- Boost - lists.boost.org

Re: [boost] [interest] Egg, function object builders and adaptors
by dan marsden 18 Jan '08

18 Jan '08

Giovanni Piero Deretta wrote: >> Kirit Sælensminde wrote: >> >> >> If interested enough, it might be enqueued to the review queue. >> >> Anyway, comments are most welcome. >> > >> > I do hope so. I'm surprised that nobody has replied to queue it up >> > already :( >> >> For the record, I and a couple more, have expressed interest already >> in a previous thread. > >I didn't reply to this specific thread, but yes, count me definitely interested. > Just to make sure, I'm definitely interested as well. Looks excellent. Cheers Dan ___________________________________________________________ Support the World Aids Awareness campaign this month with Yahoo! For Good http://uk.promotions.yahoo.com/forgood/

1 0

[serialization] failures of test_exported* tests with gcc 4.1 and higher
by Boris Gubenko 18 Jan '08

18 Jan '08

x.cpp below reproduces failures of test_exported* tests on several platforms using gcc 4.1 and higher. The tests fail with: terminate called after throwing an instance of 'boost::archive::archive_exception' what(): unregistered void cast When compiled with macro WORKS defined, i.e. when destructor of the base class is removed, reproducer succeeds. Does it give any clue? Thanks, Boris x.cpp ----- #include <fstream> #include <boost/archive/binary_oarchive.hpp> struct polymorphic_base { template<class Archive> void serialize(Archive&, const unsigned int) {} #ifndef WORKS virtual ~polymorphic_base(){} #endif }; struct polymorphic_derived : polymorphic_base { template<class Archive> void serialize(Archive&, const unsigned int) {} }; int main() { std::ofstream os("x.x"); boost::archive::binary_oarchive oa(os); polymorphic_base *rb = new polymorphic_derived; oa << BOOST_SERIALIZATION_NVP(rb); }

2 1

the Switch library by Steven Watanabe has been accepted provisionally
by Stjepan Rajko 18 Jan '08

18 Jan '08

The Switch library, submitted by Steven Watanabe, has been accepted provisionally pending a mini-review to be scheduled when possible. The purpose of the mini-review will be to assure that the suggestions outlined below have been adequately addressed. I would like to again thank Steven for submitting a very useful library, and for diligently working with the reviewers on improving its design. I would also like to thank Tobias Schwinger, Joel de Guzman, Darren Cook, Daniel (dherring@...), Felipe Magno de Almeida, and Paul Bristow for providing reviews, votes and half votes. Finally, thanks also to Dan Marsden, Larry Evans, Alexander Nasonov, Richard (legalize@...) and Mathias Gaunard for contributing to the discussion. SUMMARY OF VOTES: Tobias Schwinger and Joel de Guzman have submitted full reviews with positive votes: http://article.gmane.org/gmane.comp.lib.boost.user/32637 http://article.gmane.org/gmane.comp.lib.boost.user/32634 (Joel eventually converted his vote to "yes") One "yes" (half) vote was submitted by Paul Bristow: http://article.gmane.org/gmane.comp.lib.boost.devel/170069 One "no" (half) vote was submitted by Felipe Magno de Almeida: http://article.gmane.org/gmane.comp.lib.boost.devel/170051 Another "no" came privately - the point there was that the library didn't undergo enough scrutiny before submission, and that it should be resubmitted. Another privately submitted "no" came soon after the review opened, stating that the reviewer was not able to deduce any usability/usefulness from the library documentation, and suggested there should be a motivating example on the first page. RATIONALE FOR ACCEPTANCE: I believe the low number of full, positive reviews is justified by the relatively narrow scope that this library covers. Within that scope, however, reviewers indicated that the library is extremely useful. Also, the reviewers submitting positive votes have indicated (and proven) significant expertise in the area, leading me to place much weight on their votes. On the other hand, the negative votes did not point out any concrete flaws in the library itself that would merit rejection. RATIONALE FOR REQUIRING A MINI-REVIEW The "no" votes did, on the other hand, indicate some skepticism of the quality of the library/documentation as submitted, and the progress of the redesign that was happening during the review. Also, as a result of the review, the potential scope of the library has expanded significantly from what was originally proposed (which mostly focused on the ability to choose the number of cases at run-time). I believe that a combination of these two give sufficient reason for a mini-review - doubts about the library can be resolved, and those that might become prospective users of the library under the expanded scope will get a chance to give their opinion. I volunteered to manage the mini-review, so once the library is ready it will hopefully not remain in the review queue for too long (as it seems that finding review managers is the biggest problem ATM). To keep the burden on the community low, I propose to treat the mini-review as a partial continuation of the current review - I will consider the "yes" votes by Tobias and Joel to be left standing unless they state otherwise, and will require any "no" votes to be (re)justified given the new condition of the library. SUGGESTED CHANGES TO THE LIBRARY: * Documentation All reviewers have noted that the documentation could be improved. Specific suggestions were: * motivating example on the first page * additional examples * performance stats * more explanation of the types involved and how they are used. As the scope for the library has increased as a consequence of the review, a gentler, more detailed documentation approach might benefit the library - I believe that many people might find the new library useful, given the right motivation and examples in the documentation. * Return type Tobias Schwinger suggested that the return type should be specified explicitly in the switch_ call, rather than inferring it from the function object, and this seems to be the status quo in the newer designs. Joel de Guzman also suggested that if a version of switch_ should infer the return type, that return type should be a Boost.Variant of the types returned by each of the cases. * Default default behavior If the default behavior is left unspecified, switch_ should invoke the default constructor of the return type in the default case, rather than throwing an exception. This was suggested by Tobias Schwinger after the issue was raised by Dan Marsden. Alexander Nasonov has also indicated that a throwing default case can cause inefficiency, even when it is guaranteed not to be invoked. * Fall-through Alexander Nasonov pointed out the lack of fall-through capability in the submitted implementation. Fall-through has been incorporated into the recent designs being discussed. * Allowing void-returning function objects for the default case with a non-void returning switch Tobias Schwinger mentioned the possibility of allowing default case function objects of void return type even when the switch return type is non-void, as long as the function object call does not return. This would be useful, for example, if the function object is guaranteed to throw. After some discussion on whether determining if the return type is void can be done efficiently, it looks like the proposed solution is to use a special return type that indicates that the call will not return. Since Tobias indicated that this feature is a minor point, it could be implemented or not, depending on how well it fits into the final design. * Design There has been a lengthy discussion on the underlying design of the library, what it should offer, and whether the name of the library is appropriate. There are two types of designs that have been put forward as being appropriate and/or necessary for a Boost.Switch library. On one hand, there is the submitted design (labeled "A" in recent postings), which takes the switch cases in an MPL sequence and a single function object containing implementations of all cases. Joel de Guzman has proposed another design (labeled "B"), in which the switch cases are specified separately and allow a different function to be used for each case. Design "A" is better suited for situations in which the number of cases is determined by a compile time constant, while "B" is more appropriate for many use cases, and provides more flexibility and a more familiar syntax. The result of the discussion was that the Boost.Switch library should cover both situations, and an interface is now being developed that can be used in different ways for different uses (based on a common concept capturing case sequences). Thanks again to all that have contributed to this review, and congratulations to Steven for the accepted library. As others have stated, there is much confidence that Steven will meet the challenges successfully for the mini-review, which we will schedule as soon as the library is ready. Regards, Stjepan

2 2

interprocess
by Jan 17 Jan '08

17 Jan '08

Hi, I noticed the documentation has been updated in the last few months. How close is interprocess to been included in a release of boost? Thanks

2 2

Wiki Defaced
by Joel 17 Jan '08

17 Jan '08

Folks, A spammer just defaced the front of the wiki... http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl Joel

1 0

Attention gmane users
by Phil Endecott 17 Jan '08

17 Jan '08

I know that some people read this list via the gmane web interface, and I thought I ought to point out that http://news.gmane.org/gmane.comp.lib.boost.devel has been missing some messages recently; in particular it has none of the singleton review thread after the announcement. It is visible at http://blog.gmane.org/gmane.comp.lib.boost.devel, however. Phil.

2 1

Facebook update.
by Jan 17 Jan '08

17 Jan '08

We now have 80 members on facebook, this can serve as a good recruiting ground for new C++ talent. Boost/C++ group on facebook: http://www.facebook.com/group.php?gid=4200526117 Thanks

1 0

[Accumulators] Documentation
by Edward Diener 17 Jan '08

17 Jan '08

Attempting to build accumulators documentation from the latest SVN head by executing bjam from the /boost/libs/accumulators/doc directory yields: error: Unable to find file or target named error: 'object(file-target)@476' error: referred from project at error: /E:/Programming/VersionControl/Subversion/Boost/tools/quickbook'

3 3

Boost.Tests: Unit test defines own main() function, why?
by Jens Seidel 17 Jan '08

17 Jan '08

Hi, I would like to know why Boost's unit test framework requires that the user defines init_unit_test_suite(int, char**) (or with char *[] argument???) instead of main(). I recently had again some trouble because of it: * It broke the general autoconf test in boost.m4 as the existance of a Boost library can normally easily be tested by using boost code in main() and not init_unit_test_suite(). (Will probably be fixed soon in boost.m4.) * According to http://lists.boost.org/Archives/boost/2007/05/121601.php linkers don't look into libraries (either static or dynamic) to find main so that unit tests cannot be cimpiled as shared library. * I noticed that Debian's Boost library doesn't work for me as main() is not found in the shared library (but in the static one). I found now beside my bug report http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=456863 also http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=432921 which mainly refers to the link above. I'm not sure whether my problem is an error in Debian's packages or not but I ask why is main handled differently in Boost.test? Wouldn't it be sufficient to ask the user simple to start init_unit_test_suite() from a user specified main() to make most problems go away? Maybe it would be even OK to just start init_unit_test_suite() in the constructor of a static variable (yes, I know that the initialisation order may cause trouble with other static data). http://www.boost.org/libs/test/doc/faq.html doesn't explain it! I will now use ifdef's to easily allow using boost/test/included/unit_test.hpp as alternative to the library. Jens

10 52

[interest] Egg, function object builders and adaptors
by shunsuke 17 Jan '08

17 Jan '08

Hi, all. In a recent discussion, a library named Egg possibly left a good impression on boosters, so I propose it. Introduction: It is not so easy to define FunctionObjects especially if you want to support Boost.ResultOf and Boost.Lambda. Therefore, as Boost provides iterator_facade and iterator adaptors, Egg provides function_facade and function adaptors. Egg provides the following features: * Helpers to build FunctionObject: * egg::function and egg::function_facade provides the way to build FunctionObject which supports Boost.ResultOf and Lambda. * FunctionObject Adaptors(in other words, higher-order functions): * egg::curryN supports the currying. * egg::pipable emulates extension methods(C#). * egg::fuse emulates variadic templates in very limited way. * etc... The zipped source code is in Vault/FunctionObjects: http://tinyurl.com/34lgda The document is on line: http://p-stade.sourceforge.net/boost/libs/egg/ svn repository: https://p-stade.svn.sourceforge.net/svnroot/p-stade/trunk/boost Tested under: trunk and Boost1.34.1. An example of currying in Egg: int plus(int x, int y) { return x + y; } typedef result_of_curry2<int (*)(int, int)>::type T_curried_plus; T_curried_plus const curried_plus = BOOST_EGG_CURRY2(&::plus); void test() { BOOST_CHECK( curry2(&::plus)(4)(9) == 13 ); BOOST_CHECK( curried_plus(4)(9) == plus(4, 9) ); BOOST_CHECK( uncurry(curry2(plus))(4, 9) == 13 ); } If interested enough, it might be enqueued to the review queue. Anyway, comments are most welcome. Regards, -- Shunsuke Sogame

6 9