- Boost - lists.boost.org

Boost.Function comparisons
by David Abrahams 09 Jun '10

09 Jun '10

Hi, I know this has been a big issue in the distant past and I admit to being too lazy to pore through the old discussion history, but I was reading through the proposed C++0x standard and noticed there was no equality comparison for std::function, so I went to look at Boost. There I saw that boost::function<S> has meaningful equality comparisons with objects of the type stored in the function instance, but not with function<S> itself. And that struck me as very odd. It was clearly quite intentional, since such comparisons are intentionally “poisoned” to make sure nobody uses them. So can someone explain this to me? Whatever the rationale, I think it should be available in the documentation. Thanks, -- Dave Abrahams BoostPro Computing http://www.boostpro.com

3 3

Bug sprint daily update for Wednesday, June 9th
by Marshall Clow 09 Jun '10

09 Jun '10

As of 0800 Pacific time today, we had 894 open trac tickets. That is down by 44 from yesterday, and 95 since the start of the bug sprint! This is great news!! Let's keep up the good work! Total ticket counts: Saturday 989 Sunday 988 Monday 958 Tuesday 938 Wednesday 894 Activity summary for yesterday (I'm counting tickets by hand here, here so blame my crummy math if it doesn't add up) New tickets: 14 Modified: 27 Closed: 59 Reopened: 1 If you're looking to help, and don't know what you can do, there are a whole lot of bugs in the database reported against old versions. You can regress these, see if they still exist, and close them if they do not. Here's a report that shows all open bugs reported against version 1.34.1 (which was released a long, long time ago). <https://svn.boost.org/trac/boost/query?status=assigned&status=new&status=re…> Today, rather than the summary totals for yesterday, I have deltas; changes in the totals. If you'ld like to see the totals, they are available online at: Open ticket count by owner: https://svn.boost.org/trac/boost/report/18 Open ticket count by component: https://svn.boost.org/trac/boost/report/19 Open ticket count by type https://svn.boost.org/trac/boost/report/20 (Overwhelmingly, these are 'bugs') Open ticket count by milestone: https://svn.boost.org/trac/boost/report/21 (All changes are since the start of the bug sprint; 0800 PDT Saturday June 5th) Changed totals by library: thread -27 None -17 website -10 iterator -6 concept_check-5 function -5 multi_array -5 asio -4 random -4 lambda -3 units -3 variant -3 fusion -2 mpl -2 Building Boost -1 GIL -1 Regression Testing -1 date_time -1 filesystem -1 io -1 lexical_cast -1 mpi -1 regex -1 spirit -1 trac / subversion -1 circular_buffer 1 config 1 program_options 1 property_tree 1 serialization 1 utility 1 xpressive 1 Documentation 2 build 4 Changed totals by author: anthonyw -26 <none> -19 dave -9 grafik -9 dgregor -7 no-maintainer -6 garcia -5 chris_kohlhoff -4 ebf -3 jsiek -3 agurtovoy -2 djowel -2 dlwalker -2 matthiasschabel -2 az_sw_dude -1 bemandawes -1 bschaeling -1 danieljames -1 hkaiser -1 hljin -1 nasonov -1 speedsnail -1 steven_watanabe -1 danielw 1 eric_niebler 1 jano_gaspar 1 jewillco 1 matias 1 ramey 1 witt 1 vladimir_prus 7 Changed totals by category: Bugs -71 Feature Requests -10 Patches -7 Tasks -3 Support Requests -2 Library Submissions -1 -- Marshall

1 0

[msm] Choosing initial event in sub machines
by Michael.Herchel＠rohde-schwarz.com 09 Jun '10

09 Jun '10

Hi, I just came along another question regarding state machines and sub machines. With "typedef my_initial_event initial_event" it is possible to customize the initial event. This does not work for nested sub machines, because in those cases msm just forwards the event that triggered the state change so that we end up in the sub machines initial state. Msm does not use the specified initial event of the sub machine to trigger the initial state of the sub machine. Actually, this is exactly the behaviour what we want but is it UML conform? Do you have any idea? Best Regards, Michael

1 0

[fusion] conflict with mpl
by Christophe Henry 09 Jun '10

09 Jun '10

Hi, While looking at the trunk test results for MSM, I realized (with a shock) that many test cases were not working with g++ any more. The reason is listed in (read the last few lines): http://www.boost.org/development/tests/trunk/developer/output/minion-clang-… It appears there is a conflict between mpl and fusion for at. Editing boost/fusion/container/set/detail/deref_impl.hpp from return at<typename It::index>(it.seq->get_data()); to return ::boost::fusion::at<typename It::index>(it.seq->get_data()); Seems to solve the problem in the small example I attached. This stuff is pretty urgent for me because it breaks a lot of msm tests. How to proceed? The Monday deadline is looming ;-) Thanks, Christophe

2 1

[endian] endian flip and endian domain
by vicente.botet 09 Jun '10

09 Jun '10

Hi, I think that I have found a good name that can be used instead of the swap_in_place: "flip". This recall the flip operation used in bitset exchanging the value of a bit. On the endian context this will exchange in place the endianess of a data from/to the native endianess. As the same data can be used with different endianess depending of the the context, the flip function needs a specific parameter, let me call this the endian domain. As the data to flip can containg part that are big and others that are little endian, we need to map all the leaveds of the tree associated to the data structure. The domain mill map staticaly the data to its endianess. For fundamental types this will be reduced to a single endian value, but for structures, the map will be a tree having as leaves endian types. This mapping allows us to separate the endian concern from the data itself in a clear way. For structures that have been adapted as fusion sequences, we can define flip_to/from functions that will do the the flip in a generic way. Of course the user could always overload the flips functions to provide its own implementation. I have started to play with this, and even if I'm not a metaprogramming expert, I have reached to have a prototype implementation that I need to polish a little bit more, before I share it with you. I will need your help later to make it much more elegant. Anyway, let me know what do you think of this design. Best, _____________________ Vicente Juan Botet Escribá http://viboes.blogspot.com/

8 19

Re: [boost] [msm] Nested submachines and explicit state entry
by Christophe Henry 09 Jun '10

09 Jun '10

<snip> >I have made an example which implements the UML state chart provided by Michael. > >well it does not compiles with VC9 :(. > >it compiles as fare as the explicit entry target state is a simple >state but does not if it is a composite sate (e.g.: a sub state >machine). There are several reasons to this and it took me a while to figure out what was happening. The first one is that there are 2 errors in the example: RunningStateMachine::Inner1::direct<RunningStateMachine_::Inner1_> should be: RunningStateMachine::direct<RunningStateMachine_::Inner1> As you are directly entering RunningStateMachine, not Inner1. The second has to do with the fact that msm looks for states in the transition table. In particular, RunningStateMachine has: struct transition_table : mpl::vector< // Start Event Next Action Guard // +---------+-------------+---------+---------------------+----------------------+ a_row < InitialState , event4 , InitialState , &RunningStateMachine_::onEvent4 > // +---------+-------------+---------+---------------------+----------------------+ > {}; There is no Inner1 or Inner2 to be found, so you need to tell msm about them. Add in RunningStateMachine: typedef mpl::vector<Inner1,Inner2> explicit_creation; So now it should compile. But it doesn't because I think VC9 has a bug and seems to look for Inner1 not in RunningStateMachine 's table but in Inner1's, then complains not to find it. MinGW g++4.4 compiles the code fast and correctly. However, VC9 is not the only one with bugs, and if you add Inner1 into RunningStateMachine's table, it will not compile, so msm also has a bug and I simply forgot the case of a submachine being an explicit entry. You just happen not to notice it in this example. I attach you the corrected example in case you use g++ and do not need Inner1 in RunningStateMachine's transition table. So, we'll have to live with this restriction for the moment and I offer you an alternative solution. Replace the explicit entry by an entry point (from a design perspective, I find it better anyway). I attach you this solution too. HTH, Christophe

2 1

Re: [boost] [endian] endian flip and endian domain
by Dave Handley 09 Jun '10

09 Jun '10

Gottlob Frege wrote: > I think what I am saying is that, conceptually at least, the > endian-types come first, and the straight functions are built on top > of that (instead of the other way around as most are suggested). > Now, I wouldn't want that to impact performance, which is why I say > 'conceptually'. We can specialize the heck out of everything so that > performance wins, but I think the conceptual foundation is important. I disagree. If you have an endian type then I think you are coding in from the beginning a need for a copy in order to endian swap - even if the endian swap is a no-operation. Unless I'm missing something here. I also don't think this is bad. Remember that endian swapping happens once at the boundary of the code. It isn't something that is constantly changing type and we shouldn't be worried about the programmer losing track of what type the data is in imho. Of course, if someone can come up with a typed interface which doesn't enforce a copy then I may change my mind. Next, on naming, I think we should be very careful that we don't over-think the naming question. I'll ask a general question. What term does everyone use when talking about endian swapping? Certainly in my programming career I've never heard it called anything other than endian swapping. Also (and I know this shouldn't be taken as a definitive source) Wikipedia refers to the process of changing endianness as a byte swap. Shouldn't we use swap because that is what everyone knows the process of changing endianness as? That's my 2d, but I'm not wholly against another naming option. Dave

3 4

[mpl] #861 Detect nested template members with has_template_xxx
by Steven Watanabe 09 Jun '10

09 Jun '10

AMDG The patch attached to https://svn.boost.org/trac/boost/ticket/861 works with all the compilers I have to test with. It's fairly long, so I'm not including it here. Okay to commit? In Christ, Steven Watanabe

3 4

Bug Sprint daily update for Monday, June 7th
by Marshall Clow 08 Jun '10

08 Jun '10

As of 0800 Pacific time today, we had 958 open trac tickets. That is down by 30 from yesterday, and 31 since the start of the bug sprint. Great work, everyone! Thanks especially to Steven W, Andy W, and Viboes for all their work. Activity summary for yesterday: New tickets: 3 Modified: 38 Closed: 34 Reopened: 1 Today, rather than the summary totals for yesterday, I have deltas; changes in the totals. If you'ld like to see the totals, they are available online at: Open ticket count by owner: https://svn.boost.org/trac/boost/report/18 Open ticket count by component: https://svn.boost.org/trac/boost/report/19 Open ticket count by type https://svn.boost.org/trac/boost/report/20 (Overwhelmingly, these are 'bugs') Open ticket count by milestone: https://svn.boost.org/trac/boost/report/21 (All changes are since the start of the bug sprint; 0800 PDT Saturday June 5th) Changed totals by library: None -16 thread -12 Building Boost -2 date_time -2 GIL -1 asio -1 filesystem -1 function -1 fusion -1 iterator -1 quickbook -1 Documentation 1 circular_buffer 1 foreach 1 interprocess 1 intrusive 1 mpi 1 mpl 1 pool 1 program_options 1 serialization 1 Python 2 utility 2 build 5 Changed totals by author: <none> -25 anthonyw -11 djowel -6 az_sw_dude -2 bemandawes -1 bschaeling -1 chris_kohlhoff -1 dlwalker -1 hljin -1 speedsnail -1 agurtovoy 1 cnewbold 1 danielw 1 eric_niebler 1 jano_gaspar 1 matias 1 niels_dekker 1 no-maintainer 1 pdimov 1 ramey 1 witt 1 igaztanaga 2 danieljames 4 vladimir_prus 9 Changed totals by category: Bugs -17 Feature Requests -5 Patches 1 Support Requests 1 -- Marshall

3 4

[regex, xpressive] interesting(?) perf benchmark
by Eric Niebler 08 Jun '10

08 Jun '10

I'm always suspect of benchmarks, but I figured I'd post this anyway. http://astl.sourceforge.net/bench.7.html. It seems someone wrote a generic automata library and a regex engine on top of it, and compared its performance to boost.regex, xpressive, pcre, and greta. Of course, his library wins handily. (The purpose of benchmarking is to make yourself look good, right? ;-) He doesn't say that (a) his engine is a DFA, and he's comparing to NFAs; and (b) he's only implemented a tiny subset of regex features people would expect; e.g., no captures even, let alone backreferences: things that can be tricky or impossible to implement with DFAs. He also doesn't say what version of Boost he's using (tests run Nov. 2009) or whether he's using static or dynamic xpressive (probably dynamic; static can be ~15% faster). Anyway, it's not my intention to kick off yet another costly escalation in the regex war. God help me, I don't have the time. I post it here because I haven't come across a comparison of these four libraries by a third party before. Some quick observations: - I'm pleased that on the balance, xpressive fares well here relative to the other NFAs. In the end, though, not much separates best from worst. - All the NFA libraries seem to have large overhead that becomes very noticeable when matching short strings. I don't know why. - I'm gobsmacked that my VERY OLD library greta handily beats pcre, boost.regex /and/ xpressive on short matches. Failing to beat yourself in any perf benchmark is kind of embarrassing. - I'm tickled that for long matches, both xpressive and boost.regex (NFAs) beat his library (DFA). How'd that happen? - I find it interesting that both xpressive and boost.regex have large discontinuities in the chart at the bottom; the others don't. Huh. When more std library vendors start shipping regex libraries, I'll be interested in seeing this perf matrix expanded. -- Eric Niebler BoostPro Computing http://www.boostpro.com

10 22