July 2004 - Boost-users - lists.preview.boost.org

C++ Template Metaprogramming
by David Abrahams 25 Jul '04

25 Jul '04

For the past year Aleksey Gurtovoy and I have been working on a new book, titled _C++ Template Metaprogramming: Concepts, Tools and Techniques from Boost and Beyond_. The book has recently entered copyediting. If you're interested you can peruse some sample chapters at http://boost-consulting.com/mplbook. Regards, Dave -- Dave Abrahams Boost Consulting http://www.boost-consulting.com

5 10

Building boost into a single dll
by boost＠dalvander.com 24 Jul '04

24 Jul '04

Is it possible to build all (or some specified) boost libraries into one single dll? // Anders Dalvander

1 0

Newbie -- Regex, Unicode, XCode and MacOSX
by Steve Moore 23 Jul '04

23 Jul '04

Hello. I'm trying to incorporate Unicode regex searching in a little app I'm doing. I'm building using XCode 1.2 on MacOSX 10.3.4. The problem I'm having is that I can't seem to get boost's regex to incorporate unicode or wide-character support. In <boost/config/platform/macos.hpp> there is the line #define BOOST_NO_CWCHAR which seems to disable allowance for unicode by causing boost::wregex and boost::wcmatch symbols to not be typedef-ed. A test app. shows that these symbols are defined when I build using Codewarrior. So I'm wondering how to get unicode support for regex using XCode (gcc) on a MacOSX machine. Any suggestions? Thanks 10^6, STeve Moore.

2 3

Don't seed() mersenne_twister RNG with 0
by Benoît SIBAUD 23 Jul '04

23 Jul '04

Hi, a seed(0) with a mersenne_twister (like mt11213b or mt19937) aborts on some assertion (tested on boost 1.31). boost/random/linear_congruential.hpp:61: boost::random::linear_congruential<IntType, a, c, m, val>::linear_congruential(IntType) [with IntType = uint32_t, IntType a = 69069, IntType c = 0, IntType m = 0, IntType val = 0]: Assertion `c || x0' failed. mersenne_twister::seed(x) calls linear_congruential<uint32_t, 69069, 0, 0, /* unknown */ 0>(x), which asserts(0 || x). So x must be non-zero. I didn't see it in the doc, but I think it would be a good thing to have it written down, or to have mersenne_twister::seed() modified to avoid the assertion. Regards, -- Benoît Sibaud

1 1

filesystem and existing files
by Russell Hind 23 Jul '04

23 Jul '04

With filesystem, you can do filesystem::path P(....); filesystem::ofstream S; if (!filesystem::exists(P)) { S.open(P); } But isn't there a race condition here which could lead to a file created by another process being wiped out, or added to? Is there a way to get ofstream::open to fail if the file already exists? It succeeds on VC7.1 but in some situations I don't want it to succeed. Cheers Russell

2 4

Looking for some help with boost::thread::sleep()
by Kutch, Patrick G 23 Jul '04

23 Jul '04

I am used to putting in quick sleep() statements in my worker threads, usually a sleep(0) just to give up a timeslice, or maybe a sleep(1). Of course boost::thread::sleep takes an xtime object not a milliseconds value, so I wrote a small routine to simulate the sleep() statement using boost::thread::sleep: void inline UtilitySleep(long milliSeconds) { const int NANOSECONDS_PER_MILLISECOND = 1000000; if (0 == milliSeconds) { boost::thread::yield(); // just do a yield return; } boost::xtime time; int sec = 0; // if larger than 1 sec, do some voodoo for the boost::xtime struct if (milliSeconds >= 1000) { // convert ms > 1000 into secs + remaining ms int secs = milliSeconds /1000; milliSeconds = milliSeconds - secs*1000; sec += secs; } milliSeconds*=NANOSECONDS_PER_MILLISECOND; // normal boost time sleep stuff // get current time boost::xtime_get(&time,boost::TIME_UTC); // add # of desired secs and ms time.nsec+=milliSeconds; time.sec+= sec; // sleep until that time boost::thread::sleep(time); } It seems to always work fine unde Windows, however occasionally under Linux it appears that if I use a small value, such as sleep(2) that sometimes it takes longer to setup the values and call boost::thread:sleep() and I end up passing an xtime value that has already passed, basically blocking the thread forever. Anybody have any advice for me? Thanx! - Patrick

1 0

Peculiar uBLAS compile error...
by Nicholas Andrews 23 Jul '04

23 Jul '04

Hi, I have the following c++ function: void multi_state_kf::correction(const vector<double>& obs, const matrix<double>& dev) { compute_gain(dev); vector<double> temp = (m*obs-newX*m); x=newX+prod(temp,K); matrix<double> T=identity_matrix<double>(3,3)-K*m; matrix<double> temp2=prod(T,newP); matrix<double> temp3=prod(K,dev); P=prod(temp2,trans(T))+prod(temp3,trans(K)); } The syntax is completely hacked together as this is my first attempt at using boost's uBLAS, but it got rid of most of my errors except this one: \boost\numeric\ublas\functional.hpp(190): error C2440: '=' : cannot convert from 'const boost::numeric::ublas::matrix_binary_scalar2<E1,E2,F>::value_type' to 'double' (...) The odd (for me) part is that even when I remove all boost function calls from the function the compile error still shows up and it points me to a line in the functional.hpp file. So 1. Does anyone know what could be causing this compile error? 2. What's a better way of writing this function? Thanks in advance, Nick

1 0

test failure rate and usage
by JHAYNBERG＠cooperneff.com 22 Jul '04

22 Jul '04

I wanted to use the Boost smart_ptr classes on Solaris 8/Workshop 7 and RedHat E3/gcc 3.23. So, I downloaded the boost src (and jam and boost.build) and built the entire package. After seeing many test failures, I discovered on the regression test page that no compiler achieves 100% pass rate and some are about 30%. I have 2 questions. One, in general, how does one know when it is "safe" to use a boost library on a particular platform? For example, if a platform's failure rate is 15%, should boost be used? And, two, with just the boost source (without building any libraries), I was able to use the boost::shared_ptr in my Solaris application by merely including the right header. Apparently, I don't need to build boost to use shared_ptr which leads me to ask, in general, how do one know when a library is needed (or not needed)? Shouldn't the fact that boost is mainly templates mean there should be little or no implementation (i.e. code that needs to be compiled into a library)? For example, many template libraries are provided as header-only solutions. Thanks! Jay Haynberg Software Engineer This message and any attachments (the "message") is intended solely for the addressees and is confidential. If you receive this message in error, please delete it and immediately notify the sender. Any use not in accord with its purpose, any dissemination or disclosure, either whole or partial, is prohibited except formal approval. The internet can not guarantee the integrity of this message. BNP PARIBAS (and its subsidiaries) shall (will) not therefore be liable for the message if modified. --------------------------------------------- Ce message et toutes les pieces jointes (ci-apres le "message") sont etablis a l'intention exclusive de ses destinataires et sont confidentiels. Si vous recevez ce message par erreur, merci de le detruire et d'en avertir immediatement l'expediteur. Toute utilisation de ce message non conforme a sa destination, toute diffusion ou toute publication, totale ou partielle, est interdite, sauf autorisation expresse. L'internet ne permettant pas d'assurer l'integrite de ce message, BNP PARIBAS (et ses filiales) decline(nt) toute responsabilite au titre de ce message, dans l'hypothese ou il aurait ete modifie.

2 1

Please help me understand boost:lambda
by TP Diffenbach 21 Jul '04

21 Jul '04

I'm having some difficulties figuring out how to use boost::bind and boost::lambda Here's what I've trying to do: use transform to transform the values in one container to another, looking up the new value as a function of the old. What I'm trying to do is conceptually simple: call std::transform with a predicate that takes the original value, does a std::find_if on that value in another "container" using a bound std::greater< int > predicate, converts the returned iterator to an integral distance, and inserts that integral distance in the container that is the destination of the transform. Doing this using a hand-coded functor is easy enough, but I'm not able to figure out how to do it using boost::lambda. If I understand lambda, this should make each element of c equal to four times the corresponding element of o, and indeed this compiles as works as expected. std::transform( o.begin(), o.end(), std::back_inserter( c ), _1 = _1 * 4 ) ; This bit of code will look up a value, the index of the first element in partition with a value greater than v (or 3 if v is greater than all values in partition: int partition[] = { 10, 20, 30 } ; offset = std::find_if( partition, partition + 3, boost::bind( std::greater< int >(), _1, v ) ) - partition ; This code wraps the lookup and the subtraction that turns the returned pointer into an offset: template< template< typename > class Pred, typename It, typename T > int find_region_offset( It b, It e, T val ) { return std::find_if( b, e, std::bind2nd( Pred< T >(), val ) ) - b ; } and this code uses the wrapper, along with the transform: std::transform( o.begin(), o.end(), std::back_inserter( c ), boost::bind( find_region_offset< std::greater, int*, int>, partition, partition + 3, _1 ) ) ; So far, so good. What I don't know how to do is to dispense with the wrapper. As close as I can come is to use this class: struct identityPred { int operator()( int i ) { return i ; } } ; to do this: std::transform( o.begin(), o.end(), std::back_inserter( c ), boost::bind( std::less< int* >(), // to subtract the returned pointer from partition boost::bind( std::find_if< int*, identityPred >, partition, partition + 3, identityPred() ), partition) ) ; This compiles, but I don't want identityPred, I want boost::bind( std::greater< int >(), _1, _1X ) where _1 is the value in the partition array and _1X is the value read from container o. Ideally I'd like to do this using boost::lambda, which I think would look something like: std::transform( o.begin(), o.end(), std::back_inserter( c ), _1 = find_if(partition, partition + 3, _1 > var( _1 )) - partition ) ; where the first and last "_1"s are transform's (containers o and c's) values, and middle one is partition's values. Again, what I'm trying to do is conceptually simple: call transform with a predicate that takes the original value, does a find_if on that value in another "container" using a bound std::greater< int > predicate, converts the returned iterator to an integral distance, and inserts that integral distance in the container that is the destination of the transform. So if someone could point out to me what I'm doing wrong, or point me to more examples of using boost::lambda than are contained in the boost::lambda documentation, or could just tell me that what I'm trying to do isn't possible (and why it's not possible) I'd very much appreciate it. Thanks. PS And a technical trouble: if I include both boost::bind and boost::lambda, and I am using namespace boost::lambda to save qualifying each _1, I get a type redefinition for _1, with one definition in boost::bind, when using gcc under cygwin: b.cpp:49: error: use of `_1' is ambiguous /usr/local/include/boost/lambda/core.hpp:74: error: first declared as `const boost::lambda::placeholder1_type&boost::lambda::<unnamed>::_1' here /usr/local/include/boost/bind/placeholders.hpp:55: error: also declared as ` boost::arg<1> <unnamed>::_1' here b.cpp:57: error: use of `_1' is ambiguous /usr/local/include/boost/lambda/core.hpp:74: error: first declared as `const boost::lambda::placeholder1_type&boost::lambda::<unnamed>::_1' here /usr/local/include/boost/bind/placeholders.hpp:55: error: also declared as ` boost::arg<1> <unnamed>::_1' here Should this be happening?

1 0

date-time
by Jonathan Oulds 21 Jul '04

21 Jul '04

Could anyone explain to me why there is not a default constructor for boost::gregorian::date_time and boost::posix_time::ptime. I implemented a simple constructor where both classes are created as 'not_a_date_time'. The only problem I have had is compiling my code on other peoples machines that do not have my modification to the date_time library. jonathan

7 19