- Boost-users - lists.boost.org

Re: [Boost-users] Getting started with Boost: what did you need to know?
by Drumheller, Michael 04 Feb '06

04 Feb '06

Main stumbling point for me was bjam. It's not so much of a problem for building the libraries themselves--you don't need to know anything about it at all, really. But when you start building your own extensions, it would be helpful to have a dozen or so examples of really basic operations, like changing the destination directory, defining a preprocessor macro on the compiler command line, stuff like that. I have to say, I'm almost embarrassed to be writing this, because it seems like nobody on this list ever complains about bjam. So maybe it's just me. But to me, bjam is an incomprehensible monster, a big YALITTC (yet-another-link-in-the-tool-chain) that just makes it harder to sell boost within an organization. Michael D.

3 2

[serialization] seems like a bug in MSVS 7.1 is preventing use of the library
by cheesy4poofs 04 Feb '06

04 Feb '06

Hello, I've been trying for days to get a simple serialization example to work in our main project. I finally started taking snippets of code from the regular project into a sample project to post the code here. However, all the code in my sample project always compiles and runs. It's practically identical to our main project. At this point, I'm pretty sure I've stumbled unto a bug in the compiler. I even found a web page that seems to have the same problem I have: http://aspn.activestate.com/ASPN/Mail/Message/boost/2768351 Here's a small sample of the problem: struct Node { unsigned int handle; template <typename Archive> void save(Archive& ar, const unsigned int version) const { ar << handle; } template <typename Archive> void load(Archive& ar, const unsigned int version) { ar >> handle; } BOOST_SERIALIZATION_SPLIT_MEMBER() } In a nutshell, I have a std::map<unsigned int, Node> that I'm trying to serialize. The code to save compiles fine. But the load() method fails on compile with the STATIC_ASSERTION_FAILURE. If I comment out the ar >> handle;, it compiles fine. If I then uncomment ar >> handle; (after a good compile), it STILL COMPILES FINE. If I do a clean/rebuild of the project, it then fails with the STATIC_ASSERTION_FAILURE! I went into the iserializer.hpp:577 header of boost and found the offending line: // if this assertion trips. It means we're trying to load a // const object with a compiler that doesn't have correct // funtion template ordering. On other compilers, this is // handled below. BOOST_STATIC_ASSERT(! boost::is_const<T>::value); Based on the comment above, it seems MSVC 7.1 is having problems with correct function template ordering? On a whim, I commented out the BOOST_STATIC_ASSERT and now I see this compile error: ..\API\inc\boost\archive\basic_text_iprimitive.hpp(80) : error C2679: binary '>>' : no operator found which takes a right-hand operand of type 'const unsigned int' (or there is no acceptable conversion) ..\API\inc\boost\archive\text_iarchive.hpp(45) : see reference to function template instantiation 'void boost::archive::basic_text_iprimitive<IStream>::load<T>(T &)' being compiled with [ IStream=std::istream, T=const unsigned int ] My question is - how can I solve this so I can use serialization? The link above seems to indicate boost 1.32 works for Microsoft Visual Studio 7.1. Are there any other possible solutions that don't involve rolling back to an earlier boost version? Thanks for any help, Scott

2 1

Re: [Boost-users] walk thru multi_index
by JOAQUIN LOPEZ MU?Z 03 Feb '06

03 Feb '06

[About the documentation of Boost.MultiIndex, Evan says:] ----- Mensaje original ----- De: Evan Carew <carew(a)pobox.com> Fecha: Jueves, Febrero 2, 2006 7:56 pm Asunto: Re: [Boost-users] walk thru multi_index > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Joaquín, > > They are a little lite. I had to ask for help as well to get started. > In my case, I defined a fact pattern I needed help with and someone > kindly gave me an example, which I am now using. > Could you elaborate on that pattern? If it's general enough, I can look for a way to have it explained in the docs. When receiving specific reports about the docs I usually try to use the feedback --for instance, you complained some weeks ago about the lack of a explicit mention to index_iterator<> and family, and I added it to the docs for Boost 1.34 (you can check the preview version at the Vault.) Other than this, I'm sorry you found the documentation somewhat light. I've made sure every feature of the lib is treated, but I try also to keep the tutorial compact, and sometimes the style can be terser than desired. Also, sometimes people have questions wich are actually not related to Boost.MultiIndex --the very same problem posed by Dmitriy Bure, for instance, is more of an associative container issue, not specifically related to multi_index_container. Joaquín M López Muñoz Telefónica, Investigación y Desarrollo > Version: GnuPG v1.4.1 (GNU/Linux) > > iD8DBQFD4lXApxCQXwV2bJARAsaxAJ9rF9IBfLTUWKwh8DCUHPnCj/5gkACgmj/q > NpcBmDY2hDI+IF+ZAWQPUb0= > =EfPx > -----END PGP SIGNATURE----- > > _______________________________________________ > Boost-users mailing list > Boost-users(a)lists.boost.org > http://lists.boost.org/mailman/listinfo.cgi/boost-users >

3 3

Re: [Boost-users] Test of e-mail client - please ignore.
by Sohail Somani 03 Feb '06

03 Feb '06

> Brian Allison <brian(a)dtmr.com> writes: > > > The test below this line should be 'quoted'. Hopefully the > mail gate won't wreck it. > > > > Brian Allison wrote: > > > > Testing quoting other messages... > > Try turning off HTML messages in your client. :) I love how these threads start.

1 0

view of multiarray without knowledge of dimension
by graphicsMan69 (sent by Nabble.com) 03 Feb '06

03 Feb '06

Hi all - My basic need is to step through a portion of my multiarray at runtime. I don't know the dimension apriori, since the class I'm writing is templatized on dimension. This means that the multiarray can serve my other purposes just fine, since the dimesion is compile time constant, but I don't know the number of [] and the nesting of loops I'll need to iterate through the valid indices in my range min to max in my multiarray. Essentially, if I could specify a view of my multiarray, likely it shouldn't be too hard to iterate through the elements of that sub-multiarray, even without knowing the dimension at coding time. Anyone know of a way to get a view without knowing the dimensions at coding time? Thanks, Brian -- View this message in context: http://www.nabble.com/view-of-multiarray-without-knowledge-of-dimension-t10… Sent from the Boost - Users forum at Nabble.com.

2 2

Re: [Boost-users] Getting started with Boost: what did you need toknow?
by Nat Goodspeed 03 Feb '06

03 Feb '06

> -----Original Message----- > From: boost-users-bounces(a)lists.boost.org [mailto:boost-users- > bounces(a)lists.boost.org] On Behalf Of Drumheller, Michael > Sent: Friday, February 03, 2006 12:51 PM > To: boost-users(a)lists.boost.org > Subject: Re: [Boost-users] Getting started with Boost: what did you need > toknow? > > Main stumbling point for me was bjam. > I have to say, I'm almost embarrassed to be writing this, because it > seems like nobody on this list ever complains about bjam. So maybe it's > just me. But to me, bjam is an incomprehensible monster, [Nat] It's not just you. The more we looked at bjam, the less convinced we were that we even wanted to become good at it. When it didn't work for us right out of the box (building with MSVC 7.1 + STLport), and initial futzing showed little promise, we just constructed our own MSVC projects for the Boost libraries we needed to build, and have never looked back.

1 0

Re: [Boost-users] Bind/lambda arithmetic question
by Nat Goodspeed 03 Feb '06

03 Feb '06

> -----Original Message----- > From: boost-users-bounces(a)lists.boost.org [mailto:boost-users- > bounces(a)lists.boost.org] On Behalf Of Stephen Gross > Sent: Friday, February 03, 2006 12:52 PM > To: Boost > Subject: [Boost-users] Bind/lambda arithmetic question > > ============================= > struct Foo > { > double getX() const{ return 5; } > double getY() const { return 10; } > }; > > Foo f; > > double z = (boost::bind(&Foo::getX, boost::ref(f)) / > boost::bind(&Foo::getY, > boost::ref(f))) (); > ==================== [Nat] This might work better if you were to use boost::lambda::bind() rather than boost::bind()?

1 0

Test of e-mail client - please ignore.
by Brian Allison 03 Feb '06

03 Feb '06

2 1

Bind/lambda arithmetic question
by Stephen Gross 03 Feb '06

03 Feb '06

Ok, I want to create a functor that takes no arguments and returns a double. Furthermore, I want this functor to be composed of functors itself. For instance: ============================= struct Foo { double getX() const{ return 5; } double getY() const { return 10; } }; Foo f; double z = (boost::bind(&Foo::getX, boost::ref(f)) / boost::bind(&Foo::getY, boost::ref(f))) (); ==================== The above code, however, generates the following error: error: no match for 'operator/' in 'boost::bind(R (T::*)() const, A1) [with R = double, T = Foo, A1 = boost::reference_wrapper<Foo>](boost::ref(T&) [with T = Foo]()) / boost::bind(R (T::*)() const, A1) [with R = double, T = Foo, A1 = boost::reference_wrapper<Foo>](boost::ref(T&) [with T = Foo]())' Is there a way to make the division operation of two functors work in this example? Thanks, --Steve Stephen Gross Case Western School of Medicine Cleveland, OH "By Grabthar's hammer, by the sons of Worvan, you shall be avenged." - Dr. Lazarus

1 0

Re: [Boost-users] Binding to a static member function?
by raulh39＠ya.com 03 Feb '06

03 Feb '06

Stephen Gross wrote: >I'm trying to figure out the correct syntax for boost::bind-ing a static >member function with some arguments preset. ... > > This code works for me: #include <boost/bind.hpp> #include <boost/function.hpp> #include <iostream> struct Foo { static double go(int x, int y) { return x + y; } }; int main() { typedef boost::function<double(int)> functortype; functortype f = boost::bind(&Foo::go, _1, 4); std::cout << f(3) << std::endl; } Hope it helps.

1 0