April 2007 - Boost-users - lists.preview.boost.org

FileSystem - Test if file exists without opening the file?
by Jeff 28 Apr '07

28 Apr '07

Hello, I can see from a FileSystem example how to check for the existence of a directory. What I would like to be able to do is to check for the existence of a file but I couldn't figure out how to do it using exists(). The way that I currently check for a file's existence is by opening the file. I wonder if boost::filesystem can test for a file's existence more efficiently than having to open it like I do in the snippet below. // checking for file existence by opening the file ifstream source(fileName); if (source) ... The reason for my concern is that I am going to be reading a text document containing a list of files, and all that I need to know is if the files exist on disk or not. If there is a way to check without opening each one, that would be ideal. Thanks in Advance

5 7

[Serialization] xml_archive + templates = :-(
by Sohail Somani 28 Apr '07

28 Apr '07

Hi, Quite sensibly, xml_archive seems to fail with archive errors when class_name contains xml characters (like <) which are quite common in templates. For example, in my tests: BOOST_EXPORT_CLASS(Something<T>); // fails when deserializing because of parse error. I fix it by doing: BOOST_EXPORT_CLASS_GUID(Something<T>,"Something_T"); // works \o/ Now, do you suppose xml_archive could be smart enough to convert < to < and similarly for other prohibited characters when serializing? Even if not, is there something I can do in 1.33.1 that would make it not so redundant? Thanks, Sohail

2 2

[program_options] assertion failed on very logn option description
by Dmitriy Chumack 27 Apr '07

27 Apr '07

hi * Win32 VisualStudio 2005 boost_1_33_1 When I supply option description, which length exceeded value of options_description::m_default_line_length, to the options_description, then I have an assertion failure when I want to get help from my command line program: prog --help Assertion failure: file: xstring line: 173 and them comes very long assert expression. This seems to be a compiler specific issue, because the same code, compiled on Linux Slackware 11.0 with gcc 3.4.6 and the same version of boost, works nice. thanks in advance -- Best regards, Dmitriy Chumack mailto:saint.d.a@gmail.com

2 1

[gil] top/bottom N% value
by Christian Henning 27 Apr '07

27 Apr '07

Hi guys, here is my first implementation of the bottom N% problem. The algorithm is my own invention so please let me know when there is a better way. Also, the current implementation only works for homogeneous pixel types. Any comments are more than welcome. Christian

1 0

read_graphviz question
by Jeffrey Holle 26 Apr '07

26 Apr '07

I'm using the post 1.31.1 version of boost, which has a much nicer read_graphviz function. I'm wondering if there is a way to "ignore" unwanted properties, something akin to sending them to /dev/null. My present experience is that if an unexpected property is found, an exception is thrown. Is there a way to express that a given property is to be ignored? Better yet, is there a way to say that all unexpected properties are to be ignored?

3 3

[multi_index] REVISITED: hash_unique with differently produced hash codes
by Ovanes Markarian 26 Apr '07

26 Apr '07

Hello Joaquín, now I found the failure, to the problem described in message http://lists.boost.org/boost-users/2007/04/27211.php with std::string as a key not being found. The problem in my case was that I use Xalan to parse XML and have to convert string to single byte characters. Since std::string has some build in caching to improve performance and I decode double byte string (XalanDOMString) as follows: std::string get_string(XalanDOMString const& str) { XalanDOMString::CharVectorType v; str.transcode(v); return std::string(v.begin(), v.end()); } this code produces a string which still points to the double byte buffer and therefore this results in a different hash value for the returned string as the value would be for the basic_string<char> or C-String: An example: XalanDOMString str("Sint8"); std::string char_str(get_string(str)); Calling find in the multi-index results in a hash value 40. Calling find in the multi-index for const char* str="Sint8"; results in a hash value 13 The same result (13) can be achieved when calling on char_str.c_str(); c_str() cause the strings internal buffer to be rebuild. Since boost::hash is on the way to become a part of STL it should ensure to work with the correctly intended copy of string. What is your opinion on this issue? With Kind Regards, Ovanes

3 3

[BGL] Basic graph construction
by Alejandro Marcos Aragón 26 Apr '07

26 Apr '07

I was wondering if someone could give me a tip on how to construct the following graph. The problem is that the queue type definition does not know what "edge" is until the type definition for the graph is created. Is there a way to get around this? Thank you #include <boost/graph/adjacency_list.hpp> #include <queue> // Type definition for a priority queue of (double,edge) pairs typedef std::priority_queue<std::pair<double,int> > edge_queue; struct vertex_edge_priority_t { typedef vertex_property_tag kind; }; // Vertex properties: typedef property < vertex_edge_priority_t, edge_queue > domainVertexProperty; // Edge propeties: typedef property < edge_index_t, int, property < edge_weight_t, double > > domainEdgeProperty; // Graph used to represent possible vertices and edges within the domain. typedef adjacency_list< setS, vecS, undirectedS, domainVertexProperty, domainEdgeProperty > domainGraph; typedef graph_traits<domainGraph>::edge_descriptor edge;

2 2

[smart_ptr] Conflict with Maya?
by Paul Giaccone 26 Apr '07

26 Apr '07

Does anyone know of any evidence that the memory management in smart_ptr might be incompatible with Maya, in particular, in gcc 3.2.3? My supervisor, who is unfamiliar with Boost's smart_ptr library, but is familiar with Maya, tells me that Maya has its own memory management, and that Boost's smart pointers might somehow interfere with this. I've used Boost's smart pointers for long enough that I really, really, really don't want to go back to the horrors of having to delete memory that I allocate on the heap and spending hours tracking down memory leaks if I get it wrong. I'd really rather not do that, especially as it's a potential memory leak that I'm trying to fix (Maya plug-ins are created using a bare "new" which doesn't seem to require a corresponding delete, according to the examples I have seen). So someone please tell me that is couldn't possibly be true. Thanks.

5 16

[scoped_array] "optimization"
by Michael Marcin 26 Apr '07

26 Apr '07

Hello, I have a new programmer on my team who has been brought on as an optimization engineer. She has a lot more experience than I do but I fear some of that experience may have left impressions that have stuck with her for too long. On the first check in labeled spot optimizations there were many things changed that disturbed me. One of them was changing a class from essentially. #include <boost/scoped_array.hpp> class Context { public: Context( unsigned int width, unsigned int height ) : m_buffer( new unsigned int[width*height] ) { } unsigned int* GetBuffer() { return m_buffer.get(); } private: boost::scoped_array<unsigned int> m_buffer; }; to class Context { public: Context( unsigned int width, unsigned int height ) : m_buffer( new unsigned int[width*height] ) { } ~Context() { delete m_buffer; } inline unsigned int* GetBuffer() { return m_buffer; } private: unsigned int* m_buffer; }; First of all there is the defect of delete being called instead of delete[]. Other than that I was under the impression that there should be no difference between the two other than the former being more descriptive (it instantly tells me this is an array and the memory is owned by the instance of this class). Is this really an optimization? Thanks, Michael Marcin

9 11

[multi-index] hashed_unique and sequenced question
by Ovanes Markarian 26 Apr '07

26 Apr '07

Hello, Is there any way to find an element in the multi-index container by it's hash and proceed with all elements which were added later on in the sequence after it? With Kind Regards, Ovanes Markarian

2 1