March 2008 - Boost-users - lists.preview.boost.org

Automatically generate serialization code?
by Alex Idar 31 Mar '08

31 Mar '08

Hi there. I am looking for ideas about automatically generating the serialization helpers. Boost serialization library would be used - the question is about a simple (hopefully) and foolproof way of automatically creating the helpers. The classed that need serialization are modified on a daily basis by other developers. Maintaining the serialization code manually isn't an option - looking for an automated way. One possible approach: 1) Run a parser that scans the C++ code and identifies the data members of the classes that declare serialize() 2) Generate .inl files (one for each class to be serialized) with the boost::serialize() methods 3) Include the .inl files in the corresponding .cpps 4) Compile and link... An idea thrown by someone was to get the data members by getting the symbols from the .obj files - but this would require two compiles - one to generate the object files from where the symbols would be extracted, then another one... Definitely someone else had solved this problem (with the above approach or a different one). Please let me know your opinions. Regards, Alex ____________________________________________________________________________________ You rock. That's why Blockbuster's offering you one month of Blockbuster Total Access, No Cost. http://tc.deals.yahoo.com/tc/blockbuster/text5.com

2 2

[GSoC08] Student Application Deadline Extended
by JOAQUIN M. LOPEZ MUÑOZ 31 Mar '08

31 Mar '08

To those who are thinking of submitting an application to Google Summer of Code 08 and didn't make it on time for the original deadline... http://code.google.com/soc/2008/ Cheers, Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

1 0

[intrusive] Compile error on Solaris, SunStudio 12
by Zeljko Vrba 31 Mar '08

31 Mar '08

Can someone please confirm whether this is a bug in the code or in the compiler? == #include <boost/intrusive/list.hpp> using namespace boost::intrusive; struct C : public list_base_hook<> { int i; }; list<C> test_list; == Compiling with g++ 4.2.2 works fine, whereas compiling with CC (CC: Sun C++ 5.9 SunOS_i386 Patch 124864-02 2007/12/18) gives the following errors: "/home/zvrba/COMPILE/boost_1_35_0/boost/intrusive/list.hpp", line 107: Error: po inter is not a member of boost::intrusive::detail::generic_hook<boost::intrusive ::get_list_node_algo<void*>, boost::intrusive::default_tag, 1, 1>. "/home/zvrba/COMPILE/boost_1_35_0/boost/intrusive/list.hpp", line 1409: Wher e: While specializing "boost::intrusive::list_impl<boost::intrusive::listopt<boo st::intrusive::detail::generic_hook<boost::intrusive::get_list_node_algo<void*>, boost::intrusive::default_tag, 1, 1>, unsigned, 1>>". "/home/zvrba/COMPILE/boost_1_35_0/boost/intrusive/list.hpp", line 1409: Wher e: Specialized in boost::intrusive::list<C, boost::intrusive::none, boost::intru sive::none, boost::intrusive::none>. "q.cc", line 10: Where: Specialized in non-template code. (and 10 additional errors about something [const_pointer, iterator_traits, etc.] not being a member of ..) Thanks.

2 3

Functions that evaluate to constants
by Robert Dailey 31 Mar '08

31 Mar '08

Hi, I realize that C++03 will not permit functions to evaluate down to compile-time constants. However, I cannot wait until C++0x for this feature. So, I'm hoping that Boost will allow a temporary solution. Consider the following structure: static struct nullptr_t { template< typename t_type > operator t_type*() { return 0; } } nullptr; The compiler has everything it needs to turn this code into a single integral constant value. So, if I do the following: int* foo; if( foo == nullptr ) { } Then it should evaluate down to: int* foo; if( foo == 0 ) { } Is there any way to provide this behavior in Boost? I would like to avoid altering the actual design if possible. Perhaps this actually *does* evaluate to a constant and the problem is that I'm actually unaware of some specific rule that I didn't find in the C++03 standard. Thanks.

3 6

[1.35.0] Release Candidate 3 available
by Beman Dawes 31 Mar '08

31 Mar '08

See http://boost.cowic.de/rc/ For this release candidate the naming is exactly the same as it will be in the final release. Unless a showstopper emerges, these exact files will become the final release. Before declaring the release done, we need to verify that none of the four package files is corrupt and installs without error. I'm starting to do that now. I encourage others to also look for showstoppers, and report results on the developer or user lists. If no showstoppers surface by 9 AM tomorrow morning, US Eastern time, I'll declare 1.35.0 done, create the tag, and push the files out to Sourceforge. --Beman

7 15

Compile error boost 1.35.0 on Solaris 10/x64 Studio 12
by Anders Moe 31 Mar '08

31 Mar '08

Hi all I'm getting the following error in boost 1.35.0 on Solaris 10 using Sun Studio 12 : "/home/amoe/src/boost_1_35_0/boost/format/alt_sstream_impl.hpp", line 252: Error: Cannot assign void* to char*. Where: While instantiating "boost::io::basic_altstringbuf, std::allocator>::overflow(int)". Where: Instantiated from non-template code.1 Error(s) detected. I'd like to track down which part of my own code initiates the error, but frankly I'm not sure how to do that - the compiler output reports nothing but the above error. These are the relevant lines from alt_sstream_impl.hpp : add_size) { new_size += add_size; newptr = alloc_.allocate(new_size, is_allocated_? oldptr : 0); // ERROR } Regards, Anders

4 3

query for Boost support
by rishi keisham 31 Mar '08

31 Mar '08

Hi all, I have read the Boost supported platforms on the Boost website. I am kind of interested in knowing possibilities for Boost on more platforms. Will it be able to build the boost C++ library on MontaVista linux.. At present I am working on "mipsel-linux-g++ compiler". Thanks in advance. Thanks, Keisham --------------------------------- 5, 50, 500, 5000 - Store N number of mails in your inbox. Click here.

2 1

Defining Macro's using bjam or "config/user.hpp".
by Stephen Hewitt 31 Mar '08

31 Mar '08

I have just downloaded Boost 1.35.0 for use with MSVC9 (Microsoft Visual Studio 9.0). As some of you may be aware, Microsoft has added some features which are a little troublesome to work around with Boost. They are controlled by the following macros (and perhaps others I've missed): - _CRT_SECURE_NO_WARNINGS - _SCL_SECURE_NO_WARNINGS - _SECURE_SCL I must say, I find building Boost a pain at the best of times, but I can't seem to find a way to define these macros on bjam's command line. Putting the changes in "boost/config/user.hpp" doesn't work and errors like the following result: basic_iserializer.cpp .\boost/config/user.hpp(128) : warning C4005: '_SECURE_SCL' : macro redefinition C:\Program Files\Microsoft Visual Studio 9.0\VC\INCLUDE\yvals.h(124) : see previous definition of '_SECURE_SCL' The reason is in this instance is that "basic_iserializer.cpp" contains the following at the top: #include <cassert> #include <cstdlib> // NULL #include <boost/config.hpp> Note that <cassert> is included before <boost/config.hpp>: it's already too late then as including <cassert> results in the macros being defined. I have searched the web and found many "solutions", but I couldn't get any of them to work. Any help would be much appreciated. Steve

2 1

backwards compatibility
by Steve M. Robbins 31 Mar '08

31 Mar '08

Hello, I realize that Boost doesn't promise ABI compatibility across releases, nor even API compatibility according to the FAQ: How can the Boost libraries be used successfully for important projects? Many of the Boost libraries are actively maintained and improved, so **backward compatibility with prior version isn't always possible**. Deal with this by freezing the version of the Boost libraries used by your project. Only upgrade at points in your project's life cycle where a bit of change will not cause problems. Individual bug fixes can always be obtained from the boost repository. (emphasis mine) OK, but *in practice* is there a large amount of work adapting existing client code to a new release? I'm asking in order to gauge whether it is useful for a linux distribution (Debian) to keep multiple versions of Boost available simultaneously; e.g. 1.34.1 and 1.35. Currently Debian provides only the latest version of Boost. Thanks, -Steve

1 0

Solaris SunStudio unsupported [intrusive problems]?
by Zeljko Vrba 31 Mar '08

31 Mar '08

Why has the SunCC support been dropped from Boost 1.35? (In particular, intrusive containers are not usable, for example.)

1 0