- Boost - lists.boost.org

[math][pool]random][rational]: (somewhat) different implementations of the same algorithm.
by Yuriy Koblents-Mishke 08 May '06

08 May '06

1. The Euclide algorithm for the greatest common divisor (gcd) currently is implemented by at least four different boost libraries. Is it a good idea to edit the libraries to replace the redundant codes by #include? 2. I have few codes implementing different variants of the Stein algorithm for gcd. The codes, as tested using gcc 3.4.2 / mingw with the the optimization flag O3 under Win XP on AMD Athlon 64 3500+, are generally faster than the Euclide codes in Boost. One of the codes is about 90% faster. Is it a good idea to put the codes in Boost?

5 9

Sun Studio 10
by Mark Wilson 08 May '06

08 May '06

Has anyone built for Sun Studio 10 on Solaris (SPARC)?

2 1

[property tree review] No seperator additions
by Jarrad Waterloo 08 May '06

08 May '06

It would be nice if there were 'no seperator' variants of the the get and out methods that took an array of strings, 'const tr1_or_boost::array<key_type,T>& path', instead of 'const key_type &path'. Having this makes the api a little more object oriented instead of paths needing parsing. It can make it easier to use in more advanced scenarios as the user need not know what the seperator is and what to do if string data has that seperator. I am not saying any of the existing functions are wrong. Frankly, I like them. However, the users of this api could benefit by these minor additions. On a side note the link to some of the examples such a changing the value type seems to be broken. I would definitely like to see what I can do with the ptree_trait to allow using boost::any or the optimized cdiggins::any instead of string. // No separator template<class Type> Type get(const tr1_or_boost::array<key_type,T>& path, const std::locale &loc = std::locale()) const; // No separator template<class Type> Type get(const tr1_or_boost::array<key_type,T>& path, const Type &default_value, const std::locale &loc = std::locale()) const; template<class CharType> std::basic_string<CharType> get(const tr1_or_boost::array<key_type,T>& path, const CharType *default_value, const std::locale &loc = std::locale()) const; // No separator template<class Type> boost::optional<Type> get_optional(const tr1_or_boost::array<key_type,T>& path, const std::locale &loc = std::locale()) const; // No separator template<class Type> basic_ptree &put(const tr1_or_boost::array<key_type,T>& path, const Type &value, bool do_not_replace = false, const std::locale &loc = std::locale());

3 3

[date_time] Suggestion for duration division operator
by Beth Jacobson 08 May '06

08 May '06

Time durations can be divided by an integer to produce a duration (e.g. hours(1)/60 == minutes(1);), but can't be divided by another duration to produce an integer (e.g. hours(1)/minutes(1) == 60;). Could this be added to time_duration.hpp: int operator/(const duration_type& d) const { return (ticks_ / d.ticks_); } or maybe better, this: double operator/(const duration_type& d) const { return (static_cast<double>(ticks_) / d.ticks_); } The same thing can already be accomplished already with duration1.ticks()/duration2.ticks(), but duration1/duration2 seems more natural and fits well with the rest of the date_time operators. Beth

3 5

[SoC] what do we do when we have multiple applicants for projects
by Thorsten Ottosen 08 May '06

08 May '06

Dear List and SoC interested, I have received emails from several people that would like to do my Associtive Container project. What should we try to do in that case? Should we ecourage them all to write the same library or should we try to make up similar projects? How do we decide on who gets to what if only one may do the project? Thanks -Thorsten

3 2

Re: [boost] [asio] Draft version uploaded to vault
by Christopher Kohlhoff 08 May '06

08 May '06

I have just uploaded a new draft version of Boost.Asio to the vault ("Input - Output" directory). The changes since the last draft are: - Emulation of getaddrinfo/getnameinfo has been added. This is used to provide thread safety implementations on Mac OS X (since the supplied getaddrinfo and getnameinfo are not threadsafe on this platform), and to allow code built for Windows to transparently handle the lack of IPv6 when run on Windows 2000 or earlier. - The custom memory allocation changes discussed in this thread have been made. I.e. added size parameter to asio_handler_deallocate function, moved default implementations into boost::asio namespace. - Always call the error_handler function object even if a function succeeds. This is needed to ensure that assign_error always updates the specified error object, as used in daytime1/client.cpp tutorial program. - Various bug fixes. Cheers, Chris

2 2

Re: [boost] [Logging] logger as class's member
by Silex 08 May '06

08 May '06

I'm not sure you can do (string)"A.". Try ::createlog(string("A.") + get_name()); On 5/8/06, mos <mmosquito(a)163.com> wrote: > > Hi! > I have a class A, and A have several instance, each instance I want have > a logger (not all of A share a logger), what can I do. > I had throught like this: > > class A{ > logger m_log; > const char* get_name(); //like tom,jarry... > void CreateLog() { > m_log = ::createlog((string)"A."+get_name()); > } > } > > but seems can not get there, what can I do? > Thanks in advance for your help. > mos. > > > _______________________________________________ > Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost >

2 1

quiche receipe
by Larry Evans 08 May '06

08 May '06

Should be attached. Enjoy! Ingredients: 1 lb. Kraft mexican cheddar Jack natural finely shredded cheddar and monterrey-Jack cheese 1 lb. cottage cheese (small curd) 6 eggs 1/2 cup Bisquick Directions: Mix all the above into 9 x 13 in. floured pan or pan covered with cooking spray. Bake at 400 deg. for 15 min. Bake at 350 deg. for 40 min. or until knife comes out clean when inserted in middle of pan.

4 4

Status of BCB2006 support
by Nicola Musatti 08 May '06

08 May '06

Hallo, this is really meant for Alisdair, but I thought that others might be interested. As far as I'm concerned I believe there's not much more that I can do in the 1.34 timeframe, so I consider the remaining failures as 'expected'. Before looking into how to actually mark them as such I wanted to know if anybody is still working on Borland support for BCB2006. Cheers, Nicola Musatti

3 3

Deprecation and Boost
by AlisdairM 08 May '06

08 May '06

It has been agreed that several platforms will be deprecated in this version of Boost. To support that I have added a warning to the Borland config to alert anyone using old versions of this platform that it is being deprecated and this is the last compatible release. What I notice is the warning gets annoying pretty quickly. I propose to wrap the warning in the following test: #if __BORLANDC__ <= 0x581 && !defined(BOOST_ACKNOWLEDGE_DEPRECATED_1_34) # pragma message "..."; #endif This will allow users to disable the warning by defining the macro in their environment. I would like to extend this to a general policy, to make deprecation more visible were it occurs. Using the BOOST_ACKNOWLDGE_DEPRECATED test we could deprecate libraries or functions, as well as platforms. The test is a simple 'does this macro exist' rather than testing for a value, so should be compatible with toolchains that allow defining macros, but not giving them values. The macro is tied to the specific version that is deprecated, so in later releases we should grep for BOOST_DEPRECATED and either update the version, or do something stronger such as use #error (platform) or remove the code entirely (library) If this meets approval I will try to write up some more formal documentation for early next week, and provide something for the 1.34 release notes. For reference I believe we agreed to deprecate the following platforms in this release: borland < 5.8 [Recommend we only support 5.82] gcc 2.95 MSVC < 7.1 CW 8.3 -- AlisdairM

10 27