- Boost-users - lists.boost.org

[BGL] Using BGL for huge graphs
by Erik Arner 03 Feb '03

03 Feb '03

Hi, has anybody tried using BGL for really huge graphs? I mean graphs that are so big that there's no way to store the whole graph in memory. Any ideas on how it should be implemented? My first thought is that it should be fairly straightforward: simply construct classes that model Sequence or RandomAccessContainer for the VertexList parameter etc. The classes store part of the data in memory and read/write data to disc when necessary. Any pitfalls, gotchas or other stuff I should be aware of before embarking on this? Thanks, /Erik

4 9

More on static linkage
by Amittai Aviram <amittai＠amittai.com> 02 Feb '03

02 Feb '03

I have been trying to link the boost::regex library to a DLL (Release build) without success and I really need help. I have learned something new. Hitherto, I have put this line right above my code in one of my header files, i.e., within the "#if !defined()" block that prevents duplicate inclusion: #define BOOST_REGEX_STATIC_LINK This has no effect on the build whatsoever. The resulting DLL is no bigger and the client will not work if you rename boost_regex_vc6_mdi.dll from its place. You get an error message from the client saying that this file is missing. If, however, I place the line above as the very first line in a header file, then, instead, I simply cannot compile and build the project at all. The compile error I am getting states that there are "multiply defined symbols"! I.e., the symbols defined in the library are being defined twice. This is driving me nuts. Thanks for any help. Amittai Aviram amittai(a)amittai.com

1 0

Static linkage in VC++
by Amittai Aviram <amittai＠amittai.com> 01 Feb '03

01 Feb '03

I am trying to get my DLL to get linked statically with the boost::regex library and having _no success_. I have tried putting this line into one of the header files: #define BOOST_REGEX_STATIC_LINK And I have tried this line, alone and in combination with the above: #define BOOST_REGEX_NO_LIB I have tried putting the .lib file boost_regex_vc6_mdi.lib into the list of includes under Project->Settings->Link, with or without the above lines. I give up! And of course people want this thing yesterday. Please help? Thanks. Amittai Aviram amittai(a)amittai.com

2 3

Boost and the Lawyers: A Happy Ending (well for us anyway)
by Brian <bneal＠zeus.ia.net> 31 Jan '03

31 Jan '03

Hello again, A while back I posted a series of questions about what other people had done to make their companies lawyers happy about using boost in a commercial project. If you remember, I stated that our lawyers were worried that someone could post someone else's code to boost, passing it of as his own. Our company would then be liable for copyright infringement and other bad things. We thought at the time that we would have to contact each and every boost developer (whose libraries we use) and get them to sign something to say in effect "Yes I wrote this and you have permission to use it". This raised interesting questions like "should there be a standard boost license to make this process less painful"? Well our contracts guy got back to my team of developers. He checked out www.boost.org and did some research on boost. He came to the conclusion that boost was a widely known and used resource, and I guess "trustworthy". All he told us to do was include the boost developers copyright statements on any GUI and to embed said copyright strings into the executable along with ours. Somehow we lucked out. I hope you all have sensible people working for your company too. Thanks, BN And special thanks to the authors of the crc, shared_ptr, regex, and any libraries. You guys ROCK!

1 0

RE: shared_ptr - implicit conversion to bool
by scleary＠jerviswebb.com 30 Jan '03

30 Jan '03

> From: wasekvesely <vaclav.vesely(a)volny.cz> > > I've found this in "shared_ptr.hpp": > > // implicit conversion to "bool" > typedef T * (this_type::*unspecified_bool_type)() const; > > operator unspecified_bool_type() const // never throws > { return px == 0? 0: &this_type::get; } > > Why this is better than: > > operator bool() const > { return px != 0; } > > What more can I do with somethin like this? Actually, that's used to prevent mistakes; bool is an integral type, so if you add an implicit conversion to bool, then your smart pointer class can participate in integral promotions. Which allows lots of bad code to compile. :) For an example, implicit conversion to an integral type defines all the comparision operators with any other integral-convertible class. e.g., shared_ptr<int> x; scoped_ptr<int> y; (x == y); // does not do what you might expect... For more information on the problem, see the following reference: More Effective C++, Scott Myers :: Item 28 :: Testing Smart Pointers for Nullness or search the Boost developer archives. -Steve

1 0

shared_ptr - implicit conversion to bool
by wasekvesely <vaclav.vesely＠volny.cz> 29 Jan '03

29 Jan '03

Hi, I've found this in "shared_ptr.hpp": // implicit conversion to "bool" typedef T * (this_type::*unspecified_bool_type)() const; operator unspecified_bool_type() const // never throws { return px == 0? 0: &this_type::get; } Why this is better than: operator bool() const { return px != 0; } What more can I do with somethin like this? Thank you, Wasek

3 2

RE: [Boost-Users] Re: Why is <boost::function>::_1 not in namespa ce boost?
by Tom Matelich 29 Jan '03

29 Jan '03

> > I think I agree with Markus that it is in fact broken. It's easy to > imagine dependency scenarios like this: > > lambda <-- libA <-\ > }--> App > bind <-- libB <-/ > > The guy writing App shouldn't have to be aware of the dependency > conflict between libA and libB. Isn't this only a problem within translation units? I suppose lambda and bind might be included in headers from libA and libB into a single translation unit in App, but seems unlikely. ----------------------------------------------------------------------- DISCLAIMER: Information contained in this message and/or attachment(s) may contain confidential information of Zetec, Inc. If you have received this transmission in error, please notify the sender by return email. -----------------------------------------------------------------------

2 1

Advice on using iterator_adapter
by Rob Smallshire 29 Jan '03

29 Jan '03

Boosters, I have implemented a 'virtual container' class (i.e. it looks and behaves like it contains Objects, but it fact it doesn't) which implements an iterator interface of what are essentially bidirectional iterators. The iterators I have now are not STL compliant and I only have 'mutable' versions of the iterators, with no corresponding const iterators. Some guidance on how to make the best use of boost::iterator adapter would be appreciated to create fully STL compliant iterator/const_iterator pairs: 1) Out current iterators (i.e. the Base type which will be adapted) return smart reference objects when dereferenced, rather than true C++ references to objects. Will iterator adapter work in this situation? 2) Furthermore, can I use the iterator_adapter to adapt our home-grown mutable iterator into a STL compliant const_iterator? Do I simply modify the Reference of the iterator_adapter template type to be a const version of my smart-reference class? Many thanks, Rob

2 1

Why is <boost::function>::_1 not in namespace boost?
by Markus Werle 29 Jan '03

29 Jan '03

Hi! the placeholders in boost/bin/placeholders.hpp are put in an anonymous namespace. This is awkward, since like this they are visible even in places where I do not want them to be so, e.g.: #include <all_of_boost_lambda> #include <all_of_boost_function> // not unlikely to use both [...] { using namespace boost::lambda; std::for_each(v.begin(), v.end(), std::cerr << _1 << constant(' ')); } In such cases the "_1" is ambigous, since boost::lambda::_1 and and <anonymous>::_1 are both visible. This means that code which used to include boost_lambda stuff may change its meaning (or get ambigous) just by including another boost header. Could we not place those placeholders into namespace boost, just to get rid of the ambiguity? Any caveats I overlooked? (using my own placeholder definitions and BOOST_BIND_NO_PLACEHOLDERS is not really a satisfying solution) Markus

4 12

RE: [Boost-Users] Boost 101?
by Ferdinand Prantl 24 Jan '03

24 Jan '03

> From: Beman Dawes [mailto:bdawes@acm.org] > Sent: Thursday, January 23, 2003 02:50 > > At one time we published a header dependency chart, but as Boost grew > it became unwieldy and we no longer produce it. > > There have been past discussions on how to meet this need, and the > consensus is that usually what is required is a library (rather than > header) dependency analysis. Yes, I agree, it is almost impossible considering which files to include and which not with the current size of the boost package. It is sensible to declare explicitely library dependencies (tree) and use them. Nevertheless, I created a perl script some time ago to display header and library dependencies in my library with references to boost headers. I hacked it quickly to work with boost only; if you are interested it is at http://groups.yahoo.com/group/boost/files/hdr_depend. > This would be fairly easy to produce, using a number of Boost's own > libraries, of course:-) > > But the holdup is that it is only a medium priority task, and lately > has been pushed aside in favor of higher priority work. If > someone else wants > to take this one over, I'd be glad to hand it off. We want to do it > entirely in C++, so any Booster can maintain it. Well, currently I'm happily unpacking all boost but I like the idea; for example, for such a project I would like to use smart pointers only... I could spend some time on such a tool if anyone is interested, firstly gathering the requirements and then developing it. Ferda > > --Beman

4 3