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

Need help with XCode - simple RegEx test app crashes on OS X (Darwin)
by Thomas Tempelmann 20 Dec '08

20 Dec '08

I wanted to try out boost for the first time and ran into a basic problem: The regex test app as suggested on the installation page crashes inside the destructor of "match_results". I wonder if the libs created by the boost installer are built in an incompatible manner to what XCode generates from my own test code. I could not find any troubleshooting or special instructions for using boost with XCode, though. Here's some more details: I've installed boost 1.37, on two Macs (both Intel, one with 10.5.5, other with 10.5.6), both XCode 3.1.2 (gcc 4.01). Once installed using the official distribution, using "configure" and "make install", once using MacPorts. Both build error-free, and similarly (only difference I could see is that the former installs in /usr/local/ while the other in /opt/local). I basically followed the unix install instructions from here: http://www.boost.org/doc/libs/1_37_0/more/getting_started/unix-variants.html #link-your-program-to-a-boost-library Chapter 6 then suggests to test a little regex program. I created a new C++ Tool project in XCode, put the test code into main.cpp and changed only these three build options: HEADER_SEARCH_PATHS = /opt/local/include LIBRARY_SEARCH_PATHS = /opt/local/lib OTHER_LDFLAGS = -lboost_regex-mt The app then builds fine without errors. And I can launch it, where it then waits for input. Meaning that boost basically works. It even performs the regex query just fine. Only then, when it loops around, it crashes with a segmentation fault deep inside the destructor of "boost::smatch matches". Here's the stack trace: Exception Type: EXC_BAD_ACCESS (SIGSEGV) Exception Codes: KERN_INVALID_ADDRESS at 0x0000000067616c66 Crashed Thread: 0 Thread 0 Crashed: 0 libstdc++.6.dylib 0x92c57964 __gnu_debug::_Safe_iterator_base::_M_detach() + 62 1 libstdc++.6.dylib 0x92c579ce __gnu_debug::_Safe_iterator_base::_M_attach(__gnu_debug::_Safe_sequence_base *, bool) + 42 2 libstdc++.6.dylib 0x92c57ad4 __gnu_debug::_Safe_sequence_base::_M_detach_all() + 42 3 boost-test 0x00002c87 __gnu_debug::_Safe_sequence_base::~_Safe_sequence_base() + 23 4 boost-test 0x00002feb __gnu_debug::_Safe_sequence<__gnu_debug_def::vector<boost::sub_match<__gnu_c xx::__normal_iterator<char const*, std::basic_string<char, std::char_traits<char>, std::allocator<char> > > >, std::allocator<boost::sub_match<__gnu_cxx::__normal_iterator<char const*, std::basic_string<char, std::char_traits<char>, std::allocator<char> > > > > > >::~_Safe_sequence() + 23 5 boost-test 0x00003449 __gnu_debug_def::vector<boost::sub_match<__gnu_cxx::__normal_iterator<char const*, std::basic_string<char, std::char_traits<char>, std::allocator<char> > > >, std::allocator<boost::sub_match<__gnu_cxx::__normal_iterator<char const*, std::basic_string<char, std::char_traits<char>, std::allocator<char> > > > > >::~vector() + 29 6 boost-test 0x000034db boost::match_results<__gnu_cxx::__normal_iterator<char const*, std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<boost::sub_match<__gnu_cxx::__normal_iterator<char const*, std::basic_string<char, std::char_traits<char>, std::allocator<char> > > > > >::~match_results() + 23 7 boost-test 0x00002809 main + 215 (main.cpp:18) 8 boost-test 0x00002052 _start + 216 9 boost-test 0x00001f79 start + 41 Any ideas what's going wrong here?

2 2

Re: [Boost-users] [boost] tweaking the review process
by David Abrahams 20 Dec '08

20 Dec '08

on Fri Dec 19 2008, Fabio Fracassi <f.fracassi-AT-gmx.net> wrote: > David Abrahams schrieb: >> on Wed Dec 17 2008, Fabio Fracassi <f.fracassi-AT-gmx.net> wrote: >> >>> David Abrahams schrieb: >>>> on Mon Dec 15 2008, John Phillips <phillips-AT-mps.ohio-state.edu> wrote: >>> [...] >>>>> I don't know if I was the first to do this, but I have done it. Looking at the >>>>> current contents of the instructions to review managers, either they have changed >>>>> since I started managing reviews or I somehow hallucinated extra words to add the >>> user >>>>> list to the lists that should receive an announcement of the review. >>>> An announcement is one thing; instructing people to post their reviews >>>> in the same list is a different thing entirely. >>>> >>> You don't have to, it happens almost automatically if you are not >>> careful. Many people write their Review as a reply to the original >>> announcement, and if you are not extra careful to remove the user list >>> from the recipients your review gets send to both lists. >> >> The poster of the announcement should make it clear how to post a >> review. I never liked the idea of reviews showing up in the list with >> the subject "Review period for Boost.XXXX begins today" anyway. Also >> s/he can set the Reply-To: field. >> > > When using thunderbird as a client to the gmane news gateway the Reply to fields don't > seem to work though. > I don't know if there is a technical solution to this particular > problem. I've never had a problem with Gnus' Reply-to functionality and gmane. It's hard to imagine that TBird is broken in this way. > Beside this I agree with you, if the original anouncement clearly states that reviews > and followups should go to .devel. > Still I think that sometimes some messages might go astray and maybe we should suggest > that the review manager and/or the submitter should have an eye on the users list, so > that nothing is lost. (and maybe forward those strays to .devel) Not a bad suggestion. -- Dave Abrahams BoostPro Computing http://www.boostpro.com

1 0

Re: [Boost-users] [bind] code sample crashing MS compiler
by loadcom 20 Dec '08

20 Dec '08

In my understanding of boost::bind, your code has no problem. It's also accepted by my VS2003SP1 environment. B/Rgds Max ----- Original Message ----- From: Christian Henning To: boost-users(a)lists.boost.org Subject: [Boost-users] [bind] code sample crashing MS compiler Date: 2008-12-20 05:42:00 This following code sample will crash the MS compiler version 8. I hope there are some MS folks reading this mailing list, since I have no idea where else to post this. #include struct A { void operator() () { boost::bind( &A::lookup, *this, unsigned char() )(); } void lookup( unsigned char& c ) {} }; int main(int argc, char* argv[]) { A()(); return 0; } I know the code is not suppose to compile. Uncommenting the this pointer in the bind construct will make the code compile fine. Regards, Christian ------------------------------------------------------------------- 新浪空间——与朋友开心分享网络新生活！(http://space.sina.com.cn/ )

1 0

error C2679: binary '<<' : no operator found which takes a right-hand operand of type 'const is a user-defined type, however, I'm getting the following:'
by Joseph A. Marks 20 Dec '08

20 Dec '08

Has anyone seen an error like this? SpecialDate is a user-defined type, however, I'm getting the following: boost/test/test_tools.hpp(326) : error C2679: binary '<<' : no operator found which takes a right-hand operand of type 'const is a user-defined type, however, I'm getting the following: ' (or there is no acceptable conversion) Thanks in Advance, Joseph A. Marks e-mail: josephamarks(a)yahoo.com

4 3

Re: [Boost-users] [C++] Boost::Regex beginner question
by Alan M. Carroll 20 Dec '08

20 Dec '08

I don't know if I can help you with that. A few things to consider 1) Why do you use boost::regex but regex_replace? I would try boost::regex_replace and see if that makes a difference. 2) The boost::regex_replace function is not in the boost/regex.hpp header, it's in another file (regex_replace.hpp) in a subdirectory which is included by regex.hpp. I would verify that you have a header with that function in it. 3) You could try downloading the Boost library and see if the code compiles. Regex requires a library but if does compile you know your current headers are broken in some way. 4) I would see what happens if you don't include windows.h. Every time I have included that and any STL/Boost headers I have had problems, so I no longer use windows.h in any real source file, only in simple glue logic ones. windows.h is an almost perfect storm of incompatibility and bad practice. At 09:44 AM 12/19/2008, P.C. wrote: >thank you for the response. unfortunately, we've tried to use the >regex_replace function, but it gave an unexpected error: >'regex_replace' : undeclared identifier >we used it in the following way: >NOTE: the Regex++ we are using was downloaded from the original >author's site, not entirely with the Boost library. we can't >understand this error. if we downloaded and installed it according to >the installation instructions for VC++, how can a function be missing?

1 0

Base16, 32, 64 Encoding / Decoding
by Jeff Dunlap 19 Dec '08

19 Dec '08

I had really hoped to find something in boost but didn't. I looked on the web and found various routines to encode data and came across a pretty complete library that encodes and decodes base16, base32, base64. Most are for either base16 or base32 but CyoEncode by Cyotec covers the three of them and the source code is available online. The function parameters seem very unfamiliar to me, for example: unsigned long Base32Encode( void* dest, const void* src, unsigned long size ); unsigned long Base32Decode( void* dest, const void* src, unsigned long size ); There are no samples online so I don't know how to use it. Since I don't understand those function params above, I just came up with the code below (seems to encode and decode base64 correctly but not for base32, base16: // encode char * buf1; buf1 = new char [100]; const std::string s = "1010"; CyoEncode::Base16Encode(buf1, s.c_str(), s.length()); // CyoEncode::Base16EncodeGetLength(s.length())); std::cout << "[" << buf1 << "]" << std::endl; // decode char * buf2; buf2 = new char [100]; const std::string s2(buf1); CyoDecode::Base16Decode(buf2, s2.c_str(), s2.length()); std::cout << "[" << buf2 << "]" << std::endl; Can someone review the Base32Encode params and let me know how this is really supposed to be used? Thank you very much, Jeff -- ELKNews FREE Edition - Empower your News Reader! http://www.atozedsoftware.com

2 2

Asking for feedback for book (in German) about the Boost C++ libraries
by Boris 19 Dec '08

19 Dec '08

I've written a book in German about the Boost C++ libraries and uploaded it to http://www.highscore.de/cpp/boost/. As this is the very first version it's currently only available in HTML. I would appreciate any feedback about the content, presentation, examples, typos :) etc. Thanks, Boris

5 7

Is there a patch for "thread copy constructor is accessible" BUG in solaris.
by Maruvekere, Praveen (Contractor) 19 Dec '08

19 Dec '08

Hi , I had posted this in build list now I am posting this in users of boost list. Please help thanks in advance. -------- I have done the temporary fix to build the boost 1.37.0 as told in one of the solaris user forum thread http://forums.sun.com/thread.jspa?threadID=5326640 <http://forums.sun.com/thread.jspa?threadID=5326640> for the below said error sun.compile.c++ bin.v2/libs/thread/build/sun-5.9/release/link-static/threading-multi/thr ead.o "./boost/thread/detail/thread.hpp", line 344: Error, inaccessible: boost::thread::thread(boost::thread&) is not accessible from boost::move(boost::detail::thread_move_t<boost::thread>). 1 Error(s) detected. Do I need a patch to fix this? -PM-

2 3

boost::interprocess, shared memory and multi-core
by QuitePlace 19 Dec '08

19 Dec '08

Shared memory or other constructs from boost::interprocess will be used on multi-core computers? Will it be a performance degradation if two processes are allocated to different cores versus one core? The same with threads? Another look at this question is - should one program the inter-process/inter- thread communication first and worry about multi-core later? Or something should be planned at the development stage?

5 6

Re: [Boost-users] Buffering and ASIO
by Vadim Shmelev 19 Dec '08

19 Dec '08

> Good day. > I'm trying to read video data transferred by HTTP. I use the following function in order > to read HTTP header: > boost::asio::async_read_until(m_socket, m_headerBuffer, "\r\n\r\n", boost::bind(&CAXISParser::read_header, this, boost::asio::placeholders::error, boost::asio::placeholders::bytes_transferred)); > , where m_headerBuffer is asio::streambuf object. During header parsing i can get video sample length. Now I want to read all video data in one function call, for example: > boost::asio::async_read(m_socket, boost::asio::buffer(pSample->GetBody(), m_currentSampleLength), boost::bind(&CAXISParser::read_body, this, pSample, boost::asio::placeholders::error, boost::asio::placeholders::bytes_transferred)); >, where pSample is object with buffer for video data. There is problem that async_read_until captures > some data from video data, and that call's behavior is incorrect. How can i fight that problem? For example, how can i figure out number of bytes left in streambuf and copy them to my buffer, and then read rest with async_read? I'm will very appreciate for any advices. > It's not clear to me exactly what you're asking here. From what I can > tell, you have created a buffer of size m_currentSampleLength, and > asked ASIO to call CAXISParser::read_body when it has received this > many bytes. Is that what you intended to do? Is that what boost is > doing? Or is the problem something else? > > -----Scott. Hello, Scott. Sorry for bad english, i'm will try explain. Yep, there is problem. When i'm read HTTP header with acync_read_until, it consume header and some bytes from video data (i think). After i'm calculate video data length from header and trying read data with async_read. But it works incorrect as previous call async_read_until captures some bytes from video data and video data readed with async_read have wrong offset. How can i understand, the problem that async_read_until and async_read used different buffers for reading (asio::streambuf and custom buffer), and some video data consumed by streambuf. Question: how to get video data captured with async_read_until and read rest from tcp buffer?There is is code sample:void read_header(PSample pSample, const boost::system::error_code& error, size_t bytes_transferred){ // Calculate data length with regex boost::asio::const_buffer data = m_headerBuffer.data(); const char* beg = boost::asio::buffer_cast<const char*>(data); const char* end = beg + bytes_transferred; CalculateDataLength(beg, end); // Clear buffer std::vector<char> vec(bytes_transferred); m_headerBuffer.sgetn(&vec[0], bytes_transferred); ReadBody();} Before ReadBody call, i reject streambuf content with sgetn. But i think there ismore data in streambuf that is video data. How can i get it? Below non working sample, but it describe my wishes: // Add before read bodysize_t restBytes = m_headerBuffer.size() - bytes_transferred;if (0 != restBytes){ m_headerBuffer.sgetn(pSample->GetBody(), restBytes);} Then in ReadBody i can to read rest video data with single async_readcall. There is implementation sample: void ReadBody(NMMSS::PSample& pSample){1)//boost::asio::async_read(m_socket, m_headerBuffer.prepare(m_currentSampleLength),2)//boost::asio::async_read_until(m_socket, m_headerBuffer, "\r\n",3)//boost::asio::async_read(m_socket, boost::asio::buffer(pSample->GetBody(), m_currentSampleLength), boost::bind(&CAXISParser::read_body, this, pSample, boost::asio::placeholders::error, boost::asio::placeholders::bytes_transferred));} I trying use 3 variants there:1) Trying read to buffer all video data (dont understand why it doesn't work)2) Very slow variant, as video data is big.3) Preferrable variant P.S. One question more: how can i correctly reject data consumed by streambuf? Now i used the following strategy: streambuf buf; ..... std::vector<char> vec(size); buf.sgetn(&vec[0], size); Is there another strategy for that? For example, commit and consume strategy? I dont need streambuf content in some cases. Best regards, Vadim

3 4