March 2008 - Boost-users - lists.preview.boost.org

Ann: Floating Point Utilities Review starts today
by John Maddock 04 Mar '08

04 Mar '08

The review of Johan Rade's floating point utilities starts today. Code and docs can be downloaded from : http://www.boost-consulting.com/vault/index.php?action=downloadfile&filenam… The library consists of three parts: 1) Floating point classification routines: these are optimised implementations of the C99 and C++ TR1 functions fpclassify, isinf, isnan, isnormal and isfinite. From Boost-1.35 onwards these are already a part of Boost.Math (see http://svn.boost.org/svn/boost/trunk/libs/math/doc/sf_and_dist/html/math_to…) so if accepted the two implementations will get merged. The review here should focus on the implementation used, and testing on whatever platforms you have available - in particular are there any circumstances (compiler optimisation settings etc) where this implementation breaks? 2) Sign manipulation functions: implementations of the C99 and C++ TR1 functions copysign and signbit, plus the changesign function. Two of these (signbit and copysign) are currently undocumented members of Boost.Math, and again the two implementations will get merged if this library is accepted. Again the main focus of the review here is the implementation, and testing thereof especially in the presence of compiler optimisations. 3) C++ locale facets: these will read and write non-finite numbers in a portable and round-trippable way: that is not otherwise possible with current C++ std library implementations. These are particularly useful for number-serialisation for example. Since the design is already specified by the C++ standard for these facets, your review here should focus on implementation, testing, documentation, and perhaps where in Boost these should best be placed if accepted. These look to be a useful collection of utilities, so I'll look forward to your reviews, Regards, John Maddock, Floating Point Utilities Review Manager.

9 34

Boost::random: Using linear congruential etc.
by Hans Mull 04 Mar '08

04 Mar '08

Hi! A few days ago I started to try out the boost::random library. Now I have a problem: If you don't want to use the "special cases" like minstd_rand[0] for linear_congruential hellekalek1995 for inverse congruential algorithm, how do you use them? I have non-constant parameters (which are not know at compile time). It would be nice if you give me some examples about how to use these "general case" algorithms. Best regards, Hans

1 0

boost::iostreams zlib compression
by j.c. 04 Mar '08

04 Mar '08

Ok I think this sucks. :-) Here is my code to implement zlib compression/decompression via boost. However since is uses streams I cannot simply pass a buffer. This code is intended for compress socket buffers, but is pathetically slow. HAs anyone done this? What can I improve? The copy seems to be the bottle neck. thx jc Here is my code: #include <boost/iostreams/copy.hpp> #include <boost/iostreams/filter/zlib.hpp> #include <boost/iostreams/device/file.hpp> #include <boost/iostreams/stream.hpp> #include <boost/iostreams/filtering_stream.hpp> #include <boost/iostreams/detail/config/zlib.hpp> using namespace boost::iostreams; std::string compress(std::string data) { std::stringstream compressed; std::stringstream decompressed; decompressed << data; boost::iostreams::filtering_streambuf<boost::iostreams::input> out; out.push(boost::iostreams::zlib_compressor()); out.push(decompressed); boost::iostreams::copy(out, compressed); return compressed.str(); } std::string decompress(std::string data) { std::stringstream compressed; std::stringstream decompressed; compressed << data; boost::iostreams::filtering_streambuf<boost::iostreams::input> out; out.push(boost::iostreams::zlib_decompressor()); out.push(compressed); boost::iostreams::copy(out, decompressed); return decompressed.str(); } int main() { std::string data = compress("boo!"); std::cout << data << std::endl; std::cout << decompress(data) << std::endl; return 0; }

3 6

compiling 64 bit boost on Mac OS 10.5
by Tom Smith 04 Mar '08

04 Mar '08

Hi, I've been trying and failing to get a 64 bit build of the boost libraries on 10.5. I always seem to end up with 32 bit binaries. I've tried various permutations involving bjam address-model=64 (I last tried to build a week ago so I'm hazy on the actual command line). I only need x64, not ppc64. Cheers, Tom

3 5

iterator_adapter/iterator_tag/std::distance bug
by Martin Wartens 04 Mar '08

04 Mar '08

I used boost::iterator_adapter with a std::vector iterator, all parameters set to boost::use_default. (So this is the same as a plain std::vector iterator) With this configuration, the iterator tag is recognized as "input_iterator_tag" not "random_iterator_tag" as it must be. This hurts because the iterator tag is used to select the right std::distance function. I could solve the problem by explicitly setting the traversal type template parameter to "std::random_access_iterator_tag", while "boost::random_access_traversal_tag" also didn't work. The compiler is VC8.

2 2

serialization - xml
by Ambardekar, Mukund 03 Mar '08

03 Mar '08

Hi, I used boost serialization to write my data in binary format as in below way - std::ofstream ofs(catalogFileName.c_str(),std::ios::binary); boost::archive::binary_oarchive oa(ofs); //write oa & m_myclass; // object of CMyClass type containing some data (say name) This worked correctly. Now I wanted to write into xml format. Did the below changes std::ofstream ofs(catalogFileName.c_str(),std::ios::binary); boost::archive::xml_oarchive oa(ofs); //write oa & m_myclass; But it is giving me below error - d:\tools\boost-1feb08\boost_1_34_1\boost\archive\basic_xml_oarchive.hpp( 83) : error C2027: use of undefined type 'boost::STATIC_ASSERTION_FAILURE<x>' Can you help me to identify the problem? FYI - I have made no changed CMyClass looks like as below. And I have made no changes in it (same one used for binary and xml) class CMyClass { public: Part(const std::wstring& Name); private: std::wstring m_name; friend class boost::serialization::access; template<class Archive> void serialize(Archive& ar, const unsigned int) { ar & m_name; } }; Thanks in advance. Regards, Mukund

2 1

Minimal human-readable boost::serialization ?
by Eric Newhuis 03 Mar '08

03 Mar '08

I looks like the last time this was mentioned in this mailing list was in 2005. Has anyone considered anything like boost::archive::json_oarchive etc.? JSON looks easy to read in log files. I want name-value pairs in my serialization. But I don't want the verbosity of XML. I think JSON would be ideal.

2 1

Better debugger support for boost::optional with VS2005 compiler
by Daniel Krügler 03 Mar '08

03 Mar '08

Hello, just recently I introduced boost::optional as default type to represent optional/nullable types in a sub-project of our company, which generates C++ code out of UML models. While this worked seamless, programmers not yet used to boost::optional complained that they could not really see, what such type "contains" if they try to observe it in the debugger. I searched for a solution, only finding the discussion around the article "how to see the actual value of a 'boost::optional<__int64>' in VS2005 debugger" from 2006-07-26, which did not gave a satisfactory answer. Please ignore this message, if the here described solution (or a better one), which I worked out this weekend, is already well-known (but I would appreciate if someone could give me a link to the already known solution). My approach was to modify the unofficially supported debugger configuration file autoexp.dat, which typically is found at ${SomePath}\Microsoft Visual Studio 8\Common7\Packages\Debugger\autoexp.dat If you never have heard of this file, you should make a copy of it, before doing the modifications I discuss here. A more detailled description what can be done with this file, can be read here (but you don't need to do that, if you just want to get the results I propose here): http://www.virtualdub.org/blog/pivot/entry.php?id=120 Just use your preferred text editor to open the autoexp.dat, jump at the end of it and go five rows back. You should be +/- 1 row about the line with the content: ; This section lets you define your own errors for the HRESULT display. Go one row back (row 1957 in my file), which should be the end of the so-called [Visualizer] section. Now insert here the text provided at the end of this article and just start you debugger session (you don't need to restart the studio) - you should have a nice view on each boost::optional, even those which are optionals of references. I tested the visualizer for the following boost versions: 1.30.0 1.31.0 1.32.0 1.33.0 1.33.1 1.34.1 and all of them worked. I would like to thank the boost developers for their great contributions (where boost::optional is only one of them) and I hope you can even more enjoy boost::optional with this trick. Unfortunately I have not found similar solutions for other compilers yet, e.g for gcc, which probably is due to my lack of experience with them but I would appreciate if someone else publishes them, too. Greetings from Bremen, Daniel Krügler ;------------------------------------------------------------------------------ ; Visualizer for boost::optional, boost::optional_detail::optional_base ;------------------------------------------------------------------------------ boost::optional<*>|boost::optional_detail::optional_base<*> { preview ( #if($e.m_initialized) ( *(($T1 *) (&$e.m_storage.dummy_.data[0])) ) #else ( "?" ) ) children ( #if($e.m_initialized) ( #(value: *(($T1 *) (&$e.m_storage.dummy_.data[0])) ) ) #else ( "?" ) ) } ;------------------------------------------------------------------------------

3 3

[filesystem] bug???
by Ovanes Markarian 03 Mar '08

03 Mar '08

Hello *! A short question: removing the leaf from the path using the remove_leaf() results in my case (MS VC7.1, boost 1.34.1) in the path with the trailing slash at the end. Next call to remove_leaf() results in removing this trailing slash, which is not correct... Is it a bug, if the doc states to remove_leaf: basic_path& remove_leaf(); Effects: If has_branch_path() then remove the last filename from the stored path. If that leaves the stored path with one or more trailing slash elements not representing root-directory, remove them. Many thanks, Ovanes

1 0

A possible bug in Boost.property_tree with UNICODE XML file?
by Tan, Tom (Shanghai) 03 Mar '08

03 Mar '08

I am trying read_xml() on an xml-format config file, which starts like this: <?xml version="1.0" encoding="UTF-16"> ..... This file is well-formed and verified by "XML Copy Writer", it can also be open by IE or Filefox with no problem. I am using this code snippet to parse this file: try{ read_xml<wptree>(xmlfile,pt); } catch(xml_parser-error e) { e.what(); } and it throws an exception saying "expected <". By looking into the code I found that, this caused by the fact that the xmlfile starts with a Unicode Byte Order Mark (BOM)-0xFFFE . "fffe 3c00 3f00 7800 ......" However, read_xml() cannot handle this BOM byte and throws. It's easy to fix it by simply discard the BOM byte, but I wonder whether there is already a flag to enable/disable this that I am unaware of. Can anyone help to clarify? Thanks, tom

1 0