- Boost - lists.boost.org

website/docs: link broken
by Markus Werle 25 Jul '12

25 Jul '12

Hi! On http://www.boost.org/doc/libs/1_50_0/libs/scope_exit/doc/html/scope_exit/Ge… g_Started.html There is a link to http://www.boost.org/doc/libs/1_50_0/libs/scope_exit/doc/html/scope_exit/Tu… al.html which is broken. Error message is: 404 Not Found File "/home/grafik/www.boost.org/archives/live/boost_1_50_0/libs/scope_exit/doc /html/scope_exit/Tutorial.html" not found. Unable to find file. Markus

1 0

Re: [boost] Macro for BOOST_NO_NUMERIC_LIMITS_MAX_DIGITS10 or use BOOST_NO_NUMERIC_LIMITS_LOWEST
by Stewart, Robert 24 Jul '12

24 Jul '12

John Maddock wrote: > > >> is there a way to don't include all the deprecated macros? > > Whoa - interesting question. > > I don't know of one. > > > > However, all the deprecated macros are defined (I > > _believe_) in one place, in boost/config/suffix.hpp > > Making it so none of them are defined _should_ be > > straightforward. > > > >Could I suggest to define all of them in a specific > >boost/config/deprecated.hpp file, so that we can add a > >deprecated warning when the file is included. +1 > I think it's too soon for that - as things stand you would > basically be injecting that warning into every Boost header > that includes boost/config.hpp. Aren't there some that have been deprecated for long enough to get that sort of treatment? IOW, create a mechanism for complaining about the use of old names and a policy for how long old names are defined without complaint, how long with complaint, and how long before they are removed. For example, newly renamed macros might be kept in config/suffix.hpp for four releases before moving them to config/deprecated.hpp. Then, names in deprecated.hpp might remain for two releases before removing them altogether. > Besides I don't see what harm defining the old names does? If you can't ever remove the old names, then the set of names that can appear in Boost code grows uncomfortably large. _____ Rob Stewart robert.stewart(a)sig.com Software Engineer using std::disclaimer; Dev Tools & Components Susquehanna International Group, LLP http://www.sig.com ________________________________ IMPORTANT: The information contained in this email and/or its attachments is confidential. If you are not the intended recipient, please notify the sender immediately by reply and immediately delete this message and all its attachments. Any review, use, reproduction, disclosure or dissemination of this message or any attachment by an unintended recipient is strictly prohibited. Neither this message nor any attachment is intended as or should be construed as an offer, solicitation or recommendation to buy or sell any security or other financial instrument. Neither the sender, his or her employer nor any of their respective affiliates makes any warranties as to the completeness or accuracy of any of the information contained herein or that this message or any of its attachments is free of viruses.

2 2

[TypeErasure]
by Markus Werle 24 Jul '12

24 Jul '12

Hi Steven! I have one question: proto and spirit use _1, _2, etc. TypeErasure uses _a, _b, _c ... What was your reason to introduce yet another placeholder set? Markus

4 6

[move] Implementing operator= in terms of pass-by-value and swap
by Krzysztof Czainski 24 Jul '12

24 Jul '12

Hello, Implementing the assignment operator (that is, the copy-, move- and other assignments) in terms of pass-by-value and swap is a common idiom [1]. {{{ T operator=( T b ) { swap(*this,b); return *this; } }}} If T is copy-constructible, the above works as a copy-assign. If T is movable, the above works as a move-assign. And for anything convertible to T the above assign will work (right?). Furthermore, this will also work for movable but not copyable classes (right?). It may not be the optimal version of assignment in some cases. In rare cases it might even be incorrect - for types like std::vector, which aren't expected to free reserved memory upon assignment. But assuming swap is a cheap operation, it should be close to optimal in many if not most cases. And in cases, where RVO kicks in, it may be even better than providing separate overloads of operator= for lvalues and rvalues [2]. Boost.Move docs [3] suggest writing the separate overloads of operator=, and in C++03 move emulation the macro BOOST_COPYABLE_AND_MOVABLE(classname) provides a third overload (necessary for the emulation to work in C++03). Should the user choose to write one overload of operator= in the pass-by-value version, he needs to use a different macro, that doesn't provide the additional overload of operator=. I found such a macro: BOOST_COPYABLE_AND_MOVABLE_ALT(classname), but no docs about it. I have the following suggestions, regarding Boost.Move: 1. Providing a macro, BOOST_MOVABLE(classname), that only provides the two needed conversion operators. Or if that is the purpose of BOOST_COPYABLE_AND_MOVABLE_ALT(classname), I couldn't find docs about it. And to avoid a macro entirely, maybe it is possible to provide a CRTP base class 'movable', that would provide these? 2. Now, that there's a separate macro, defining a movable class, it may make sense to add a separate macro for making a class noncopyable. And we already have a base class for that, if one wants to avoid a macro. 3. It would be nice if the docs [3] mentioned something about the alternative of providing only one operator= passing by value. [1] http://en.wikibooks.org/wiki/More_C%2B%2B_Idioms/Copy-and-swap [2] http://cpp-next.com/archive/2009/08/want-speed-pass-by-value/ [3] http://www.boost.org/doc/libs/1_50_0/doc/html/move/implementing_movable_cla… Regards, Kris

1 0

Re: [boost] [type_erasure] [docs]contrast with boost::any
by Larry Evans 23 Jul '12

23 Jul '12

On 07/23/12 15:37, Larry Evans wrote: > On 07/18/12 00:13, Lorenzo Caminiti wrote: >> Hello all, >> >> *** The review of Steven Watanabe's proposed Boost.TypeErasure library >> begins on July 18, 2012 and ends on July 27, 2012. *** >> > The page: > > http://steven_watanabe.users.sourceforge.net/type_erasure/libs/type_erasure… > > says: > > x is approximately equivalent to a boost::any > > however, it doesn't say how te::any differs from boost::any. > > The attached shows that a boost::any can change it's > type, and, as such, is somewhat like boost::variant. > However, te::any, once it's created, cannot change > it's type, AFAICT. At least that's what the > attached code indicates when compiled with: > > #define SHOW_TE_FIXED_TYPE > A simplified version of that code (which doesn't require demangled_type_name), is attached. The attached also suggests that, when multiple placeholders are in the concept, the unary any *requires* a binding argument. This is shown by trying to compile the attachment with: #define SHOW_BINDINGS_MUST_BE_COMPLETE In contrast, the example here: http://steven_watanabe.users.sourceforge.net/type_erasure/libs/type_erasure… shows no binding is needed because the only placeholder in the concept is the default, _self. HTH. -regards, Larry

1 0

[Review Request] array
by Brian Smith 23 Jul '12

23 Jul '12

Hi All, Based on the assumption that, almost no feedback is good feedback, I'd like to submit the array class for review. The requisite files are hosted at, https://github.com/BrianJSmith/Array Comment: The array class is a C++11 compatible implementation of static multidimensional arrays. Kind Regards Brian

5 12

[TypeErasure] Forward constructors and binded types
by Vicente J. Botet Escriba 23 Jul '12

23 Jul '12

Hi,, In the example from the doc typedef mpl::vector< copy_constructible<_a>, copy_constructible<_b>, copy_constructible<_c>, constructible<_a(const _b&, const _c&)> > construct; std::vector<double> vec; int i = 10; double d = 2.5; tuple<construct, _a&, _b, _c> t(vec, i, d); any<construct, _a> v(get<1>(t), get<2>(t)); I don't see which type is constructed and stored in v and how this type is given. That is I don't see how _a is binded. Could you help me? Best, Vicente

9 32

[TypeErasure] BOOST_TYPE_ERASURE_MEMBER
by Julien Nitard 23 Jul '12

23 Jul '12

Hi Steven, all BOOST_TYPE_ERASURE_MEMBER is only mentioned in the reference, it could be useful to add it to the tutorial: I guess that most user concepts will start from member functions of a given type. A sample on how to specialize it for types not having the said member function would be also welcome. Finally, why did you choose not to include the member's signature in the macro ? Isn't this giving the way for hard to find mistakes ? Regards, Julien

2 2

[website] Server problems
by Daniel James 23 Jul '12

23 Jul '12

Hello, As you might have noticed, many of the php based pages were down this morning. The cause appears to be that the directory '/home/grafik/www.boost.org/' didn't have read and execute permissions for all, so php couldn't read a file within it. I fixed it by giving read permission - nothing in that directory is private so I don't think that will be a problem. Hopefully that's okay? I don't like messing round with these things too much, as I don't know the details about the setup. Does anyone have any idea what changed to cause this issue? Maybe an upgrade or configuration change? Could it use group permissions before but can't now for some reason? Or perhaps something accidentally changed the permission on the directory? It'd be useful to understand what happened. Daniel

1 0

[spirit][phoenix] Spirit v2 conflicts with Phoenix v3
by Andrey Semashev 23 Jul '12

23 Jul '12

Hi, Recently I tried to port my code to Boost.Spirit v2 and Boost.Phoenix v3. After including the following headers: #include <boost/phoenix/core.hpp> #include <boost/spirit/include/qi_core.hpp> I got the following error: In file included from ./boost/spirit/home/phoenix/core.hpp:13:0, from ./boost/spirit/include/phoenix_core.hpp:14, from ./boost/spirit/home/support/make_component.hpp:15, from ./boost/spirit/home/support/meta_compiler.hpp:20, from ./boost/spirit/home/qi/domain.hpp:15, from ./boost/spirit/home/qi/parser.hpp:15, from ./boost/spirit/include/qi_core.hpp:16, from libs/log/src/text_file_backend.cpp:51: ./boost/spirit/home/phoenix/core/actor.hpp:46:16: error: redefinition of ‘struct boost::phoenix::detail::error_expecting_arguments’ ./boost/phoenix/core/actor.hpp:41:16: error: previous definition of ‘struct boost::phoenix::detail::error_expecting_arguments’ ./boost/spirit/home/phoenix/core/actor.hpp:65:12: error: redefinition of ‘struct boost::phoenix::actor<Expr>’ ./boost/phoenix/core/actor.hpp:161:12: error: previous definition of ‘struct boost::phoenix::actor<Expr>’ I take it the problem is that Boost.Spirit always include Phoenix v2 which causes name clashes with v3. Is this a known problem? Shouldn't Boost.Spirit include Phoenix v3 headers instead? I can see v3 redirects includes to v2 if BOOST_PHOENIX_USE_V2_OVER_V3 macro is defined (BTW, there's an error in boost/phoenix/stl.hpp: BOOST_PHOENIX_USE_V2 macro is checked instead). This happens in the release branch.

2 6