October 2006 - Boost-users - lists.boost.org

Memory problem using shared_ptr w/ STL-containers, gcc
by Robert Fendt 20 Oct '06

20 Oct '06

Hi, I have a somewhat peculiar problem concerning the shared_ptr template. I use an STL container to hold several shared_ptr values. When the container is destroyed, all the destructors are called correctly, but the memory is not freed. Furthermore, shared_ptr usage introduces a 100x memory overhead (or more) in my case. Compiler system is GCC. Consider the following short example: #include <iostream> #include <string> #include <deque> #include <boost/shared_ptr.hpp> class A { public: int a; int b; }; int main() { std::string tmp; std::cout << "Enter something to continue" << std::endl; std::cin >> tmp; { std::deque<boost::shared_ptr<A> > list; for (int i = 0; i < 10; i++) { boost::shared_ptr<A> ptr(new A()); list.push_back(ptr); } std::cout << "Enter something to continue" << std::endl; std::cin >> tmp; } std::cout << "Enter something to continue" << std::endl; std::cin >> tmp; } Between the first and the second stop, the program size in memory grows by about 150k... and that's for 10 objects which are <20 bytes large, individually. And to top it off, it leaks memory. Between stops two and three the memory *should* be freed since the scope is left. I can include a destructor doing debug output in the class which shows that the objects are indeed destroyed. But the memory stays allocated. I suspect this is a GCC problem, but I have actually no clue... I tested it on a recent Debian "Unstable" and an Ubuntu "Breezy". GCC versions varying from 3.2 to 4.1. Or do I use the template in a wrong way? Suggestions would be very welcome... and if this is a known bug in GCC, a feasible work-around would be great. Thanks in advance, Robert

4 5

depth_first_search compilation problem
by Saptarshi Ghosh 19 Oct '06

19 Oct '06

Dear all, I am facing a problem with boost::depth_first_search. I couldn't compile the example http://boost.org/libs/graph/example/dfs-example.cpp after changing line no 42: typedef adjacency_list < listS, vecS, directedS > graph_t; to: typedef adjacency_list < listS, listS, directedS > graph_t; I need to use < listS, listS, directedS > because of performance issue. I have used gcc 3.2.2 and gcc 4.1.0 without any success. Any suggestion how to solve the problem ? regards, Saptarshi __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

2 1

Re: [Boost-users] Support for VS 2005 C++/CLI
by John Dunn 19 Oct '06

19 Oct '06

Peter Dimov wrote : > Does it fail with both the static and the dynamic VC runtime? I don't think you can use the static runtime with CLI. When I try, I get the following error - cl: Command line error D8016 : '/MTd' and '/clr' command-line options are incompatible Both the release and debug versions fail in the same manner.

3 2

Re: [Boost-users] Support for VS 2005 C++/CLI (Peter Dimov)
by John Dunn 19 Oct '06

19 Oct '06

Peter Dimov wrote : > Did you rebuild the Boost.Thread library with the new SP? You probably > still have a dependency on the old VC runtime in it if you didn't, > whereas the application now depends on the new runtime. Yes, I did rebuild boost after installing the new SP. I couldn't figure out how to get bjam to clean boost so I just deleted the boost_src/bin directory and re-installed. The .lib files all have new datestamps so I believe they were rebuilt correctly.

2 1

Re: [Boost-users] Boost-users Digest, Vol 1062, Issue 2
by james.jones＠firstinvestors.com 19 Oct '06

19 Oct '06

From: Arnaldur Gylfason <arnaldur.gylfason(a)decode.is> > Let's say I have to parse "(_1 + _2)*_1" and "(_1/_2)*_1". I only get > these command strings at runtime so I don't know the expressions at > compile time. I don't think there's any hope of making it work this way. Ultimately you really need to use run-time expression structures if you're building expressions at run-time. This is not hard to do, however. My advice would be to use boost::variant. - James Jones Administrative Data Mgmt. Webmaster 375 Raritan Center Pkwy, Suite A Data Architect Edison, NJ 08837

1 0

Re: [Boost-users] enable_shared_from_this & inheritance
by Nat Goodspeed 19 Oct '06

19 Oct '06

> -----Original Message----- > From: boost-users-bounces(a)lists.boost.org [mailto:boost-users- > bounces(a)lists.boost.org] On Behalf Of Peter Dimov > Sent: Thursday, October 19, 2006 12:39 PM > To: boost-users(a)lists.boost.org > Subject: Re: [Boost-users] enable_shared_from_this & inheritance > > Peter Dimov wrote: > > > I have no idea why this compiles. It shouldn't, B has two > > enable_shared_from_this<> bases, and the constructor will only > > initialize one of them. Only one of FooA and FooB works (FooA in this > > case). I'll investigate. > > The code compiles on g++ and como, but neither FooA nor FooB works. Looks > like an MSVC (and Borland) bug. It would be nice if I could find a way to > make it fail at compile time instead of not working, though... [Nat] <silly question> Could the ref count be put into a virtual base class so that only one instance would appear in a given leaf class, regardless of how many enable_shared_from_this<> bases it has? </silly question>

2 1

enable_shared_from_this & inheritance
by Alex Ott 19 Oct '06

19 Oct '06

Hello all I have one problem with enable_shared_from_this - i have one Base class, that inherit from enable_shared_from_this<Base>. But i also need to do shared_from_this from classes inherited from Base class. Without explicit inheritance from enable_shared_from_this for Derived code fail, but when i do enable_shared_from_this<Derived>, my code fail with "error: reference to 'shared_from_this' is ambiguous" How i can resolve this issue? -- With best wishes, Alex Ott http://alexott.blogspot.com/ http://content-filtering.blogspot.com/ http://xtalk.msk.su/~ott/

4 7

[Boost.Serialization] How to make CString behave like a primitive in xml archive?
by Markus Werle 19 Oct '06

19 Oct '06

Hi! I cannot figure out how to make some class behave like a primitive and let boost::serialization rely on ostream/istream operators for xml input. This makes sense to me for e.g. String representations other than std::string. It seems it does not suffice to define the iostream operators. As an example here is my try to get things right for an MFC/ATL CString: std::istream & operator>>(std::istream &is, CString & S) { std::string s; is >> s; S = s.c_str(); return is; } #ifndef BOOST_NO_STD_WSTREAMBUF std::wistream & operator>>(std::wistream &is, CString & S) { std::wstring s; is >> s; S = s.c_str(); return is; } #endif BOOST_CLASS_IMPLEMENTATION(CString, boost::serialization::primitive_type) Output works fine due to automatic conversion of CString to char* but input fails. Debugging shows the end tag is "eaten" and appended to the value. This means that in struct C { CString SomeString; }; template<class Archive> inline void serialize(Archive & ar, C & c, const unsigned int version) { using boost::serialization::make_nvp; ar & make_nvp("SomeTag", c.SomeString); }; C c; c.SomeString = Test; reading the archive yields SomeString == "Test</SomeTag>" and an exception ("stream error") is thrown by basic_xml_archive. Any hint how to get things fixed here? This would be nice in cases a string represenation is already defined. Or should I work around this with binary_object? And how? regards, Markus

2 5

[graph] Doc für custom properties
by Jens Müller 19 Oct '06

19 Oct '06

http://boost.sourceforge.net/libs/graph/doc/using_adjacency_list.html#sec:a… says: "Creating your own property types and properties is easy; just define a tag class for your new property. The property tag class will need to define num with a unique integer ID, and kind which should be either edge_property_tag, vertex_property_tag, or graph_property_tag." But the sample tag class below don't define num. Has this become obsolete? The book (on page 52) only mentions an enum, where the value (without _t) is set to "a unique #", then BOOST_INSTALL_PROPERTY is called, the example on the above page doesn't state a value here as well.

1 0

Can't build Python library on Solaris
by chris botos 19 Oct '06

19 Oct '06

Hi, I sent this once, but I'm not sure that it made it. Please accept my apology if it is a repeat. I'm new to this list and to Boost. I'm interested in using Boost to wrap C++ objects in Python. The environment is Solaris 10, x86. I did the following: Downloaded boost_1_33_1.tar and extracted it to /opt/boost. >setenv PYTHON_ROOT /usr/local > cd to /opt/boost/boost_1_33_1/libs/python/build > bjam I get lots of error messages, repeated throughout the compile. They are generally grouped around the same line: '->' cannot appear in a constant-expression '&' cannot appear in a constant-expression template argument 1 is invalid value is not a member of '<declaration errors>' ...failed gcc-C++-action ... If I "bjam" from /opt/boost/boost_1_33_1, I get the same errors when it gets to the python library. I would most appreciate any help with this! Regards, Chris Botos

1 0