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

boost tokenizer
by hicks 17 Jan '02

17 Jan '02

Dear boost users I am trying to use the boost tokenizer with Borland Builder C++ 5 compiler. This far compiles fine: boost::tokenizer<> tok(std::string(ac0)); All further references to tok such as tok.assign(std::string(ac0)) or tok.begin() result in an error message " Structure requires on the left side of . or .*" Any advice would be welcome. Craig Hicks Engineer, KGK, Tokyo Japan

2 3

Boost Array Question
by Duane Murphy 16 Jan '02

16 Jan '02

Below is a simple example that I cannot get to compile. I would like to create a subclass of boost::array in order to give it some additional functionality (I would like a associative array like interface). When I compile this, I get an "illegal initialization" error for test. I am using Metrowerks Codewarrior 7 on Mac OS X. Is there some special incantation required? I suspect that I am in some kind of specialized area and I cant get there from here. template < class T, std::size_t N > struct test_array : public boost::array< T, N > { }; static boost::array< char,1 > array_test = {{ 'a' }}; static test_array< char, 1 > test = {{ 'a' }}; It appears that if I aggregate instead of inherit that this might work (at least it compiles :-) template < class T, std::size_t N > struct test_array { boost::array< T, N > array_; }; static boost::array< char,1 > array_test = {{ 'a' }}; static test_array< char, 1 > test = {{ 'a' }}; // The funny thing is this works also: static test_array< char, 1 > test2 = { 'a' }; TIA, ...Duane

2 6

tokenizer
by rradh 16 Jan '02

16 Jan '02

Hi, I have string line("a_1 = b_1 ( c_2 )") tokenizer< > tok(line); for (tokenizer<>::iterator beg=tok.begin(); beg!=tok.end();++beg) { cout << *beg << endl; } gives me a 1 b 1 c 2 I want to be able to specify delimiters like in strtok so that I get a_1, b_1 and c_2 but couldn't figure out how to specify the tokenizer function. Any suggestions? Regards Rajesh

3 3

graph: depth first search
by tobi_schuele 15 Jan '02

15 Jan '02

Hi, I'd like to use depth_first_search to check whether a particular vertex is reachable from the root vertex. Is there a way to stop the search when the vertex is found? In other words, can a visitor be used to cut the search tree? Sorry, if this is a newbie question. Thank you very much in advance for your help. Tobias

2 1

[newbie] segmentation fault on running dijkstra example on RedHat 7.2
by Mohammad M. Purbo 15 Jan '02

15 Jan '02

Dear All, I'm new to BOOST. I tried to compile the [BOOST_HOME/libs/graph/example/dijkstra-example.cpp] provided on the distribution with the following comand: gcc -I/home/purbo/include -lstdc++-3-libc6.2-2-2.10.0 -o dijkstra-example dijkstra-example.cpp no error on compiling/linking, but when i ran it, i got: Segmentation fault I used: RedHat 7.2 gcc 2.96 Anybody knows the problem? Thank you in advance, M. Purbo

1 0

A Static Map?
by Duane Murphy 14 Jan '02

14 Jan '02

I have learned a lot about templates and library programming from reading and using the Boost code. Recently I have been finding myself doing something that I would call Static Maps. Basically I would like to use an array of structs as if it were a map or associative array. Does boost have anything like this or can anyone recommend something like that? Here is a simple example: struct entry { const char* key; const char* value; }; const entry entries[] = { { "one", "Hello" }, { "two", "World" }, { "three", "!" } }; { some_type the_map = make_static_map( entries ); const char* word = the_map[ "one" ]; } This is a real poor example, but shows the idea. The struct could be any types that are less-than comparable; maybe there is some special arrangment of key and type. But the idea is there. I would like to be able to lookup items in a static array of items using array notation (other than numerically indexed). For extra credit, the interface would be writeable also; aka a fixed array space map. Any idea? Comments? Suggestions? ..Duane

2 2

Any interest in a weak pointer?
by wayneconrad 14 Jan '02

14 Jan '02

Is there any interest in a weak pointer that plays well with shared_ptr and its friends? Wayne Conrad

5 14

RE: [Boost-Users] Any interest for a parser class?
by Bjorn Karlsson 14 Jan '02

14 Jan '02

> From: Patrick Guio [mailto:patrick.guio@fys.uio.no] > > > Dear all, > I have written a parser class and I wanted to know whether it > could be of > any interest to be included in the boost system. Here is a > short description > of it. > This class is meant to be used in two ways, either to be > included in another > class as an inherited class and/or to be included in the main routine. > [snip] > > This is brief example to show how the class Parser can be used. > it can be used to handle more complex inherited classes. > If it can be of any interest, let me know and I will provide > you with more > details. > > Sincerely, > Patrick > I suggest that you also post this message to the Boost mailing list (http://groups.yahoo.com/group/Boost/) While Boost-Users certainly is a good place to get insightful opinions and reflections, the focus is to handle issues related to using existing Boost libraries. Also, library submissions and reviews are performed by the Boost list, so you'll need to go there sooner or later. Regards, Bjorn [Non-text portions of this message have been removed]

1 0

Any interest for a parser class?
by Patrick Guio 14 Jan '02

14 Jan '02

Dear all, I have written a parser class and I wanted to know whether it could be of any interest to be included in the boost system. Here is a short description of it. This class is meant to be used in two ways, either to be included in another class as an inherited class and/or to be included in the main routine. The first use is for classes containing fields to be initialised with some values. So far it handles intrinsic C++ types and STL vectors of intrinsics. As an example, the class Parser can be used in the class Foo as class Foo : public Parser { Foo::Foo(int nargs, char *args[]); private: enum { var1_enum, ... }; int var1; .... }; The Foo constructor that uses the Parser feature looks like Foo::Foo(nargs, args, ....) : Parser(nargs, args), var1(default_var1), ... { // Register class Parser::RegisterClass("Foo"); Parser::RegisterPackage(package, version, copyright); // Insert options and aliases to parse a value to var1 // the Any constructor is used to register the default value Parser::InsertOption(var1_enum,"var1","var1_type","var1_comment", Any(var1)); // Insert alias to option Parser::InsertOptionAlias(var1_enum, "var1_alias1"); Parser::InsertOptionAlias(var1_enum, "var1_alias2"); .... // Possibility to add a "debug level option" that takes an int argument, // the default value is 0 Parser::ParseLevelDebugOption("debug_level_option_name"); // the member function int Parser::DebugLevel() can be used thereafter to // check the this value if (Parser::DebugLevel() > 50) cout << "Debug level is " << Parser::DebugLevel() << endl; } Then a member function Foo::ParseParameters() method could look like void Foo::ParseParameters() { // The ParseOption family member functions return a boolean indicated // whether the var1 has been parsed if (Parser::ParseOption(var1_enum,var1)) cout << "var1 = " << var1 << " parsed." << endl; // Other ParseOption member functions are // * Parser::ParseOption(var1_enum,var1,"filename") // parse the filename "filename" instead of the command line // * Parser::ParseOptioninFileonCmdLine(var1_enum,var1) // if the option -i/--input[=]paramter_file is on the command line then // parse first the parameter_file, thereafter the command line. // * Parser::ParseOptiononCmdLineinFile(var1_enum,var1) // if the option -i/--input[=]paramter_file is on the command line then // parse first the command line, thereafter the parameter_file // * Parser::ParseHelp() // if the option -h/--help is parsed, print to standard output a message // with class_name information and a list of all the registered/inserted // options for that class in the predefined format : // option_name[,option_aliases] option_type option_comment // and return a true boolean // * Parser::ParseVersion() // if the option -v/--version is parsed, print to standard output a // message containing information package, version, copyright for the class // * Parser::ParseTemplate() // if the option --create-template is parsed, print to standard output a // message with class_name information and a list of all the // registered/inserted options for that class in a format to be used // later with the -i/--input containing information about variable and // aliases, type of the variable, comment/description and default value } The second way of using is from the main routine and is very similar to the first way int main(int nargs, char *args[]) { try { Parser parser(nargs, args); parser.RegisterProgram(args[0]); parser.RegisterPackage(package, version, copyright); Foo foo(nargs, args); // Insert options if necessary ... // Parse defaults if (parser.ParseHelp()) { foo.ParseHelp(); return EXIT_SUCCESS; } if (parser.ParseVersion()) { foo.ParseVersion(); return EXIT_SUCCESS; } if (parser.ParseTemplate()) { foo.ParseTemplate(); return EXIT_SUCCESS; } // Parse option for program if necessary ... // Parse class option foo.ParseParameters(); ... return EXIT_SUCCESS; } catch(ParseException &x) { cerr << x.what() << endl; return EXIT_FAILURE; } } The try/catch is used to handle potential errors at parsing and uses a class ParseException. This is brief example to show how the class Parser can be used. it can be used to handle more complex inherited classes. If it can be of any interest, let me know and I will provide you with more details. Sincerely, Patrick ====================================================================== Patrick Guio Institute of Physics, University of Oslo P.O. box 1048, Blindern, N-0316 Oslo Tel : (+47) 22 84 40 60 - Fax : (+47) 22 85 56 71 E-mail : patrick.guio(a)fys.uio.no URL : http://folk.uio.no/~patricg

3 2

iterator_adaptors Comment
by Duane Murphy 14 Jan '02

14 Jan '02

There is a comment in iterator_adaptors: public: // implementation details (too many compilers have trouble when these are private). Policies& policies() { return m_iter_p.second(); } const Policies& policies() const { return m_iter_p.second(); } Base& iter() { return m_iter_p.first(); } const Base& iter() const { return m_iter_p.first(); } This seems to imply that iter() and policies() ought to be private. I would like to suggest that having access to the underlying iterator is valuable. This is especially true of transform and projection iterators which allow you to locate a record by searching on a field. Access to the underlying iterator gets access to the rest of the record. Is there some other way of accomplishing the same thing without using iter()? Thanks, ...Duane

2 2