July 2008 - Boost - lists.preview.boost.org

[functional/hash] Failing link_ext_test on Windows both trunk and release
by Beman Dawes 18 Jul '08

18 Jul '08

link_ext_test has a link failure on Windows both trunk and release. See http://beta.boost.org/development/tests/release/developer/functional-hash.h… We need to clear this ASAP. --Beman PS: The error is an unresolved external to "test_main" in Boost.Test, so this may be a Jamfile problem rather than a problem with the library's actual code.

2 1

Re: [boost] Is our subversion server/archive 1.5 yet?
by Douglas Gregor 18 Jul '08

18 Jul '08

On Sun, 2008-06-29 at 16:12 -0400, Daryle Walker wrote: > Last week, a was bouncing around sites and found out that Subversion > 1.5.0 has gone final. I've downloaded it because I want to try its > built-in merging support (to fix up Boost.Integer in a branch). Has > our subversion system been updated to 1.5? No, we're still using Subversion 1.4 on the server. > Obviously, if the client and server use subversion versions 1.x and > 1.y respectively, the transactions will proceed as the lower of those > versions. Therefore, both individual clients and the server must be > at least the 1.5 level to use built-in merging. There's another > requirement, though: if the server is upgraded to 1.5, the > repositories it manages aren't automatically upgraded. If a > repository isn't upgraded to the 1.5 format, then the server will act > in 1.4-compatibility mode as needed. > > Is there a schedule for these upgrades to be done (if not done > already)? Is it possible to tell from the client side which mode a > server/repository uses? I actually don't know how to tell from the client-side which version of Subversion the server is running. As for the upgrade... we're looking into Subversion 1.5, but it appears to be a nontrivial process to perform the upgrade and migrate svnmerge.py merge information appropriately, so we need to read more, do some test runs, check that other projects haven't run into serious problems, etc. So much depends on the Subversion repository that we need to be careful. When we're ready to move, we'll send out an e-mail to the mailing list. - Doug

6 8

[fusion] pair documentation issues
by Steven Watanabe 17 Jul '08

17 Jul '08

AMDG The documentation of fusion::pair has a number of problems http://www.boost.org/doc/libs/1_35_0/libs/fusion/doc/html/fusion/support/pa… template <typename First, typename Second> struct first; Should be template<class Pair> struct first; Similarly for second. Also, there is a missing ` in the expression semantics table, which looks horrible as a result. In Christ, Steven Watanabe

2 1

[system] documentation bug, error_code::value_type
by Arvid Norberg 17 Jul '08

17 Jul '08

The boost.system documentation mentions a value_type typedef in boost::system::error_code several times in: http://www.boost.org/doc/libs/1_35_0/libs/system/doc/reference.html#Class-e… In ther header file there is no such typedef, just an int is used. I've filed a ticket here: http://svn.boost.org/trac/boost/ticket/2125 -- Arvid Norberg

1 0

[lambda] Composing lambda expressions
by Robert Jones 17 Jul '08

17 Jul '08

Given a bit of code like this, struct S { bool pred( ) const; }; typdef vector<pair<string, S *> > V; V :: const_iterator i; i -> second -> pred( ); I'd like to rewrite the last line as a lambda expression. Looking at the documentation, I think this is done with bind( & S :: pred, bind( & V :: value_type :: second, _1 ) )( * i ); which works. But I'd rather be writing something like ( _1 -> second -> pred )( * i ); Am I right in thinking that Boost.Lambda just doesn't work like this, and the frst form is how it's done? Thanks, Rob.

3 4

[fusion] constructing all elements of the fusion vector with the same argument
by Stjepan Rajko 17 Jul '08

17 Jul '08

Hello, I recently ran into a need to initialize a fusion vector with a single argument which is then used to construct all of the vector elements. For example, something that would allow: fusion::vector<int, float, double> ones(initialize_all_with(),1); // at_c<0>(ones) == 1 // at_c<1>(ones) == 1.0f // at_c<2>(ones) == 1.0; I hacked fusion::vector so that the above works. Is this addition of interest to anyone else? There is another (perhaps more elegant) solution in implementing a new fusion "array" Sequence (one where all element types are identical) which can construct all elements from a single constructor argument, and then using that sequence to do the above initialization. E.g., fusion::vector<int, float, double> ones(fusion::array<int, 3>(1)); ... or (another solution) a make_array function that returns a fusion::vector: fusion::vector<int, float, double> ones(fusion::make_array<3>(1)); It would be great if one of these solutions (or a different solution) were added to fusion. I can take a stab at putting something together, if any of them are acceptable. That is, unless this is already possible in some way that I can't see at the moment :-) Regards, Stjepan

5 12

Possible MingW ASIO bug with Boost 1.35.0 g++ 4.2.1
by Jason Turner 17 Jul '08

17 Jul '08

Some ASIO test code is crashing in a way that I do not understand. GDB is unable to debug it (no stack available on crash) and the error is a rather obtuse "This application has requested the Runtime to terminate it in an unusual way. Please contact the application's support team for more information." Test code which crashes: #include <boost/asio/ip/address.hpp> #include <iostream> int main() { std::cout << boost::asio::ip::address::from_string("1.1.2.3") << std::endl; } By changing the asio file included, the code no longer crashes: #include <boost/asio.hpp> #include <iostream> int main() { std::cout << boost::asio::ip::address::from_string("1.1.2.3") << std::endl; } This is windowsxp professional, gcc version 4.2.1-dw2. I'm unable to test this on Linux at the moment, so I do not know if it affects other platforms. -Jason -- http://emptycrate.com Games, Programming, Travel & other stuff

2 2

[iochain] mutable_buffer issues
by Robert Dailey 17 Jul '08

17 Jul '08

Assume the following code, which I wrote in an attempt to test very basic functionality of Boost.Iostream: namespace io = boost::io; void foo() { std::vector<boost::uint8_t> data; io::source<io::iobyte> src( io::chain( io::read_mem( data ) ) ); io::iobyte rcvbuffer; const io::mutable_buffer<io::iobyte> mutbuffer( &rcvbuffer, sizeof( rcvbuffer ) ); src.read( mutbuffer ); } The first issue I have is with mutable_buffer. Notice how it requires a template argument that *must* be the exact type passed into the io::source template parameter, in this case io::iobyte. Passing any other type in here results in lots of compiler errors on MSVC 9. Basically my biggest issue with this is that I can't pass an "int" into the mutable_buffer here if I want to read 4 bytes, because the type passed in must match io::iobyte. In my opinion, mutable_buffer should accept *any* type and just read in the number of bytes requested. Without this I fail to see the usefulness of mutable_buffer. The second issue has to do with the type of the first parameter in io::source::read(). Notice it takes a "const mutable_buffer&". Why is it taking a const mutable_buffer if the buffer is mutable? This seems conceptually wrong. I would expect it to take a "mutable_buffer&".

1 1

[math][constants] Documentation?
by Michael Marcin 17 Jul '08

17 Jul '08

I found the math constants library in the 1.35.0 release and it looks like it will be useful to me. I can't find it documented anywhere however. It is not in a detail folder so I assume it is a public library. I've found documentation for a Math Constants library on yahoo groups from what looks like the initial review of this library. http://lists.boost.org/Archives/boost/2003/06/48724.php Is this the same library? Where did the documentation go? Thanks, Michael Marcin

3 2

[iochain] Suggestion for chain
by Robert Dailey 17 Jul '08

17 Jul '08

Hi, Currently I'm doing this: namespace io = boost::io; std::vector<boost::uint8_t> data; io::source<io::iobyte> src( io::chain( io::read_mem( data ) ) ); Notice how I have to specify io::chain() explicitly. Would it be possible for it to be implicitly constructable from a stub? This would allow me to clean up the code a bit to this: namespace io = boost::io; std::vector<boost::uint8_t> data; io::source<io::iobyte> src( io::read_mem( data ) ); Doing this results in compiler errors right now. Finding the solution to the problem (as obvious as it seems) took a long time for me.

1 0