September 2009 - Boost-users - lists.preview.boost.org

boost::filesystem::rename() documentation incorrect
by Paul 25 Sep '09

25 Sep '09

Hi all, this documentation is incorrect: Otherwise, if to_p resolves to an existing file, it is removed. This is from Boost 1.40.0, and yet the code for rename() calls rename_api() which then does the following: if ( fs::exists( status_api( to, dummy ) ) ) return error_code( EEXIST, system_category ); ie, if to_p resolves to an existing file, it will always return/throw an error... NOT remove the existing file, as was documented. Or am I reading things wrong? cheers Paul

1 0

A problem on using boost timer (different time intervals were showed). Can any one with kindness help me?
by fmingu 25 Sep '09

25 Sep '09

I am using boost timer in C++ builder 2009 to test the time used in my program.there are the lines: .................... boost::timer t; double timeresult; ................... t.restart(); ........... //my program main body timeresult=t.elapsed(); ShowMessage( " the time consumed by my program is "+FloatToStr(timeresult)); ......................... the program was compiled and linked and worked well. But I find a problem: every time the message showed is not the same time interval, sometimes it is 0.015 and sometimes it is 0.016 and sometimes it is 0. I do not know how it was happened. Does it happen due to the multi-process and multi-thread mode by windows? I do not know how to solve it. My program is quite strict on the time consumed by the computer. Can any one with kindness help me?

7 8

[lambda] Defining a predicate for std::find_if if the key to a std::map is std::pair
by Ryan McConnehey 25 Sep '09

25 Sep '09

I have the following std::map signature. typedef std::pair<std::string, int> cPair; typedef std::map<cPair, std::vector<classA> > cMap; cMap myMap; When entering values into the map, the std::string is the only part of the key that needs to be unique. Upon retrieving map values both values of the key need to be compared. I've tried to define my insert comparison as follows, but am uncertain how to define the predicate with boost::lambda. std::string givenName = "some_text_value"; int givenInt = 3; cMap::const_iterator itor = std::find_if( myMap.begin(), myMap.end(), boost::lambda::bind( ??? , _1) == givenName ); What goes at the question marks? Ryan

5 12

[spirit] parsing vsv files with embedded quotes/line breaks
by Sean Farrow 25 Sep '09

25 Sep '09

Hi: I know there is a spirit example to parse csv or any list-separated file. What I need to do is to be able to parse files with embedded "," characters, quotes and line breaks in. How would i go about doing this? Regards Sean.

2 1

[MPL] at<> function with too many arguments
by Ryan McConnehey 25 Sep '09

25 Sep '09

I have a mpl::map as follows, with a template parameter of a passed in type and a template parameter of "supported_type" that is a class. using boost::mpl; enum { value = 1 + supported_type::value }; typedef typename insert<typename supported_type::type_map, pair<T, int_<value> > >::type type_map; I'm trying to create a function that searches the map for a given type and returns a numeric value if it's found or -1 if it's not found. template <typename S> int typeValue(void) { return at< type_map, S, int_<-1> >::type::value; } The documentation says I can specify a default value for "at" that is returned if the given key isn't found. The compiler, though, says that I've provided too many arguments. Have I called "at" incorrectly? Is the documentation incorrect? Is the fault with the compiler (VS 2005)? Ryan

2 1

[python] Ownership of python objects.
by Germán Diago 24 Sep '09

24 Sep '09

Hello. I'm trying to do the following thing: 1.- Create an object 2.- Put it in an scene (for a game) 3.- To be able to delete it from the scene (this is done from C++) My problem is that if I do the following: p = Object("blabla") Scene.addToScene("myobject", p) //Here, the object is deleted by C++, which is causing problems p.removeFromScene("myobject") I know a workaround for this. It would be to make a wrapper for objects created from python and not deleting the object if it was created from there. But I would like if there is some other (more general) way to avoid this. Thanks in advance.

2 1

[MultiIndex] documentation for macros
by Igor R 24 Sep '09

24 Sep '09

Hello, Is there any description of the various macros that control MI behavior, like BOOST_MULTI_INDEX_ENABLE_SAFE_MODE etc.? I tried to search through the library reference, but haven't managed to find any. Thanks.

3 4

Re: [Boost-users] [Regex] Group of number at line end
by Roland Bock 24 Sep '09

24 Sep '09

Reiner Hutsch wrote: > -----Ursprüngliche Nachricht----- > Von: boost-users-bounces(a)lists.boost.org > [mailto:boost-users-bounces@lists.boost.org] Im Auftrag von John Maddock > Gesendet: Mittwoch, 23. September 2009 18:45 > An: boost-users(a)lists.boost.org > Betreff: Re: [Boost-users] [Regex] Group of number at line end > >> i am looking for a regular expression which extracts a group of number >> from >> the end of a line like the following: >> >> "This line may contain digits (9) but ends with numbers like 1 2 3 4 5" >> >> When using a regex like "^(.+)(\\s+(\\d+))+$" I receive >> "This line may contain digits (9) but ends with numbers like 1 2 3 4" and >> "5". >> >> Can someone help me? > > What about: "^(.+)[^\\d\\s](\\s+\\d+)+$" > > HTH, John. > > Hi John, > > sorry, but this gives also only the last number. > > Reiner. If you want the group, not the each number separately, this should work: ^.*[^ 0-9] *([ 0-9]+)$ Testing with perl: perl -e '$x = "This line may contain digits (9) but ends with numbers like 1 2 3 4 5"; $x =~ s/^.*[^ 0-9] *([ 0-9]+)$/$1/g; print "$x\n";' Output: 1 2 3 4 5 Regards, Roland

1 0

redirect flow with Breadth-First-Search - problems with the residual capacities
by Benjamin Rupp 24 Sep '09

24 Sep '09

Hi, I want to redirect Flow, which I have earlier calculated with Edmonds-Karp, with Breadth-First-Search (Aim is to determine the maximal free capacity of an edge by redirecting all possible flow to other edges) When I call Edmonds-Karp the Maximal Flow is calculated. If I then print out the values of the residual capacities of the edges, all edges with a capacity of 0 (this are the reverse edges) have a residual capacity of 0 instead of a capacity equal to the flow of the reverse edge. I wrote the following code to correct the values: long speicher[num_edges(g)]; for(tie(e_iter,e_end)=edges(g); e_iter!=e_end; ++e_iter) if(get(e_index,*e_iter) < num_edges(g)/2){ speicher[e_index[*e_iter]] = residual_capacity[*e_iter]; speicher[e_index[*e_iter]+(num_edges(g)/2)] = flow[*e_iter]; } for(tie(e_iter,e_end)=edges(g); e_iter!=e_end; ++e_iter) put(residual_capacity,*e_iter,speicher[e_index[*e_iter]]); But this is very time inefficient. Does anyone have an idea why the reverse edges do not have the correct residual capacity? Another strange phenomenon is that if I allocate the correct residual capacity only to the reverse edges because the others are correct this change to the residual capacities is not saved. for(tie(e_iter,e_end)=edges(g); e_iter!=e_end; ++e_iter) if(get(e_index,*e_iter) < num_edges(g)/2){ put(residual_capacity,reverse_edge[*e_iter],flow[*e_iter]); Thanks Benjamin

1 0

[bind/function] memory allocation question
by Tim Blechmann 24 Sep '09

24 Sep '09

hi all, i would like to use boost.bind and boost.function in a real-time application, where heap memory allocation is not allowed. according to the docs, boost.function may allocate memory on the heap, but uses a small object optimization, to avoid heap allocation. i'd be curious, if there is a way to figure out, whether heap allocation may occur or not ... i would like to be able to use boost::function objects, initialized by boost::bind, something like: void callme (int some, int args) {} boost::function<void (void)> fn = boost::bind(callme, 1, 2); fn(); thanks, tim -- tim(a)klingt.org http://tim.klingt.org Which is more musical, a truck passing by a factory or a truck passing by a music school? John Cage

1 0