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

Re: [Boost-users] Exception when reading from an archive
by kauai 05 May '05

05 May '05

--- Eyal Susser <eyal.susser(a)gmail.com> wrote: > Hi, > I'm using the serialization library, but have encountered > problems. > > I have an archiver that I use with a stringstream to > serialize > objects. I write some ints to the archive using "ar << > my_int", and > finally I archive a derived class wi ar << derived. So > far, so good. > > In a different process, I create an input archiver, read > the integers, > and then pass the archiver to a different method, where I > create an > instance of the class I serialized and use "ar >> > instance". This > throws an exception. > > The code for the object created is one *.so, and teh code > that does > the creating does it with a templated class factory in a > different > *.so. Is that why I'm gettin the exceptions? Is there a > solution? > > The exception is boost::archive::archive_exception > what(): stream error. > > Thanks. > > _______________________________________________ > Boost-users mailing list > Boost-users(a)lists.boost.org > http://lists.boost.org/mailman/listinfo.cgi/boost-users > Are you registering all your classes you are serializing? Your input archiver may not know about all class types in your archive. Use the macro BOOST_CLASS_EXPORT(ClassName) to export a GUID for every class. This should be in a .cpp file outside of any functions. include <boost/serialization/export.hpp> to use it. This include should be placed after any archiver includes. Also within your class (if it is a container for other classes) you will need to do this in your serialize function: ar.register_type(static_cast<ClassName*>(NULL)); // if they are pointers Have fun, -K __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

1 0

smart 'memory' ?
by Gottlob Frege 04 May '05

04 May '05

I find sometimes the journey is as entertaining as the destination, so let me indulge with how I got to what I'm thinking... So, I'm looking at some code. there is an object like FooManager, that manages a list<Foo *>. It has functions like RemoveFoo(Foo *), AddFoo(Foo *), etc. (Ah, so many thoughts - why not just "Remove", instead of "RemoveFoo", since we know the context, why even wrap the list at all (this one is easy - the FooManager does more than just wrap a list), etc). Anyhow, RemoveFoo(Foo *) linearly searches the list for the matching Foo * (by pointer comparison). 1. So I think, hey, really, this should be a set, not a list. And by looking at what else we do with the list, yeah, a set would be better. 2. And really, shouldn't it be a list of smart_ptr<Foo> or somesuch? (probably intrusive_ptr, as Foo has its own ref-counting, but whatever.) 3. And what about this pointer comparison? 4. etc. Just so much rewriting could be done. How do you decide where to stop! So, about the pointer comparison: Since an allocator can (and, surprisingly to some, often does) reuse the same pointer that was previously allocated and freed, how do I even know that the Foo * that should be removed, is really the "same" Foo * that is in the list. (Well, in my case, if I assume decent programming, and since the Foo's are ref-counted, no one should have a pointer to a deleted Foo, but that is assuming decent programming). So, my first thought is that each Foo should have a unique, non-recycled ID (GUID, ever incrementing int/long/etc, or something). Then I would know if I _really_ had the same Foo. And so I could do that. But we are now actually at the point where I will probably skip the whole idea and assume decent programming exists and there is no real problem. Then, I started wondering - I've had this problem before, thus where's the general solution. And what would it be? And *should* it be (ie should it exist)? So that's my question. What would it be. Would it be an allocator that included space for, and initialized, the ID, and then also special pointer type, where the pointer comparison checked the ID? Is it just dumb to try to be robust against bad programming (to this extent)? Would the Object ID be useful beyond that - ie to distinquish between: 1. these 2 objects are equal and 2. these 2 objects are the same object (ie there is only 1 object, not 2) Thoughts?

1 0

RE: [Boost-users] The Boost Community
by Brian Braatz 04 May '05

04 May '05

Excellent ideas. I hope this "happens" and is not another good idea that never gets implemented. :) > -----Original Message----- > From: boost-users-bounces(a)lists.boost.org [mailto:boost-users- > bounces(a)lists.boost.org] On Behalf Of Foster, Gareth > Sent: Wednesday, May 04, 2005 1:41 AM > To: boost-users(a)lists.boost.org > Subject: RE: [Boost-users] The Boost Community > > Hello all, > > As a C++ developer, and a developer who is coming to appreciate the > usefulness of Boost, I think this is a great idea. > > There isn't really a focal point on the web for C++ developers, at least > not > in the sense that Java developers have in java.sun.com, Mono developers > have > in go-mono.net, and as menitoned by somebody else, PHP developers have in > php.net. Boost certainly has enough clout to make this happen. > > The idea of having the following C++/Boost oriented things on one site is > fantastic. > > * blogs for people with influence over the development of the C++ standard > (think planet), > * blogs for heavy boost contributors (think planet), > * documentation for C++/Boost with user anotations + > * forum > * news section feat. latest compiler news and c++/boost developments etc > > + see recent work murray cumming did in adding anotation facilities > to gtkmm docs > > I am sure there are more ideas, but that summarises the best of them I > hope. > > Useful links that might help get this off the ground. > > - http://www.planetplanet.org/ > // will help with blogs > - http://oswd.org > // will help with site design > - http://www.murrayc.com/temp/gtkmm_book_with_comments/html/ch13s02.php > // murray's annotation effort > > It would be nice to get the site looking good right from the outset, > whilst > a "Name Competition" might not merrit a two month wait, I think perhaps a > site design competition would. Look at sites like the Firefox page, a real > central C++/Boost resource looking that professional would have to be good > for C++ as a language. > > I hope my comments and thoughts add something to the discussion. > > All the best, > > Gaz > > > -----Original Message----- > > From: Fernando Cacciola [mailto:fernando_cacciola@hotmail.com] > > Sent: 03 May 2005 17:42 > > To: boost-users(a)lists.boost.org > > Subject: [Boost-users] The Boost Community > > > > Hello people, > > > > In 1998, Bemans Dawes wrote the "Proposal for a C++ Library > > Repository Web > > Site" (you can find the paper in > > (www.boost.org/more/index.htm) which > > lunched the Boost we all know and love today. > > The focus of the site was and is the production of high-quality > > peer-reviewed C++ libraries. > > Over the years, Boost grew as a community of experienced C++ > > developers with > > the will and means to share their expertise and work > > contributing to the > > highly respected Boost Libraries. Today, Boost is a somewhat > > recongisable > > part of the C++ programming community at large. > > > > Today, the way I see it, Boost as a community encompases only > > the boost > > developers, which are just a fraction of the larger C++ > > crowd. There are > > certainly Boost users, but I don't think we can say both > > developers and > > users form a community. I think is worth and in-the-spirit to > > reach out for > > ALL C++ programmers of ALL LEVELS around an "extended" Boost > > Community whose > > main sharing point is not just the submission of high-quality > > libraries but > > also the efficient and effective use of the language. > > > > A Boost Community site would focus not directly in the > > Libraries produced by > > Boost but on C++ itself as the great language we all love and > > on fostering > > proper use of it (which includes mixing it with other languages when > > appropiate). > > It would be volunteer-driven, just like the Boost libraries > > are, so the site > > would be moderated but open for external edition (I think a > > wiki would > > work). > > I envision it having home page links to the fundamental > > resources for C++ > > programmers (CUJ,ACCU, "the" books, etc), blogs from > > developers and users > > (related to state of the art C++ and/or the boost libraties), > > important C++ > > related news (not just boost-library related news), reviews, etc... > > essentially, everything the will of volunteers can bring on to the > > community. The motto being "fostering proper use of C++". > > A minimum of moderation would be required but I assume not much. > > > > I know there are other programmer communities around (even > > C++ focused). The > > reason I propose this here (around Boost) is becasue I think that a > > community centered around the Boost Libraries but reaching > > out could stand > > out for its quality in the same way the Boost Libraries do. > > I also know that a newsgroup or mailing list is a form of > > community, and we > > have c.l.cpp.m and c.s.cpp. But they lack the verstile structure and > > dynamism of the today's virtual communities. Also, somehow I > > believe that > > centering a community around the Boost Libraries would spread > > their quality > > into the rest of the site; to some degree at least. > > > > I welcome comments, ctricticism and I call interested people > > to put forward > > concrete ideas. > > > > Best > > > > Fernando Cacciola > > SciSoft > > > > > > > > _______________________________________________ > > Boost-users mailing list > > Boost-users(a)lists.boost.org > > http://lists.boost.org/mailman/listinfo.cgi/boost-users > > > _______________________________________________ > Boost-users mailing list > Boost-users(a)lists.boost.org > http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

RE: [Boost-users] Re: The Boost Community
by Brian Braatz 04 May '05

04 May '05

> -----Original Message----- > Sent: Tuesday, May 03, 2005 6:29 PM > > "Eric Goebelbecker" <egoebelbecker(a)gmail.com> escribió en el mensaje > news:7611ef5a05050311052664ea@mail.gmail.com... > > I'm new to Boost and have been lurking on the list for a short time. > > > > This sounds like a good idea to me. Although I mean lean toward a > > portal type site, only because wikis have been compromised frequently > > of late, and a volunteer site doesn't need the additional headaches. > > > > I may be able to help out with web hosting space. I have an "extra" > > account and a low-budget hosting company. > > > > Best, > > Fernando Cacciola > SciSoft [Brian Braatz Writes:] #1 GO GO GO ! This is an excellent idea #2 I suggest if you cannot get a donated host, that you setup a PayPal link on the site. This would allow people to donate. I for one would donate money to keep the site running.

1 0

RE: [Boost-users] The Boost Community
by Foster, Gareth 04 May '05

04 May '05

I might be wrong here, but I think the gist of the discussion was that coming up with something bigger than just a documentation Wiki was the way forward. I am new to boost, I had no idea there was a Wiki. A site with documentation as well as blogs and news on C++ and Boost would pull more traffic than a Wiki alone, in doing so it would hopefully become established, this would in turn drive the expansion of the Wiki, leading to a general trend towards the resource being indispensible for C++ developers. Ideally, you want the situation where C++ student of one year, A, upon being asked by C++ student of 1 month, B, where he can find std::cout documentation on line, will instantly point to one place, as he would with PHP etc. I'll admit, as soon as I saw the word "skeptic", the enormity of the task dawned on me, maybe it is more than a little pie in the sky, then again, it would be great if successful. The anoying part of this is surely that all the tools to build a great community site are always there, its just the theming and pooling of the technologies that is tough. Gaz > -----Original Message----- > From: Jeff Garland [mailto:jeff@crystalclearsoftware.com] > Sent: 04 May 2005 09:10 > To: boost-users(a)lists.boost.org > Subject: Re: [Boost-users] The Boost Community > > I'll just respond in general to the thread. It's great to > see some enthusiasm > from the boost user community. I'm afraid to say, however, > that I'm a bit of > a skeptic. Having started the Boost Wiki in 2001 after a > burst of similar > discussion and the initial creation of the Boost-user mailing > list, I consider > the wiki to be basically a failed experiment. That's not to > say that the > Boost_User Wiki hasn't been very useful to many people > (especially Boost > developers). But I'd say that there has never been sustained > "user" support to > help grow the Wiki and make it more useful for Boost users. > The initial > vision was that users would contribute 'tutorial' and other useful > information. It's been sparse at best. > > So my thought is that before we go and grow another > technological solution and > diluting our resources even more, we need to really think > about what the > 'users' want to achieve and why the current site doesn't meet > those needs. Why > don't more "users" contribute to the current Wiki? What is > it that users > really want to contribute to Boost? Why not build on some of > the good things > on the current wiki (like the 'effective xyz' pages? Why not > band together > and help work on the current Wiki? If it's additions to docs there are > certainly many ways that can be done currently and is done > all the time by a > variety of users. It's alot of work to do it well I believe > -- more than I > can do. And if we want to augment with other technologies we > can -- I'm fully > open to expanding the hosting I already provide. > > As for the wiki spam problem, I believe it to be a non-issue > at this point. > The problem really started to take off in mid-2004, but a > series of measures > have been implemented to reduce and resolve the problem. Yes, > spam still > appears from time to time, but it is quickly and completely > reversed. The > spammers that get thru initially get blocked out because the > new content > filter stops them from posting links to spammer sites after > an admin update. > The more organized and dangerous spammers seem to have moved > on to easier > targets. And at this point, any open content system is going > to have to face > this problem -- so be prepared. > > Jeff > > > > > _______________________________________________ > Boost-users mailing list > Boost-users(a)lists.boost.org > http://lists.boost.org/mailman/listinfo.cgi/boost-users >

3 2

RE: [Boost-users] genetic programming
by Lyon, Richard 04 May '05

04 May '05

Hi, Are you talking about genetic programming or optimisation using genetic algorithms? There is a big difference. There are some pretty good c++ libraries available for genetic algorithms. Galib is one that springs to mind. Solving graph problems using genetic algorithms is a bit hit and miss. Some randomising techniques have been used to solve things like mst efficiently. These are different to a genetic algorithm. As I understand it genetic programming is the concept of using genetic algorithms to write code which will solve an optimisation problem. Rather a difficult problem. Well that enough bignoting, thanks ... Richard -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Martin Jynge Sent: Wednesday, 4 May 2005 10:31 PM To: boost-users(a)lists.boost.org Subject: [Boost-users] genetic programming Hi. Ayone interested in gp using boost?

2 1

serialized memory allocation using boost::pool
by Davi de Castro Reis 04 May '05

04 May '05

Hi, I am trying to improve the performance of a std::list instance in a multi-thread application using boost::pool, but the syncronization overhead is killing me. The point is that all access to my std::list are serialized, so I would like my pool to not use mutexes when retrieving memory in the pool, which is what I suppose is happening (got evidence through valgrind). From the template parameters of boost pool, I think this is possible, but I couldn't figure out how to do this. Any tips here? pseudo code: struct X; std::list<X *, boost::fast_pool_allocator<X *> > l; lock(); //fill l with 1000 X * instances unlock(); lock(); for (uint32 i = 0; i < 1000; ++i) { X *x = l.front(); l.erase(l.begin()); l.push_back(x); } unlock(); Thanks in advance, Davi de Castro Reis

1 0

Boost.Test and Visual Studio .Net 2003
by Robert Cober 04 May '05

04 May '05

I am trying to learn how to use boost.test. I have a solution with two projects in it: MyApp - this is the application I am building Test - this is the project for the test framework. Everything builds ok. Simple tests run fine, errors can be clicked on, all is well. My question has to do with how the Test project can refer to the code within MyApp. If I want to refer a class or function defined in MyApp, do I have to manually add the appropriate .h and .obj files to Test? Is there any way to have Test be able to automatically refer to the files of other projects within this workspace? Thanks in advance for any help......

3 2

Re: [Boost-users] genetic programming
by Martin Jynge 04 May '05

04 May '05

Quoting steve ahlgren <steveahlgren(a)gmail.com>: > I've messed around with genetic optimization routines and have > applied > them to a number of different applications with excellent success. I > would love to see a "real" toolkit available in something like boost. It shure would be fancy. > What algorithms have you implemented? Well, the old function regression was my first implementation, but since then the code has changed a lot. I'm currently working on the Serengeti-problem (with lions and a gazelle, you know). It's not running yet, because of design dilemmas.. urk.. There are two things I have difficulties with: first i'd like to put values in the edges. But my implementation uses different value types in one graph, like a double and a vector of doubles. How does one implement that in boost? Second, it seems that I have to use indices with each node, in order to put the right edges in the right slots. So what kind of container should be used? > I'll have to dig through my old > code to see what I have lying around. I'd love to see what you've got. > Great idea; I know! regards, Martin Jynge

1 0

genetic programming
by Martin Jynge 04 May '05

04 May '05

Hi. Ayone interested in gp using boost? I've seen a few earlier posts on trees, but my main interest is a true graph approach.. and it's a little difficult implementing nodes having several outputs (as opposed to a single one i tree-gp). Looping is also a concern as you might guess. I don't have any usable code yet as I'm still learning boost, but I will soon try to port my non-boost code to boost (likely with some help, though) Any input would be highly appreciated :) regards, Martin Jynge --

1 0