- Boost-users - lists.boost.org

[iostreams] Inconsistent state of stream
by Nicola 03 May '10

03 May '10

Hi, can someone tell me if the behaviour of the enclosed example is a bug? In my test, the last two assertions fail, but I don't understand why (Boost 1.42.0). Thanks, Nicola #include <boost/iostreams/copy.hpp> #include <boost/iostreams/filtering_stream.hpp> #include <boost/iostreams/filter/bzip2.hpp> #include <cassert> #include <fstream> #include <sstream> struct bz2istream : public boost::iostreams::filtering_stream<boost::iostreams::input> { public: bz2istream(std::string const& path) : pBz2file(path.c_str(), std::ios_base::in | std::ios_base::binary) { push(boost::iostreams::bzip2_decompressor()); push(pBz2file); } private: std::ifstream pBz2file; }; int main() { bz2istream bzippedfile("foo.bz2"); std::ostringstream os; boost::iostreams::copy(bzippedfile, os); assert(!(bzippedfile.rdstate() & std::ios_base::goodbit)); // Ok assert(bzippedfile.rdstate() & std::ios_base::eofbit); // Fails assert(!(bzippedfile.good())); // Fails }

2 1

[UTF] Let my debugger catch unhandled exceptions
by Adam Nielsen 03 May '10

03 May '10

Hi all, I'm trying to track down a problem in my code that the Boost test framework has isolated, however I'm finding it difficult to track it back to a line in the source. The Boost test framework likes to catch all exceptions, which means when you run the code through a debugger it doesn't break at the point the exception becomes unhandled, so you can't do a backtrace to see where it came from. There is the option --catch_system_errors=no which works for some exceptions, but not these ones. I'm using gdb and telling it to "catch throw" is painful because there are dozens of exceptions thrown around correctly before it gets to the unhandled one I'm interested in. Is there some easy way I can tell UTF not to catch any exceptions? Many thanks, Adam.

3 4

[spirit] Parsing a value multiple times
by Ravi 03 May '10

03 May '10

Hi, Is there a standard qi paradigm for parsing a value multiple times? Suppose a configuration file has three parameters "a", "b" and "c" each of which take integers. Here's how to parse them in arbitrary order: struct abc { int a; int b; int c; }; .. // Inside the qi grammar with attribute abc start = +( (lit("a") >> lit("=") >> int_[bind(&abc::a,_val)=_1] >> lit(";")) |(lit("b") >> lit("=") >> int_[bind(&abc::b,_val)=_1] >> lit(";")) |(lit("c") >> lit("=") >> int_[bind(&abc::c,_val)=_1] >> lit(";")) ); With the grammar above, the final value of "a" comes from the the last line in the configuration file of the form "a = 23;" overwriting all previous statements. This (later values overriding earlier ones) is a fundamental requirement for this application. Is there a better way to write the rule above? Can semantic actions be avoided (perhaps by use of FUSION_ADAPT_STRUCT)? As far as I can tell, using the "^" operator requires that the members of struct abc be boost::optionals. This grammar does not need optionals because the instance of struct abc that is passed to the parsing function is already initialized with sane defaults. Second, composing grammars of this form does not seem to scale well. For example, suppose the configuration above is extended by one parameter, "d", for a different "composite" configuration file format. The easiest way to incorporate the new parameter is to copy and paste the old rule: struct abcd : abc { int d; }; .. // Inside the qi grammar with attribute abcd start = +( (lit("a") >> lit("=") >> int_[bind(&abcd::a,_val)=_1] >> lit(";")) |(lit("b") >> lit("=") >> int_[bind(&abcd::b,_val)=_1] >> lit(";")) |(lit("c") >> lit("=") >> int_[bind(&abcd::c,_val)=_1] >> lit(";")) |(lit("d") >> lit("=") >> int_[bind(&abcd::d,_val)=_1] >> lit(";")) ); The solution above is a clear DRY violation. Is there a solution that avoids copy-and-paste? Ravi

1 0

[mpl] merging 2 maps
by Igor R 02 May '10

02 May '10

Hello, typedef mpl::map < mpl::pair<K1, V1>, mpl::pair<K2, V2> > m1; typedef mpl::map < mpl::pair<K1, V3>, mpl::pair<K3, V4> > m2; In the current version of boost, what's the right way to merge the 2 above mpl maps - to get 1 map that contains: mpl::pair<K1, V1>, mpl::pair<K1, V3>, mpl::pair<K2, V2> mpl::pair<K3, V4> Thanks.

2 3

[date_time] Decoding ISO 8601 strings with timezones?
by Allen Cronce 01 May '10

01 May '10

Hi all, I'm trying to decode ISO strings into ptime values. But when I initialize what should be two equivalent ptimes, one in UTC and the other with a timezone designation, the resulting ptimes do not compare. It's as if the ISO extended format input facet is ignoring timezone values. I'm using boost 1.42.0. I tried instancing an input facet with "%Y-%m-%dT%H:%M:%S%F%Q" instead of calling set_iso_extended_format, but this didn't work either. Can someone suggest a solution? Also, all the sample code I've seen in this area allocates a time_input_facet but does not delete it. Does the stream take care of deallocating the input facet when it goes out of scope, or will this result in a memory leak? Here's the code and resulting assert output: --snip-- ptime from_iso_extended_string(const string& inDateString) { // If we get passed a zero length string, then return the "not a date" value. if (inDateString.empty()) { return not_a_date_time; } // Use the ISO extended input facet to interpret the string. std::stringstream ss; time_input_facet* input_facet = new time_input_facet(); input_facet->set_iso_extended_format(); ss.imbue(std::locale(ss.getloc(), input_facet)); ss.str(inDateString); ptime timeFromString; ss >> timeFromString; return timeFromString; } void boostTimeTest(void) { cout << "Test decoding an ISO 8601 string with different timezones for equivalency:" << endl; string utcTimeStr("2010-05-01T18:00:00.000Z"); string localTimeStr("2010-05-01T11:00:00.000-07:00"); ptime utcTime = from_iso_extended_string(utcTimeStr); ptime localTime = from_iso_extended_string(localTimeStr); cout << " Original utc string: " << utcTimeStr << endl; cout << " Decoded utc string: " << to_iso_extended_string(utcTime) << endl; cout << " Original local string: " << localTimeStr << endl; cout << " Decoded local string: " << to_iso_extended_string(localTime) << endl; // The following asserts because the resulting ptimes are not equivalent assert(utcTime == localTime); } -------------------------------------------- Test decoding an ISO 8601 string with different timezones for equivalency: Original utc string: 2010-05-01T18:00:00.000Z Decoded utc string: 2010-05-01T18:00:00 Original local string: 2010-05-01T11:00:00.000-07:00 Decoded local string: 2010-05-01T11:00:00 Assertion failed: (utcTime == localTime), function boostTimeTest --snip-- Best, -- Allen Cronce

1 0

[serialization] void_caster fix ...
by Bogdan 01 May '10

01 May '10

Just want to confirm I tested the change (61558) made by Robert Ramey to void_caster and it fixes the crashes I previously experienced. Thank you Robert!

2 1

How do you use the find_iterator to do in-place replacements?
by G S 01 May '10

01 May '10

Hi all. I have the split_iterator working as intended on a string, but I don't see how to use the range it returns to change the string directly without copying it. For example, I want to do an erase_all on a range, but that function won't take a range. Anybody have an example of how to do this? Thanks!

2 6

[Dynamic Bitset] Comparableness between different sized sets
by Etienne Philip Pretorius 01 May '10

01 May '10

Hello List, Just a quick question, why can you not compare between different sized dynamic bitsets? As I am trying to add some operators to it in my application so that it can become a large interger type.... (ie in excess of 256 bits) boost::dynamic_bitset<> operator++( boost::dynamic_bitset<>& x, int ) { bool c = true; for(boost::dynamic_bitset<>::size_type i = 0; c && (i < x.size()); i++) { if(x[i]) { c = true; }else{ c = false; } x[i].flip(); } if(c) { x.resize(x.size()+1,true); } return x; }; If I have a for loop and the post incremental operation increases the size of the bitset then operator < bails with an assertion failure : /usr/include/boost/dynamic_bitset/dynamic_bitset.hpp:1288: bool boost::operator<(const boost::dynamic_bitset<Block, Allocator>&, const boost::dynamic_bitset<Block, Allocator>&) [with Block = long unsigned int, Allocator = std::allocator<long unsigned int>]: Assertion `a.size() == b.size()' failed. Aborted Etienne -- This message has been scanned for viruses and dangerous content by Pinpoint, and is believed to be clean.

1 0

Boost Graph: determining type of bundled property
by Trevor Harmon 30 Apr '10

30 Apr '10

Hi, I wrote some Boost Graph code that uses a custom pointer type as the vertex: adjacency_list<vecS, vecS, directedS, Foo*, property<edge_index_t, std::size_t> > I now need to expand the graph so that the Foo pointer is just one of several vertex properties. I assume bundled properties are the way to go: http://www.boost.org/doc/libs/1_42_0/libs/graph/doc/bundles.html I also need my vertices to be of different types (but all as children of a base type), such that my vertex bundles look something like this: class BaseVertex { }; class EntryVertex : public BaseVertex { }; class ExitVertex : public BaseVertex { }; class FooVertex : public BaseVertex { Foo *foo; // ... other properties follow ... }; My goal is to be able to iterate through the vertices and perform different actions depending on the type of vertex. (e.g., if vertex type is EntryVertex do this, if vertex type is ExitVertex do that, etc.) However, I don't know how to determine a vertex's type. The bundles document mentions something about vertex_bundle_type<Graph>::type, but I have no idea how to use this. I couldn't find any code in the Boost examples that use it, either. Any tips to get me started? Also, note that I've added an index property to my graph's edges, declared explicitly in the adjacency_list instantiation (as seen above). Would it make sense to convert this to a bundled property as well, or should I leave it as-is? (I don't think I'll ever need any edge properties other than the index.) Thanks, Trevor

3 6

handling of daylight saving time...
by Archie14 30 Apr '10

30 Apr '10

My software records time events on the computer and it does it using whatever timezone and daylight saving time settings are on this computer. I realized too late that, while storing timestamps, I am not accounting for this daylight adjustment process. So, if hypothetically speaking, the daylight adjustment happens today at 9:00 am and I am writing my timestamps every second I will have something like that: … 8.58 8.59 9.00 10.01 (without timesaving should be 9.01) 10.02 (without timesaving should be 9.02) … When I am reading my timestamps back I am constructing timezone and using it for timestamps manipulation std::string gmt = (boost::format("TMZ%1% TMS1:00:00,M3.2.0/02:00:00,M11.1.0/02:00:00") % gmtoffset).str(); , but this one-hour gap persists, as I now think it should. The question is: is it possible to do manipulation with datetime library to remove this gap without hard-coding one hour removal after specific date/time of daylight saving time?

3 4