- Boost - lists.preview.boost.org

Re: [boost] Generative programming support in boost
by JOAQUIN LOPEZ MU?Z 01 May '04

01 May '04

----- Mensaje original ----- De: Dave Gomboc <dave(a)cs.ualberta.ca> Fecha: Sábado, Mayo 1, 2004 4:04 am Asunto: [boost] Generative programming support in boost > I'm revisiting Czarnecki and Eisenecker's work (Generative > Programming:Methods, Tools, and Applications) from 2000. I'm > wondering what is > available in boost that exploits (or even advances) on the techniques > and code examples given there. > > It occurs to me that both the concept_check and the mpl library have > something to offer in this respect. At the very least, there are few > obvious things (e.g. if, apply_if, etc. in mpl). Is anyone actively > applying boost code in this area? > FWIW, multi_index_container crucially depends on MPL to construct the container out of a typelist of index specifiers (mpl::fold stuff). Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

1 0

RE: [boost] Generative programming support in boost
by Reece Dunn 01 May '04

01 May '04

Dave Gomboc wrote: >I'm revisiting Czarnecki and Eisenecker's work (Generative Programming: >Methods, Tools, and Applications) from 2000. I'm wondering what is >available in boost that exploits (or even advances) on the techniques >and code examples given there. >It occurs to me that both the concept_check and the mpl library have >something to offer in this respect. At the very least, there are few >obvious things (e.g. if, apply_if, etc. in mpl). Is anyone actively >applying boost code in this area? Have you had a look at Boost.Spirit and the related Phoenix/Fusion libraries? Spirit is a lexer/parser generator that allows you to write BNF grammars in C++ that are close to the actual BNF grammar declerations, as well as providing a set of grammar recognisers like real numbers. See the documentation for more information. Spirit allows you to bind actions to the grammar (e.g. summing the numbers read from a file/user input). There is closure support, alliowing a grammar rule to have data bound to it, and the Phoenix/Fusion library provides a mechanism for writing lazily evaluated C++-like code (e.g. if, while as well as STL actions such as push_back and assign). e.g. std::vector< float > v; ... rule<> = real_p[ push_back_p( v ) ] >> ',' >> real_p[ push_back_p( v ) ] NOTE: The Spirit documentation contains this example with a more detailed explanation. There is also the Lambda library - which is currently being rewritten to sit on top of the Fusion framework - that allows you to use Lambda expressions in C++ code, e.g.: std::for_each( c.begin(), c.end(), n += _1 ); // see Lambda docs for correct usage There has also been a review of the FC++ library (see the mail archives for details) that attempts to provide a functional language framework, and although it has not yet been accepted into Boost, there are some powerful concepts like Lazy Lists. It provides its own Lambda implementation and equivalent of Boost.Function/Boost,Bind. Even though this is functional programming (like Haskell) there is a heavy reliance on template metaprogramming as well as adding to the existing generative framework. If you want to find out how extensively the MPL library is used in Boost, do a search for "mpl::". I know that the Serialisation library is using it. Also, there are several workarounds for the current lack of typeof/auto in various libraries that need it. My I/O formatters library (boost-sandbox/outfmt) uses a technique similar to this to map a type to a classification (e.g. container, array, n-ary) in order to build the appropriate formatter so that it can be correctly written to/read from a C++ stream. The documentation is currently out of synch with the implementation and examples, but I am revising it at the moment. Also, the STL provides generative support with iterators/iterator_traits, algorithms and functors. The Boost library has enhanced these with the Iterator Adaptors library, a better iterator_traits and other stuff. Hope that helped. Regards, Reece _________________________________________________________________ Express yourself with cool emoticons - download MSN Messenger today! http://www.msn.co.uk/messenger

1 0

BOOST_TRY, can we make it into CVS?
by Pavel Vozenilek 01 May '04

01 May '04

I would like to ask someone with write access to CVS to create file boost/details/exception_helpers.hpp (or so) with content: #include <boost/config.hpp> #include <boost/detail/workaround.hpp> #if !(defined BOOST_NO_EXCEPTIONS) # define BOOST_TRY { try # define BOOST_CATCH(x) catch(x) # define BOOST_RETHROW throw; # define BOOST_CATCH_END } #else # if BOOST_WORKAROUND(__BORLANDC__, BOOST_TESTED_AT(0x564)) # define BOOST_TRY { if ("") # define BOOST_CATCH(x) else if (!"") # else # define BOOST_TRY { if (true) # define BOOST_CATCH else if (false) # endif # define BOOST_RETHROW # define BOOST_CATCH_END } #endif circular_buffer, serialization and multi_index_container may use this code. It was discussed on the list and reaction was rather positive. TIA /Pavel

2 1

looping construct
by michael toksvig 30 Apr '04

30 Apr '04

A while back, Erik Niebler was testing the waters for inclusion of his very, very clever looping construct into boost, but nothing really came of it. Erik's technique is described here: http://www.nwcpp.org/Meetings/2004/01.html Yup, macros are involved and they are yucky, and yup, there already exists a for_each in the standard that works just fine if you care to move your loop body away into a function, and yup, the name BOOST_FOREACH could use a little love. But there is no denying the notational clarity and succinctness of e.g.: foreach(double &d, container) d += 2; Versus the rather redundant: for (list<double>::iterator i = container.begin(); i !=container.end(); ++i) *i += 2; (Note that the latter stops working if container is represented using a vector or a built-in array instead) /Michael Toksvig

11 52

Can't access repositories using :ssh:
by Victor A. Wagner Jr. 30 Apr '04

30 Apr '04

Starting yesterday (it's still Thursday here) I started getting errors in my update and build scripts. When I chased it down, I was getting: WARNING - POTENTIAL SECURITY BREACH! The server's host key does not match the one PuTTY has cached in the registry. This means that either the server administrator has changed the host key, or you have actually connected to another computer pretending to be the server. The new key fingerprint is: ssh-dss 1024 4c:68:03:d4:5c:58:a6:1d:9d:17:13:24:14:48:ba:99 If you were expecting this change and trust the new key, hit Yes to update PuTTY's cache and continue connecting. If you want to carry on connecting but without updating the cache, hit No. If you want to abandon the connection completely, hit Cancel. Cancel is the ONLY guaranteed safe choice.1¦ I tried it manually a few times, then surrendered and said yes.... that didn't solve the problem now I get: C:\Projects\boost>cvs update -dPA Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Protocol error: 'Directory' missingE Cannot access /cvsroot/boost/CVSROOT No such file or directory I get this behavio(u)r on both boost and boost-sandbox a simpler command yields (logging in just to make sure...): C:\Projects\boost>cvs login Logging in to :ssh:vawjr@cvs.boost.sourceforge.net:22:/cvsroot/boost CVS password: ******* C:\Projects\boost>cvs co -c Cannot access /cvsroot/boost/CVSROOT No such file or directory Victor A. Wagner Jr. http://rudbek.com The five most dangerous words in the English language: "There oughta be a law"

3 3

cycle_iterator update
by Neal D. Becker 30 Apr '04

30 Apr '04

Here is an update to cycle_iterator. This is based on the original design by Gennadiy Rozental. After fairly extensive performance testing I am getting good results (compared to my old implementation which was hand-written, not based on boost iterators), after adding operator[]. There are two versions, both give about the same performance in my tests. One version, which I prefer, modifies the underlying base iterator for operations such as increment. The other implementation never touches the base iterator, but modifies an integral offset value. Also included is a Ring container adaptor class, which adapts a container (e.g., std::vector) to a ring. Basic operations are there, but more can be added. Also a simple test program.

4 9

boost.test XML output error
by Alex Odinaev 30 Apr '04

30 Apr '04

Hi, When I was trying to export test log to XML format I found that unit test framework doesn't convert '&' symbol to appropriate XML sequence. & shows up in log file if you have something like BOOST_CHECK(a&&b). Everything inside () goes to XML directly. Thanks, Alex Odinayev

1 0

bind.hpp needs a cwpro8 workaround
by David Abrahams 30 Apr '04

30 Apr '04

Pro8 is having trouble parsing these templates. Any ideas, Peter? template<class R, class T> inline _mfi::dm<R, T> unwrap(R T::* * pm, int) { return _mfi::dm<R, T>(*pm); } template<class R, class T> inline _mfi::dm<R, T> unwrap(R T::* const * pm, int) { return _mfi::dm<R, T>(*pm); } --- ### mwcc Compiler: # In: C:\boost\boost\bind.hpp # From: named_params_test.cpp # ------------------------------ # 131: template<class R, class T> inline _mfi::dm<R, T> unwrap(R T::* * pm, int) # Error: ^ # ')' expected ### mwcc Compiler: # 132: { # Error: ^ # declaration syntax error ### mwcc Compiler: # 133: return _mfi::dm<R, T>(*pm); # Error: ^^^^^^ # declaration syntax error ### mwcc Compiler: # 136: template<class R, class T> inline _mfi::dm<R, T> unwrap(R T::* const * pm, int) # Error: ^^^^^ # ')' expected ### mwcc Compiler: # 137: { # Error: ^ # declaration syntax error # Too many errors printed, aborting program -- Dave Abrahams Boost Consulting http://www.boost-consulting.com

2 11

[type traits, mpl] small fix for BCB 6.4
by Pavel Vozenilek 30 Apr '04

30 Apr '04

BCB 6.4 doesnt't like when top level namespace is prefixed with :: type_traits/is_const.cpp contains line BOOST_TT_AUX_BOOL_TRAIT_DEF1(is_const,T,::boost::detail::cv_traits_imp<T*>:: is_const) If the leading :: is removed to BOOST_TT_AUX_BOOL_TRAIT_DEF1(is_const,T,boost::detail::cv_traits_imp<T*>::is _const) BCB stops complaining. There are few other places there which may be changed similarly. ---------------- The same problem appears in MPL's value_wknd.hpp: # if !defined(BOOST_MPL_MSVC_60_ETI_BUG) # define BOOST_MPL_AUX_VALUE_WKND(C) ::boost::mpl::aux::value_wknd< C > needs to be changed to # define BOOST_MPL_AUX_VALUE_WKND(C) boost::mpl::aux::value_wknd< C > These problems appeared when trying to compile Collection Traits library. /Pavel

2 2

Can libs/regex/build/gcc.mk use $(AR) instead of just "ar"
by Brad Spencer 30 Apr '04

30 Apr '04

When cross compiling, I need to use an "ar" with the correct target and not just "ar". I've patched boost/libs/regex/build/gcc.mk locally as shown in my attachment, but all I really did was replace use of "ar" with $(AR), which I can then override when invoking make. Perhaps this is reasonable to apply to the mainline copy? [BTW, I'm still locally patching iterator/iterator_facade.hpp because of conflicts with system header macros on i386-pc-solaris2.8. That thread seemed to die.] Thanks! -- ------------------------------------------------------------------ Brad Spencer - spencer(a)infointeractive.com - "It's quite nice..." Systems Architect | InfoInterActive Corp. | A Canadian AOL Company

2 1