unit test example 1" ); test->add( BOOST_TEST_CASE( &free_test_function ) ); return test; } I compile using g++ -Wall -I/usr/local/include/boost-1.34 -o unit_test unit_test_test.cpp -L/usr/local/lib64 -lboost_unit_test_framework-gcc41-mt-d If i define BOOST_TEST_DYN_LINK either in the cpp file before or after the unit_test.hpp include, or specifify it using -D on the command line I get the same results. I tried to look through the headers responsible for defining BOOST_TEST_DYN_LINK but theres a lot of references to "msvc auto-linking" which I of course do not use. Should I be defining BOOST_ALL_DYN_LINK ? I recall it not being set somewhere because boost.build v2 doesn't like it or something. Really i can't make heads or tails of what I should be defining, whether I shouldve defined something when building the unit test libs, or whether I should be defining something building my own. The docs say something about defining BOOST_TEST_DYN_LINK but as stated above, it doesn't seem to make a difference. Chris

2 3

Boost.Test and mpatrol
by KLINIK Markus SDI2-G (AREVA NP GmbH) 30 Apr '07

30 Apr '07

Hi, there seems to be a problem if Boost.Test and mpatrol are used together. The following program hangs and constantly uses more and more memory. How to reproduce the behaviour: System: Windows XP Prof. SP2 MS Visual Studio 7 mpatrol-1.4.8 boost-1.33.1 <code> #define BOOST_AUTO_TEST_MAIN #include <boost/test/auto_unit_test.hpp> BOOST_AUTO_TEST_CASE(testA) { int* a = new int[100]; } </code> After "sourcing" vsvars32.bat the following command was used to compile & link the sample. cl -Zi -EHsc -ML -Ic:\boost\include\boost-1_33_1 test.cpp -link -libpath:c:\local\lib mpatrol.lib -libpath:c:\boost\lib libboost_unit_test_framework-vc7-s.lib Static linkage of mpatrol results in the same behaviour: cl -Zi -EHsc -ML -Ic:\boost\include\boost-1_33_1 test.cpp -link -libpath:c:\local\lib libmpatrol.lib imagehlp.lib -libpath:c:\boost\lib libboost_unit_test_framework-vc7-s.lib Same with -MT, mpatrolmt.lib and libboost_unit_test_framework-vc7-mt-s.lib The combinations -MLd, mpatrol.lib and libboost_unit_test_framework-vc7-sgd.lib -MTd, mpatrolmt.lib and libboost_unit_test_framework-vc7-mt-sgd.lib do not result in the erroneous behaviour, but memory debugging is buggy because mpatrol and the M$ debug malloc implementation are incompatible (according to the mpatrol documentation/mailing list). I've posted the bug description here because I don't know where to start. Even "it is their fault" pointers would be helpful. Thanks in advance, Markus Appendix: real quick howto install mpatrol (to make it work for this example): - download from http://www.cbmamiga.demon.co.uk/mpatrol/files/mpatrol_1.4.8.tar.gz - unpack to c:\src - cmd prompt to C:\src\mpatrol\build\windows - "source" vsvars32.bat (YMMV): $ "c:\Programme\Microsoft Visual Studio .NET\Common7\Tools\vsvars32.bat" - $ nmake -f NMakefile all - copy mpatrolmt.lib to c:\local\lib - copy libmpatrolmt.lib to c:\local\lib - copy mpatrolmt.dll to a directory in $PATH (or local directory)

2 1

[SPAM:## 74%] [boost] [Review] The review of the Boost.Guid library starts today (April 30th)
by Hartmut Kaiser 30 Apr '07

30 Apr '07

Hi all, The review of the proposed Boost Guid library written by Andy Tompkins starts today (April 30th, 2007) and ends May 10th, 2007. You can download the library here: http://www.boost-consulting.com/ vault/ (file guid_v7.zip). --------------------------------------------------- About the library: Guid is a library providing an implementation of Globally Unique Identifiers. Guid's are used in many facets of programming, including databases, networks, remote procedures, serializations, transactions, identifying documents, and identifying classes. This library is efficient and is able to create name-based and random-based guids. --------------------------------------------------- Please always state in your review, whether you think the library should be accepted as a Boost library! Additionally please consider giving feedback on the following general topics: - What is your evaluation of the design? - What is your evaluation of the implementation? - What is your evaluation of the documentation? - What is your evaluation of the potential usefulness of the library? - Did you try to use the library? With what compiler? Did you have any problems? - How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? - Are you knowledgeable about the problem domain? Regards Hartmut Review Manager _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/ listinfo.cgi/boost

1 0

[-s- lib versions][vc80] new user Q: what's the difference between lib*vc80-mt-1_33_1.lib and lib*vc80-mt-s-1_33_1.lib
by John McMahon 30 Apr '07

30 Apr '07

Hi I'm a new boost user. I downloaded the pre-built Visual Studio 2005/vc80 version of boost from http://www.boost-consulting.com/products/free and read the instructions at http://www.boost.org/more/getting_started.html , specifically the section on library runtimes. I have a question; In the context of vc80, and for components which have libraries, what's the difference between -s- and non-s static libs, e.g. for 'thread' -rw-r--r-- 1 John None 516994 Mar 25 2006 libboost_thread-vc80-mt-1_33_1.lib -rw-r--r-- 1 John None 845330 Mar 25 2006 libboost_thread-vc80-mt-s-1_33_1.lib thanks in advance John McMahon

2 1

[filesystem] Can't create absolute paths from strings on windows
by Ken Klose 29 Apr '07

29 Apr '07

cout << "init path " << initial_path().string() << endl; cout << "init path exists " << exists(initial_path()) << endl; prints: init path C:/Documents and Settings/me/workspace/3.2/DciApiTest init path exists 1 But if I try to construct the same path from a string: I get: path init_path2(initial_path().string(), no_check); // I've also tried 'native' checker cout << "init path 2 " << init_path2.string() << endl; cout << "init path 2 exists " << exists(init_path2) << endl; I get: init path 2 C:Documents and Settings/ken.klose/workspace/3.2/DciApiTest init path 2 exists 0 The / after the : gets lost and therefore the file isn't found. Thanks in advance! BTW: Using MinGW.

1 2

iostreams and compression
by Aljaz 29 Apr '07

29 Apr '07

Hello I've been trying to figure out another way to compress data from/to std::stringstream. Here is the source how I do it now: std::stringstream decompressed; decompressed << "somestring....."; std::stringstream compressed; { boost::iostreams::filtering_streambuf<boost::iostreams::input> out; out.push(boost::iostreams::zlib_compressor()); out.push(decompressed); boost::iostreams::copy(out, compressed); } Is there any way around this? I want to avoid boost::iostreams::copy because it takes 40% of this source execution time. How can I compress data directly to stringstream (to compressed in this example) without copying it? Thank you very much for help Regards Aljaz

1 0

Re: [Boost-users] [thread] Segfault constructing a thread
by Ken Klose 28 Apr '07

28 Apr '07

Peter, Thank you very much. Your suggestions have fixed it. Ken "Peter Dimov" <pdimov(a)mmltd.net> wrote in message news:<008b01c789b2$21b0fe20$6607a8c0@pdimov2>... > Ken Klose wrote: > > I've got a singleton object that is trying to run its operator() > > method in a separate thread, but I'm getting a segfault when I > > construct the thread object: > > > > PublishedListener::PublishedListener(string directory) > > { > > _thread = new thread(*this); // segfault here > > } > > This line makes a copy of *this. It's very likely that this isn't what you > want. Use ref(*this) to store a reference, and consider disabling the copy > constructor and the assignment operator of the class.

1 0

Re: [Boost-users] [thread] Segfault constructing a thread
by Ken Klose 28 Apr '07

28 Apr '07

Thanks for replying! The operator()() method is just a stub at the moment: void PublishedListener::operator()() { this->run(); } void PublishedListener::run() { cout << "Hooray! My thread is running" << endl; } I tried stepping through the construction of the thread and it looks like the problem is occurring somewhere in the construction of the function0<void>. As a test I changed the constructor to this: PublishedListener::PublishedListener(string directory) { cout << "1" << endl; function0<void> runnable(*this); cout << "2" << endl; _thread = new thread(runnable); cout << "3" << endl; } And I get the problem between 1 and 2. Thanks, Ken PS - I see Peter Dimov suggested I use ref. I'm going to try that now and reply to his post. "Digvijoy Chatterjee" <digvijoy.c(a)gmail.com> wrote in message news:<2f7228250704271831y58640424ld32d3cfe0c1bbc50(a)mail.gmail.com>... > On 4/27/07, Ken Klose <ken.klose(a)future-skies.com> wrote: > > > > I've got a singleton object that is trying to run its operator() method in > > a > > separate thread, but I'm getting a segfault when I construct the thread > > object: > > > > PublishedListener::PublishedListener(string directory) > > { > > _thread = new thread(*this); // segfault here > > } > > > > PublishedListener.h: > > class PublishedListener > > { > > public: > > static PublishedListener &getInstance(); > > void run(); > > void operator()(); > > boost::filesystem::path getItemDirectory(); > > virtual ~PublishedListener(); > > private: > > PublishedListener(std::string directory); > > static PublishedListener *_instance; > > boost::thread *_thread; > > boost::filesystem::path *itemDirectory; > > }; > > > > Thanks in advance! > > > Are you sure the implementation of operator()() isnt seg faulting, the > thread should be able to invoke it properly and i dont see any reason for > the thread instantiation to segfault.. > HTH > Digz >

1 0