- Boost-users - lists.boost.org

boost::iostreams - user defined format flags
by Alan M. Carroll 08 Aug '05

08 Aug '05

I'm trying to use the boost::iostreams library but I've run in to a problem that makes that difficult. The basis is that I want to write manipulators to manipulate the behavior of the filters in a filtering iostream. The standard mechanism for such manipulators are user defined format flags for streams, accessed via xalloc() and iword()/pword(). These are not available in a filter. It is also not feasible to store the formatting state in the filters, as there are not accessible from the filtering stream (and hence inaccessible to the manipulators). I exchanged an e-mail with the author in which he states that a forthcoming version of boost::iostreams will provide access to the filters from the stream object. Manipulators could then reach in to the filter objects and modify them to adjust formatting behavior. However, while this is desirable for other reasons, I'd prefer to be able to use iword/pword from the filters, i.e. be able to call these methods on the Sources and Sinks (which, presumably, would forward them to the actual stream). It seems more "standard" to me, although one could argue for storing state directly in the filters from a clean design point of view. Does anyone else have an opinion on this? Or have a suggestion for an alternate method of creating manipulators to adjust the behavior of the filters using stream local state? I searched around for previous discussion of this issue but didn't find one. If there was, a pointer to it would be much appreciated. P.S. For those interested, the details of what I want follow. I'm doing some work with XML which involves user input. Because one never knows what a user might put in his data, I need to encode/decode problematic characters such as angle brackets. The problem with having streams to do the encode/decode is that one doesn't always want this performed on the data. For instance, during output one needs to send literal angle brackets around the element tags and I don't want to get in to trying to second guess the encoder to work around it. The solution I decided to implement is to have manipulators that turn the coding on and off, so one would do something like this: xml::coding::ostream s; s << "<tag>" << xml::coding::on << node.content() << xml::coding::off << "</tag>"; This allows low level code fine grained control over what gets coded while leaving the details up to the stream provided by the application. The hassle of setting up strings for in place conversions is also avoid, along with the problem of making sure every body is using the same coding algorithms. However, it doesn't help all that much on input as a parser has to read from a non-filtered stream to correctly detect the tags. Element content will end up in a string and it's no easier to use a filtering stream on that than to just call a decode function.

1 0

[bcp]minimal option
by Thomas Matelich 08 Aug '05

08 Aug '05

I was really shocked when a bcp shared_ptr foo/ copied 274 files. I expected a lot, but not that many. Then I read the doc's more closely and it says that if you depend on any header from a lib, you get the whole lib. Could it be possible to add a -minimal option to have only the actual headers used included? Thanks, Tom

3 3

Re: [Boost-users] [Serialization] 1.33 RC1 Problem linking void_cast_register
by Richard Jennings 08 Aug '05

08 Aug '05

Thanks for your response on this. I have tried moving the void_cast_register to the derived class constructor (without success) and I have looked further in to my use case. This was probably brought about by me removing the archive and serialization headers from the precompiled header file that Borland uses because there were other compilation problems in keeping them in. The use case is as follows: Classes A and B are derived from class Z. Class M has a pointer to class Z as member data. This member data is set to a pointer to an instance of A or B as appropriate. Class S serializes M. Class S includes the Archive headers as it creates the archive class instance. S includes the headers for M that includes the headers for Z. However S does not include the headers for A or B, because it does not need to, and in fact does not need to know of the existence of A or B as they are used polymorphically. In our case S does not use A/B/Z but others use them via M. I have a free function called RegisterTypes in a module that has BOOST_CLASS_EXPORT(A) and void_cast_register<A,Z>(0,0) in it. It's run once at application start. Which means that A.hpp is never included in a compilation unit with any archive header and therefore the void_cast_register function is not generated (as I understand it). I realise that serializing classes via a pointer to the base class and void_cast_register<T> is a seldom used corner of serialization. I can fix the problem by including archive headers where RegisterTypes is defined even though they may not be needed, and then the application links ok. I'm not sure if there is a need for a change, you are best to judge. I hope this explains my use-case. Richard -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Robert Ramey Sent: 04 August 2005 17:57 To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] [Serialization] 1.33 RC1 Problemlinkingvoid_cast_register I believe this is a side effect of our efforts to be sure that the mere inclusion of a class that includes serialization headers doesn't trigger (through auto-linking) the requirement for inlcuding a library module. The idea is that one would never need "void_cast_register<T>" to be generated unless there is an archive somewhere involved. So that raises the question of what the context is here. Your test illustrates the link error but it doesn't illustrate the use case. That is, what is the context of explicitly invoking void_cast_register without an archive header present? I'm not saying its wrong. Its just that I don't see how this comes up. When we implemented changes which resulted in this symptom - we enivisioned a pattern of usage like that shown below. In this way, the *.hpp files could be included in other programs without triggering auto-link and requiring linking with the boost library to include code that in fact will never be used. Were we missing something here? What is it? Robert Ramey Z.hpp ==== #include <boost/serialization/is_abstract.hpp> #include <boost/serialization/serialization.hpp> #include <boost/serialization/export.hpp> class Z ... }; BOOST_IS_ABSTRACT(Z) A.hpp ==== #include ... // other serialization headers #include "Z.hpp" class A : public Z { A(){ boost::serialization::void_cast_register<A, Z>(0,0); } }; BOOST_CLASS_EXPORT(A) B.hpp ==== #include ... // other serialization headers #include "Z.hpp" class B : public Z { ... B(){ boost::serialization::void_cast_register<B, Z>(0,0); } }; BOOST_CLASS_EXPORT(B) main.cpp ====== #include <boost/serialization/text_oarchive.hpp> #include <boost/serialization/text_iarchive.hpp> #include "Z.hpp" #include "A.hpp" #include "B.hpp" int main(int argc, char* argv[]){ return 0; } _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

2 1

[serialization] 1.33 RC1 vc1.7 serializing containers - include order
by Zrubavel Zodikovich 08 Aug '05

08 Aug '05

The following code produces error saying that archive is included after serialization. What is the correct place to include map.hpp and other predefined implementations when making a class with container members serializable ? a.h ----- #include <boost/serialization/serialization.hpp> #include <boost/serialization/map.hpp> class A { public: template<class Archive> void serialize(Archive & ar, const unsigned int version) { ar & someMap; } std::map<int, int> someMap; }; ----- Thanks

3 11

[1.33.0] Release candidate #1 is available
by Doug Gregor 08 Aug '05

08 Aug '05

Boost 1.33.0 release candidate #1 is now available here: http://www.osl.iu.edu/~dgregor/boost-1.33.0/ There are two sets of files important for this release: - boost_1_33_0_rc1.*: The actual release candidate - boost_1_33_0_(docbook|fo|man).*: The documentation in other formats Please download and unpack the archives, browse the documentation, try the release with your favorite compiler, etc. Thanks! Doug Gregor 1.33.0 Release Manager

5 12

Fw: boost::enable_if : usage clarification
by Suresh T 08 Aug '05

08 Aug '05

hi There is a correction in my snippet that i attached for enable_if. The correct snippet is //primary class template <std::size_t size , template<std::size_t size,typename E = void > class USER_TYPE = select_type > class my_class { public: //Primary class //Note that the select_type is default for USER_TYPE }; //Specialized class template <std::size_t size , template<std::size_t size,typename E = void > class USER_TYPE > class my_class< size,SER_TYPE> { public: //The commented code line below does not work for VC7.1 // typedef typename USER_TYPE<size>::type data_type; //THIS DOES NOT WORK with vc7.1 typedef typename USER_TYPE<size,void>::type data_type; //THIS WORKS }; --------------------------------------------------------------------------------------------------------- Rgds ----- Forwarded by Suresh T/BTC/SC/PHILIPS on 08/08/2005 02:03 PM ----- hi I am using the boost::enable_if for selection of types. Something similar to below template<std::size_t Size,class Enable = void> struct type_select{ }; template<std::size_t Size> struct type_select< Size, typename boost::enable_if < bool_ < (size <= 8) > > >::type >{ typedef typename char type; }; template<std::size_t Size> struct type_select< Size, typename boost::enable_if < bool_ < (size > 8) > > >::type >{ typedef typename int type; }; The conditions are based on the first parameter size. The normal usage would be something like type_select<10>::type; My question is : If i explicitly use the second parameter ( by passing void) will the correct specialization be still used i.e, if i use say type_select<6,void>::type, will it resolve to "char" in the above case. Please note: I was using the normal way (without the void) for my code on linux+gcc without any problems My problem came up when i tried to move my code to windows + vc7.1. The vc7.1 compiler needs the second argument (in my usage shown below ) else it gives an error. The results are as desired, but i would like to know whether this usage is correct. Below is a snippet where the vc7.1 compiler gives a problem with the usage. Note type_select is not directly used but is used as a default type for USER_TYPE parameter template <std::size_t size , template<std::size_t size,typename E = void > class USER_TYPE > class my_class < size, USER_TYPE = select_type //select_type is a default parameter > { public: //The commented code line below does not work for VC7.1 // typedef typename USER_TYPE<size>::type data_type; //THIS DOES NOT WORK with vc7.1 typedef typename USER_TYPE<size,void>::type data_type; //THIS WORKS }; Thanks for clarification in advance.. --------------------------------------------------------------------------------------------------------- Rgds, Suresh

1 0

boost::enable_if : usage clarification
by Suresh T 08 Aug '05

08 Aug '05

hi I am using the boost::enable_if for selection of types. Something similar to below template<std::size_t Size,class Enable = void> struct type_select{ }; template<std::size_t Size> struct type_select< Size, typename boost::enable_if < bool_ < (size <= 8) > > >::type >{ typedef typename char type; }; template<std::size_t Size> struct type_select< Size, typename boost::enable_if < bool_ < (size > 8) > > >::type >{ typedef typename int type; }; The conditions are based on the first parameter size. The normal usage would be something like type_select<10>::type; My question is : If i explicitly use the second parameter ( by passing void) will the correct specialization be still used i.e, if i use say type_select<6,void>::type, will it resolve to "char" in the above case. Please note: I was using the normal way (without the void) for my code on linux+gcc without any problems My problem came up when i tried to move my code to windows + vc7.1. The vc7.1 compiler needs the second argument (in my usage shown below ) else it gives an error. The results are as desired, but i would like to know whether this usage is correct. Below is a snippet where the vc7.1 compiler gives a problem with the usage. Note type_select is not directly used but is used as a default type for USER_TYPE parameter template <std::size_t size , template<std::size_t size,typename E = void > class USER_TYPE > class my_class < size, USER_TYPE = select_type //select_type is a default parameter > { public: //The commented code line below does not work for VC7.1 // typedef typename USER_TYPE<size>::type data_type; //THIS DOES NOT WORK with vc7.1 typedef typename USER_TYPE<size,void>::type data_type; //THIS WORKS }; Thanks for clarification in advance.. --------------------------------------------------------------------------------------------------------- Rgds, Suresh

1 0

Re: [Boost-users] question: boost::array does not have compile time checked access
by Merrill Cornish 07 Aug '05

07 Aug '05

>>> the places where you would use fixed arrays, >>> you probably also know at what indexes certain data resides. If you use fixed arrays in that way, then you are using them much like structs where, say, [0] mean THIS and [1] means THAT and [2] means something else etc. If that were the case, then individually named variables would likely work better. The more common reason for fixed arrays is that the array is matching something in the real world: I'm doing a 256 element Fast Forier Transform, I'm processing census data for 50 US states, I'm tallying sales for the 12 months. Even if you can attach a specific meaning for individual elements, most of the processing is still done either by iterating over the array or by accessing an element based on an index number extracted from incoming data. In general rogramming the occassions where you would be accessing the array by a constant index is still very rare compared to accessing the elements by a variable index. Even if you did have an application where almost all of the array accesses were with constant indexes, then you would also certainly create a set of constants so that you would, say, extract the state name from an array of address strings by saying [WYOMING] rather than [49]. You set the constants up once and you're done. No more array bounds access errors. Even if the initial set of constant definitions had an error, once fixed, it's fixed forever. Again, the number of places where checking array bounds on fixed length arrays is too small to justify work involved. Merrill

2 1

Boost.Iostreams
by Steven T. Hatton 07 Aug '05

07 Aug '05

I just happened to open Konqueror to the root of a recent boost CVS, and discovered an HTML file with some information I've never seen before. Among that information was a link to the Boost.Iostreams library. I don't see any reference to it on the Boost homepage, and when I searched the site I didn't find much there. But I've found it elsewhere on the Internet. I'm confused as to the status of the library. Is it a continuing project? I sure hope so. -- Regards, Steven

2 3

[1.33.0] Release candidate #2 is available
by Douglas Gregor 07 Aug '05

07 Aug '05

Boost 1.33.0 release candidate #2 is now available at the usual place: http://www.osl.iu.edu/~dgregor/boost-1.33.0/ Please build it, test it, and report back to the mailing list. If no major problems are found before August 9th, this release candidate will become Boost 1.33.0. Doug Gregor 1.33.0 Release Manager

1 0