- Boost - lists.preview.boost.org

[msm] history pseudo states
by Christophe Henry 07 Dec '09

07 Dec '09

Hi Michael, >>"To be frank, History is not a completely satisfying concept. First >>of all, there can be just one history pseudostate and only one >>transition may originate from it." >I'll not contend with the views of history yet. They can be a very >powerful concept. What I'm unclear about is the statement that >there can be just one. According to the UML specification there >can be just one per composite state. That of course makes perfect >sense. MSM has SubmachineStates instead of composite states; however, >the same "rule" applies to both. >Does MSM obey this definition or is there a further constraint? >My confusion might be partially attributed to not having finished >reading all of the docs. Sorry for the confusion which I created. Let me try to correct it. Yes of course, I meant one per composite state. And no it doesn't make perfect sense (at least for me) so there'll be a small constraint: History is not a pseudo state for MSM. I find the concept in UML highly dissatisfying. It simply doesn't fit well with orthogonal regions. What if I want to "remember" more than one region? With UML, you're out of luck: one history per composite, one transition originating from it. Want more? Sorry we can't help you. And with deep history, it is even worse. You don't even see for what region of the subcomposite it applies. And I don't see how adding a "*" is supposed to help. So MSM replaced this unfinished and unsatisfying concept with a policy. This allows you much better history configuration and easy deep history. I hope I managed to correct the created confusion. Christophe

2 1

Two FSM libraries, one interface
by Christophe Henry 07 Dec '09

07 Dec '09

Hi David, >Anyway, I just got tired of being a grumpy old passive man, so decided to put some of my money >(or time) where my mouth is. I will write more about this side effect of my analysis in a blog post later I find it great and worth of respect that you spent time making a useful contribution to the discussion by going the extra mile (or even 2-3) to present your point. While I am by no mean a friend of the preprocessor, I find your solution interesting and encourage you to push it a bit further by supporting at least guards. If you allow a few comments: - I suggest you to give the action a name because you won't always manage to name an action actState1EventState2, for example with 2 transitions with the same source and target. - this would also remove the need for direct code writing, which could break the preprocessor - What about adding some entry/exit? I'm sure you manage to pack this one too. Otherwise I find the idea surprising but fun and think that it could gain being pushed a bit further. If you do, you could even tempt me to reuse it in an eUML front-end ;-) Christophe

2 1

[msm] internal/external conflicting transitions
by Christophe Henry 07 Dec '09

07 Dec '09

Hi Michael, Looks like you are thoroughly testing MSM and I find it great :) >The docs have a section on "Conflicting Transitions". It has >two conflicts described: > >1. an event enables multiple transitions because > multiple guards test valid, > >2. an internal and external transition are both enabled > >I see a clear definition for the behavior of #1 in the >discussion; however, it is not clear that #2 will follow >the same rule. I was maybe unclear with "internal" so let's see if we can converge. With internal, I mean internal to a composite. Suppose we have a fsm named Top containing a state State and a composite Sub. Top has a transition on Event from Sub to State. Sub has 2 states, SubState1 and SubState2 and a transition on Event from SubState2 to SubState1. If the fsm is in the state Sub::SubState2 and Event is detected, we now have a conflict between Sub->State and SubState2->SubState1. UML defines this as a conflict and orders that SubState2->SubState1 wins. It also surprised me a bit until I implemented exit points, which are in themselves a conflict (they are defined as a connection between a transition inside a composite and one outside, triggered by the same event). So, if SubState1 were an exit point, it would still work. >It was actually a little surprising to me that the order <of the transition definitions affected #2. (Well, maybe >not surprising ... but not what I wanted). I used >the SimpleTutorial to test the behavior with: > > a_row < Stopped , play , Stopped , &p::test_conflict >, > a_row < Stopped , play , Playing , &p::start_playback >, > a_row < Stopped , open_close , Open , &p::open_drawer >, > >and > > a_row < Stopped , play , Playing , &p::start_playback >, > a_row < Stopped , play , Stopped , &p::test_conflict >, > a_row < Stopped , open_close , Open , &p::open_drawer >, > > >The last transition defined "wins" in a conflict. This is were I'm unsure if I follow you. SimpleTutorial has no composite, so no internal transition so we must be talking about different things. I can, however, still answer about the order of evaluation. UML orders that conflicts be solved by guards which are mutually exclusive, so your machine is currently ill-formed (no guards) and UML says the behavior is undefined and I, as implementer, am allowed to punish you like I wish ;-) Seriously, if you have guards and they are mutually exclusive, your machine is correct and all will work as expected, no matter the order of evaluation. If you still want to know how it works, with pleasure (this is where normal users could stop reading ;-) ). First, for every event used inside a composite, MSM adds a forwarding row at the end of your table (at the end, so they have no influence on the state index calculation system). Then, for a given event, MSM iterates at compile-time through the STT, from back to front and replaces the conflicts by a new row containing all of the conflicts (this is were your compiler suffers). Thus, transitions internal to a composite are given higher priority in the evaluation. This also means that for "normal" conflicts, the last transition wins. I was already asked if I would change it. It is implementation-specific so you should not rely on it, but it is very unlikely that I will ever want to change it. I hope I brought enough elements to bring clarity. If not, please let me know. Christophe

2 1

Re: [boost] Meta State Machine (MSM) review
by Christophe Henry 07 Dec '09

07 Dec '09

Hi Robert, >I wonder if the library can provide a debugging mode that would provide a helpful default implementation >of no_transition that could help identify the unknown transition rather than requiring the user to write something? >Such a function might access internal details, perhaps only provided in the said debugging mode, >that would be clearer about the source state and event for which no transition was provided than would user-level code. Right now you only get an assert in debug mode. I also wonder if more would be possible. About the event, sure, it is already done, about the source state, I can't promise anything, but I will have a look at it. If I understand correctly, you would be satisfied with an assert or error message bringing a bit more debug information? Christophe

2 1

Interprocess regexes
by Mike Spertus 07 Dec '09

07 Dec '09

I would like to store regexes in shared memory, but it seems like Boost regex falls just barely short of supporting this. In the basic_regex documentation, it says "The definition of basic_regex... is based very closely on class basic_string and fulfils the requirements for a constant-container of charT." This may be technically correct, however basic_string and other containers take an allocator template argument, while basic_regex does not. If basic_regex had an allocator it would reap all the same benefits that other string and container types do. In particular, it would be possible to use boost::interprocess to place them in shared memory. Since the allocator would be defaulted to allocator<charT>, I think it would IMO be compatible with the vast majority of existing code (not quite all because template<class charT, class traits> void f(basic_regex<charT, traits>) wouldn't work). Does this make sense? Do others agree that it would make regexes more like other strings and containers as advertised? Thanks, Mike

7 16

Re: [boost] [msm] Review
by Christophe Henry 07 Dec '09

07 Dec '09

Hi Franz, > SC 19.63630 s > Rhapsody 4.95109 s > MSM 0.90754 s Thanks for the review and for the very interesting comparison with Rhapsody, which I found impressive for several reasons: - Rhapsody is, I think, a pretty costly tool (I heard 15k€) so I expected some cool fsm framework and instead the event dispatching really is just a simple switch on states of if/else on event ids, so a pretty basic implementation (though it would be interesting to see how more complicated machines are generated). Of course, as it is made for embedded development, this is understandable. - although this is such basic C-like implementation, Msm is 5 times faster. This is fascinating, the rootState_processEvent is so simple, one can wonder how it is possible to be faster. I find this a good example of the power of metaprogramming. I love C++ :) Thanks, Christophe

17 47

[msm] choice (point) pseudo state
by Christophe Henry 07 Dec '09

07 Dec '09

Hi Michael, >I do not see the concept of a choice point. Did >I miss it? It looks like it can be modeled with >a state having only anonymous transitions and >appropriate guards. Correct, there is currently no choice point as it can be easily emulated and I'm still scratching my head about what to do with it. I'm actually unsure whether and how it should be implemented. It'd be a normal state with an incoming anonymous transition (as you mentioned) and several guards. Even the "else" is optional. So I admit I'm a bit confused if I should simply provide a typedef or if something makes it different from a state. That's the reason why it's not implemented yet. If you or anybody else sees a good use for it and why it's more than a normal state, I'd be quite thankful and could implement it. Christophe

1 0

GGL review (JTF)
by Jonathan Franklin 07 Dec '09

07 Dec '09

=== What is your evaluation of the design? === The design seems to adequately capture important geometric concepts, such as dimensionality and coordinate systems, as well as numerical precision, etc. The 4-type overload for assign (center + radius for a sphere) deviates from the semantics of the 2-/3-arg versions, which I dislike. I have not reviewed any of the extensions, so perhaps there is a good reason doing this... Other than the fact that more people will probably use center+radius than 4-d polygons. Then again, assign is parameterized by the geometry type, so this is probably just an omission. Issues w/ units continue to nag at the back of my mind. However, specifying units for specific geometries (e.g. radians), and requiring consistency of units within a given geometry seems appropriate. Any design issues that would prevent strong robustness guarantees (without drastic refactoring) in the future should be identified and fixed prior to release. I *think* we are okay here, but have not put sufficient effort to make that call (because I don't need it). === What is your evaluation of the implementation? === I didn't have time to review the implementation, other than general poking-around in the code. I was happy to note that some of the functionality already works for k-dimensional euclidean geometries, even though it is technically not supported at the moment. And by k-dimensional, I mean k=4. :-) I didn't have time to push the limits. === What is your evaluation of the documentation? === In a nutshell, the documentation is inadequate, and IMO must be completed prior to release. I'm not sure whether it is best to require a mini-review, or some other process. A few anecdotes that illustrate the problem: The example code on the main page didn't compile, which isn't a problem in and of itself. However, it took an inspection of random code in the separate Examples section, and inductive reasoning to determine that I needed some additional includes beyond ggl.hpp to make the snippets work. I suggest a direct link to a working example program anywhere example code snippets are provided. I also had a bit of difficulty getting several algorithms to work, like union/intersection. The concepts were decently documented, however, translating the concepts to working code took some leg-work. The doxygen documentation seems incomplete. I expected a tab for functions, like union_inserter, but was only able to drill down to the functions by looking at the header docs. An algorithms page seems apropos. However, I couldn't find a page listing the available algorithms. I arrived at the union_inserter page via the doxygen-generated union.hpp header page. Average and worst case performance, as well as any known/suspected numerical issues or other bugs (such as the non-minimal result returned by union_inserter) IMO *must* be listed in the docs for each algorithm. This is the trade-off for not providing a 100% robustness guarantee. Even then, the pre/post conditions for algorithms must be specified, which seemed to be lacking from the docs. A note (personal opinion) regarding Doxygen documentation: Doxygen is best suited for a reference manual. For a User's/Programmer's guide, not so much. === What is your evaluation of the potential usefulness of the library? === This is a generally useful library across *many* domains. In particular, I am excited to make use of this library at work in our graphics/visualization infrastructure code, as well as our actual GIS features. In my academic research, I plan to make use of this library in statistical machine learning/data mining algorithms. However, it is a shame that the library lacks a first-class spatial index capability in it's present state (e.g. what is up for review). Even a relatively naive reference implementation using an R-tree would be useful to me. === Did you try to use the library? === Yup. * With what compiler? g++ (Ubuntu 4.4.1-4ubuntu8) 4.4.1 * Did you have any problems? The internal headers #include <ggl/blah.hpp>, which requires a -I<BOOST_ROOT>/boost directive, in addition to the usual -I required to find the boost headers. This needs to be fixed prior to release, such that only -I<BOOST_ROOT> is required. I am assuming that this is broken for the same reason that Barends chose not to put ggl in the boost namespace. Other than having to use the 'find' command to locate the appropriate headers for inclusion, and other documentation-related issues listed elsewhere in this review, I didn't experience much difficulty using the library. === How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? === On average, a couple hours a day, over the course of a week. Admittedly, I squandered too many hours justifying my own use case on the ML, which may have been better spent reviewing the implementation, and/or writing useful test cases. === Are you knowledgeable about the problem domain? === I have been a casual user (and occasional developer) of Computational Geometry and GIS algorithms at my current job for ~ 4 years. I have also taken a few (recent) graduate-level C.G. and numerical analysis courses. === Miscellany === The existing GGL mailing list(s) should probably be moved to boost/boost-users, and/or a boost-ggl list. Unless there is a compelling reason to maintain an external list (i.e. paid support, etc) , this will best serve the GGL users/developers. === Do you think the library should be accepted as a Boost library? === Yes, contingent on meeting a minimum level of documentation, as determined by the review manager and/or community. Thanks for putting so much work into this library! Jon

5 7

[Review] GGL review starts today, November 5th
by Hartmut Kaiser 07 Dec '09

07 Dec '09

Hi all, The formal review of the Generic Geometry Library (GGL) starts today, November 5, 2009 and will finish November 15, 2009. GGL is being developed by Barend Gehrels, Bruno Lalande and Mateusz Loskot, with substantial input from this Boost mailing list. ------------------------------------------------ About the library: GGL defines concepts for geometries and implements some algorithms on such geometries. GGL is header-only, and can be applied in all software where geometry plays a small or a larger role. Users of the library can use just only one function, such as distance: int a[2] = {1,1}; int b[2] = {2,3}; double d = ggl::distance(a, b); Library users can also use the library in combination with std::vector, boost::tuple's and boost::ranges, such as: std::vector<boost::tuple<double, double, double> > line; line.push_back(boost::make_tuple(1, 2, 3)); line.push_back(boost::make_tuple(4, 5, 6)); line.push_back(boost::make_tuple(7, 8, 9)); double length = ggl::length(line); GGL can also be used in combination with custom or legacy geometries, adapting them to the library specializing traits classes or using registration macro's. Formally, GGL contains a dimension-agnostic, coordinate-system-agnostic and scalable kernel, based on concepts, meta-functions and tag-dispatching. On top of that kernel, algorithms are built: area, length, perimeter, centroid, convex hull, intersection (clipping), within (point in polygon), distance, envelope (bounding box), simplify, transform, convert, and more. The library is also designed to support high precision arithmetic numbers, such as GMP. The GGL might be used in all domains where geometry plays a role: mapping and GIS, gaming, computer graphics and widgets, robotics, astronomy... The core is designed to be as generic as possible and support those domains. However, for now the development has been mostly GIS-oriented. The GGL team proposes an extension model, very similar to what is used by GIL. The proposal as it is in the Boost Sandbox included three extensions: - SVG because it is used in the samples and to generate the documentation - WKT because it is used in the tests - The geographic coordinate system because it is used in some of the examples, also showing how other coordinate systems can be implemented There are more extensions, not included now, a.o.: - Spatial index - Map projections The proposed GGL has seen four previews, many discussions and many changes in design and implementation, based on those discussions. Previews were published on this list January '08, March'08, October'08 and February'09. GGL can be found in the sandbox https://svn.boost.org/svn/boost/sandbox/ggl/formal_review, including source code, examples, unit tests and documentation. Documentation can also be found online: http://geometrylibrary.geodan.nl/formal_review ------------------------------------------------------- Everybody on this list is invited to participate in this formal review. I hope to see your review of this library, including your vote, and I welcome your participation in the discussions on the Boost mailing list. Please always state in your review, whether you think the library should be accepted as a Boost library. Additionally, please consider giving feedback on the following general topics: - What is your evaluation of the design? - What is your evaluation of the implementation? - What is your evaluation of the documentation? - What is your evaluation of the potential usefulness of the library? - Did you try to use the library? With what compiler? Did you have any problems? - How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? - Are you knowledgeable about the problem domain? Regards Hartmut Review Manager ------------------- Meet me at BoostCon http://boostcon.com

20 78

[msm] internal/external conflicting transitions
by Michael Caisse 07 Dec '09

07 Dec '09

Christophe - The docs have a section on "Conflicting Transitions". It has two conflicts described: 1. an event enables multiple transitions because multiple guards test valid, 2. an internal and external transition are both enabled I see a clear definition for the behavior of #1 in the discussion; however, it is not clear that #2 will follow the same rule. It was actually a little surprising to me that the order of the transition definitions affected #2. (Well, maybe not surprising ... but not what I wanted). I used the SimpleTutorial to test the behavior with: a_row < Stopped , play , Stopped , &p::test_conflict >, a_row < Stopped , play , Playing , &p::start_playback >, a_row < Stopped , open_close , Open , &p::open_drawer >, and a_row < Stopped , play , Playing , &p::start_playback >, a_row < Stopped , play , Stopped , &p::test_conflict >, a_row < Stopped , open_close , Open , &p::open_drawer >, The last transition defined "wins" in a conflict. I believe that there is strong historical precedent that internal transitions are always considered before external transitions. I suspect that if this is the behavior that I desire then I can order the transitions so that internals are listed last. Perhaps some additional clarification for case #2 in the docs would be helpful. Thanks - michael -- ---------------------------------- Michael Caisse Object Modeling Designs www.objectmodelingdesigns.com

1 0

[msm] choice (point) pseudo state
by Michael Caisse 07 Dec '09

07 Dec '09

Christophe - I am modifying my code generator to target MSM so that I can evaluate some larger machines and run existing test suites that I use. I do not see the concept of a choice point. Did I miss it? It looks like it can be modeled with a state having only anonymous transitions and appropriate guards. michael -- ---------------------------------- Michael Caisse Object Modeling Designs www.objectmodelingdesigns.com

1 0

[msm] history pseudo states
by Michael Caisse 07 Dec '09

07 Dec '09

Hello Christophe - I am trying to understand your implementation of history. In the MSM documentation it states: "As explained in the BoostCon09 paper <cid:part1.00040908.05020506@objectmodelingdesigns.com>, Histories are weird. First, there can be just one." In the boostcon09 paper it also states [page 5]: "To be frank, History is not a completely satisfying concept. First of all, there can be just one history pseudostate and only one transition may originate from it." I'll not contend with the views of history yet. They can be a very powerful concept. What I'm unclear about is the statement that there can be just one. According to the UML specification there can be just one per composite state. That of course makes perfect sense. MSM has SubmachineStates instead of composite states; however, the same "rule" applies to both. Does MSM obey this definition or is there a further constraint? My confusion might be partially attributed to not having finished reading all of the docs. Thanks - michael -- ---------------------------------- Michael Caisse Object Modeling Designs www.objectmodelingdesigns.com

1 0

[msm] eUML guard/action location
by Michael Caisse 07 Dec '09

07 Dec '09

Christophe - Just a quick question about eUML syntax. I think the concept is great. The placement of the guard and actions doesn't strike me as natural. CurrentState() + cool_event() = DestState() [Guard()]/(action()) There are a couple of reasons for this. One is the conceptual ordering of how the UML machine works. 1. In CurrentState 2. Receive cool_event 3. Check for transitions matching cool_event 4. Found match 5. Verify guards 6. Guards pass 7. Handle CurrentState exit action 8. Take transition and handle transition action 9. Handle DestState entry action 10. Reside in DestState Another reason that I find it unnatural is that the event/guard/action is associated (in my mind) first with the transition leaving the CurrentState. The transition obviously is associated with the DestState but the event/guard portions "occur" while in the CurrentState. UML already (via its notation) pulls together the event and action. What do you think about this instead? CurrentState() + cool_event()[Guard()]/(action()) = DestState() I think it more closely matches the UML syntax. Best regards - michael -- ---------------------------------- Michael Caisse Object Modeling Designs www.objectmodelingdesigns.com

1 0

Two FSM libraries, one interface
by David Bergman 07 Dec '09

07 Dec '09

NO, this is not yet another attempt of mine to irritate people by claiming that the interfaces of Statechart and MSM are similar ;-) When evaluating MSM, both from an interface and performance (read "abuse" ;-)) point of view, I decided to create a common wrapper for both libraries, since I host this awkward notion that the mapping from such a (much more abstract and less verbose) wrapper to the underlying libraries should not vary that much. Be that due to the interfacial commonalities between the two libraries or my great mapping skills :-) Anyway, I just got tired of being a grumpy old passive man, so decided to put some of my money (or time) where my mouth is. I will write more about this side effect of my analysis in a blog post later, as well as provide access to the actual header file. I just want to gauge the interest for such an abstraction layer *for non-complex FSM needs*. I do not claim any commutation, but at least we have a decent sub diagram with arrows from my abstract interface to each one of Statechart and MSM for non-complex cases. I currently imaginatively call this common interface "abstract_fsm.h". It preprocesses a common description to either Statechart or MSM, decided by setting either "USE_SC" or "USE_MSM". It is currently limited to non-nested machines without guards, and only handles the specification, not execution. But, the end result is pretty cool. So, hold on to you LISP hats (there will be parentheses...) and see for yourself. This is a simple FSM app using that abstraction: #include <iostream> #include "abstract_fsm.h" FSM_DEF(Machine, // The type of the FSM (Play)(Stop), // the events // The transitions (including the states) ((Stopped, ((Play, Playing, { std::cout << "starting to play" << std::endl; })) ((Stop, Stopped, { std::cout << "stopping" << std::endl; })))) ((Playing, ((Stop, Stopped, { std::cout << "staying in Stopped" << std::endl; })))) ) int main(int argc, const char* argv[]) { Machine mach; mach.initiate(); mach.process_event(Stop()); mach.process_event(Play()); mach.process_event(Stop()); } Everything is contained in that FSM_DEF construct, including events and transitions (including the source states). One cool thing is that the actions are right there. I have no clue as to the general usability of this abstract layer, but it serves the immediate purpose of testing runtime performance of the two libraries by merely using USE_SC or USE_MSM. This application is preprocessed to this Statechart app (by using -E -DUSE_SC...): #include <boost/statechart/event.hpp> #include <boost/statechart/state_machine.hpp> #include <boost/statechart/simple_state.hpp> #include <boost/statechart/transition.hpp> struct Play : boost::statechart::event<Play> {}; struct Stop : boost::statechart::event<Stop> {}; struct Stopped; struct Playing; struct Machine : boost::statechart::state_machine<Machine, Stopped> { void actStoppedPlayPlaying(const Play& evt) { std::cout << "starting to play" << std::endl; } void actStoppedStopStopped(const Stop& evt) { std::cout << "stopping" << std::endl; } void actPlayingStopStopped(const Stop& evt) { std::cout << "staying in Stopped" << std::endl; } }; struct Stopped : boost::statechart::simple_state<Stopped, Machine> { typedef boost::mpl::list< boost::statechart::transition<Play, Playing, Machine, &Machine::actStoppedPlayPlaying> , boost::statechart::transition<Stop, Stopped, Machine, &Machine::actStoppedStopStopped> > reactions; }; struct Playing : boost::statechart::simple_state<Playing, Machine> { typedef boost::mpl::list< boost::statechart::transition<Stop, Stopped, Machine, &Machine::actPlayingStopStopped> > reactions; }; int main(int argc, const char* argv[]) { Machine mach; mach.initiate(); mach.process_event(Stop()); mach.process_event(Play()); mach.process_event(Stop()); } And when preprocessing to an MSM app (by using -E -DUSE_MSM...): #include <boost/msm/back/state_machine.hpp> #include <boost/msm/front/state_machine_def.hpp> struct Play {}; struct Stop {}; struct Stopped : boost::msm::front::state<> {}; struct Playing : boost::msm::front::state<> {}; struct Machine_ : boost::msm::front::state_machine_def<Machine_> { typedef Stopped initial_state; void actStoppedPlayPlaying(const Play& evt) { std::cout << "starting to play" << std::endl; } void actStoppedStopStopped(const Stop& evt) { std::cout << "stopping" << std::endl; } void actPlayingStopStopped(const Stop& evt) { std::cout << "staying in Stopped" << std::endl; } typedef boost::mpl::vector< a_row<Stopped, Play, Playing, &Machine_::actStoppedPlayPlaying> , a_row<Stopped, Stop, Stopped, &Machine_::actStoppedStopStopped> , a_row<Playing, Stop, Stopped, &Machine_::actPlayingStopStopped> > transition_table; void initiate() {} }; typedef boost::msm::back::state_machine<Machine_> Machine; int main(int argc, const char* argv[]) { Machine mach; mach.initiate(); mach.process_event(Stop()); mach.process_event(Play()); mach.process_event(Stop()); } It is worth noting that once I had wrestled with Boost.Preprocessor and created a mapping to Statechart, the MSM mapping took 14 minutes, and a lot of meta functions were common between the two cases. /David

1 0

Does anyone know what happens to all we write at home?
by Vladimir Batov 07 Dec '09

07 Dec '09

Does anyone know what happens to all we write at home? We habitually put the 'boost' license on such work. However, can your employer claim that *anything* you do belongs to "him" and to slap you with a violation of Confidentiality Agreement or Intellectual Property Agreement? V.

3 2

[rdb, fusion, profile_templates] 0.2.03 - compilation speedup
by Jean-Louis Leroy 06 Dec '09

06 Dec '09

I have uploaded version 0.2.03 to the vault. No new features this time but improvements in compilation speed. The test_odbc project took more than 4.5 minutes to build in debug mode (msvc9, ThinkPad x200s, Intel L9400 with 2 cores @ 1.86 Ghz, 4 Gb RAM) and...17 minutes on a 2Gb RAM computer ! This was becoming unbearable. After a run through profile_templates - thanks to Steven Watanabe for creating this tool ! - I replaced the fusion::maps that were used to gather clause information in sql statement templates with a simpler, less flexible, less orthogonal and less capable version of my own. The result was a 41% speedup (calculated as (old-new)/old) on the 4 Gb computer and 85% on the 2Gb one. The number of template instantiations, as reported by profile_templates, dropped from 86330 to 62771. If someone is interested I can post the complete profiling reports. I plan to focus on datetime support now... Jean-Louis

2 2

Re: [boost] [msm] Review
by Christophe Henry 06 Dec '09

06 Dec '09

Hi David, >>> 2. In Statechart, the definition of a state is coupled to the machine in which it resides >>> while in MSM the state can be defined totally in isolation of the machine, and, thus, reused in other machines >>> 3. In Statechart, the transitions are coupled to the from-state, and defined therein, >>> whereas in MSM, the transitions are coupled to the machine itself; this makes it a bit tricky to reuse a state in Statechart >> >> This is an important interface difference, isn't it? > Yes, which is why I brought it up; see (*) below. Thanks for doing it, I see this as an important point. >* - Yes and no... As I explained in my note to Dave, I did not mean API when I wrote that dreaded sentence about the interface similarity, >I meant it in a more abstract sense. BUT, for simple cases, where we do not deal with transition guards or nested machines, >there actually is a quite close similarity even at a syntactical level. Agreed. But isn't the interface similarity mainly due to the very tightly specified problem domain and to the fact that we all use the excellent MPL tool? Of course, simple state machines using MPL will all look slightly the same, as will do those based on OO-principles with each other. When a new implementation (even based on the same tools) brings important advantages, shouldn't it be considered for inclusion? >They treat guards differently. Yes. I'll come back to it in a minute. >> Let's now add a conflicting transition, which means another possible >> transition on the same event and the same source state, with guards >> deciding which (if any) transition will be taken. >We are now arguably leaving theoretical FSM's and entering pragmatic machines :-) Sure. Theoretical FSM's are anyway about the same for all. My point is to show that differences between MSM and Statechart are subtle but play an important role. >> With MSM you'd have, for example: >> typedef mpl::vector< >> row<State1, Event, State2, &Machine_::action, &Machine_::some_condition>, >> row<State1, Event, State3, &Machine_::action, >> &Machine_::some_other_condition> >> > transition_table; >> >> Could you now write the Statechart version for us? > No ;-) Then, I'm not an expert with Statechart but allow me to try to help ;-) I think (please correct me if I'm wrong) it'd look about like: sc::result react(const Event& evt) { if (some_other_condition(evt)) { action(evt); return transit<State3>();} else if (some_condition(evt) { action(evt); return transit<State2>();} } And this is one important difference, not just a syntactic one. With SC, the structure of the state machine is hidden in the code, which will make your conversion tool harder to write on the direction SC->MSM as the tool will not just be able to parse the code but also understand it. Add a few more actions or other piece of code and I wish your tool luck ;-) OTOH you get more flexibility. So, my point is, there are more differences that syntactic ones, as long as you make more complicated machines. I still agree with you on the point that interfaces are not "widly different" but I put this on the account of a clearly specified problem domain and the fact that most of Boost uses the same MPL tool (due its huge power. Thanks Dave, thanks Aleksey ;-) ). >Yes, there are syntactical differences and decouplings of state vs machine which makes MSM a nicer tool; >and, didi I mention that MSM is in my simple test cases six times faster? You did and I thank you for adding your own tests to the review. As an extra note, if you are ready to give up some state features and need just an enum for the states, you might want to take the FSM in the MPL book. It would probably even be possible to make from it a backend which could at some point partly support the different MSM front-ends. MSM does not yet support this solution in order to deliver a back-end supporting the state features specified by UML (again the restricted problem domain) but it is one of the backends I consider for addition. Do we agree to leave the dreaded sentence behind us and replace it with one saying that both libraries solve the same problem using a wildly different implementation with similarities in the syntax? Christophe

2 1

[chrono] type_traits/common_type and integer/ratio
by Vicente Botet Escriba 06 Dec '09

06 Dec '09

Hi, I'm looking on how to continue with the Boost.Chrono library and I would like to know if * Boost.TypeTraits could take care of the common_type class template. * Boost.Integer could take care of the ratio class template. In this way Boost.Chrono will contain only the time related facilities, duration, time_point and Clocks. In order to minimize the task and been able to purpose as soon as possible the library for review I would like to concentrate on the implementation of the standard proposal and let for a future release the facilities Beman added as * Class template timer, with typedefs: * `system_timer` * `monotonic_timer` * `high_resolution_timer` * Process clocks and timers: * `process_clock`, capturing real, user-CPU, and system-CPU times. * `process_timer`, capturing elapsed real, user-CPU, and system-CPU times. * `run_timer`, convenient reporting of process_timer results. What do you think about this approach? Best, Vicente -- View this message in context: http://old.nabble.com/-chrono--type_traits-common_type-and-integer-ratio-tp… Sent from the Boost - Dev mailing list archive at Nabble.com.

6 19

Re: [boost] [msm] Review
by Christophe Henry 06 Dec '09

06 Dec '09

>> Repeating: the MSM/Statechart overlap in features and interface is quite >> unique to Boost. > > > They definitely overlap in features and also in intent, but my understanding > is that the interfaces are vastly different, and that's an important > distinguishing feature. Am I missing something? No you aren't. At first look, they seem similar, but as soon as you concretely start writing a state machine with both libraries, like Juraj could tell you (he did it), you'll start seeing that the differences are there. Of course, both use states, events and transitions but that's about all. MSM bases itself on the transition table, while with Statechart, the user writes his decisions in the react function. Composites are full state machines with MSM, a simple state with Statechart. MSM uses more metaprogramming, Statechart follows a solution with multiple translation units. Etc. etc. The point is, the interfaces differ, the normal usage even more and the "correct" way to use each (in the authors' eyes) much more as both follow a completely different philosophy, as many heated discussions showed. Christophe

4 4

arithmetic meta functions and traits
by Zachary Turner 06 Dec '09

06 Dec '09

Is there any use for something like this? Following compiles on MSVC 9.0. Despite the simplicity of the final template, it was *NOT* easy to get this working given that there's about 10 ways you should be able to do this according to the standard, but poor SFINAE support makes it next to impossible to get your compiler to accept anything. Since it doesn't do anything highly fancy, I would wager it works on GCC and other compilers as well, although it needs to be tested. struct no_type { }; #define MAKE_BINOP_METAFUNCTIONS(op, binop, traits) \ template<typename T, typename U> no_type operator op (T, U); \ template<typename T, typename U> \ struct traits \ { \ typedef BOOST_TYPEOF(T() op U()) type; \ }; \ template<typename T, typename U> \ struct binop \ { \ private: \ typedef typename traits<T,U>::type result_type; \ public: \ static const bool value = !boost::is_same<result_type,no_type>::value; \ }; MAKE_BINOP_METAFUNCTIONS(+, is_addable, addition_traits); MAKE_BINOP_METAFUNCTIONS(-, is_subtractable, subtraction_traits); MAKE_BINOP_METAFUNCTIONS(/, is_dividable, division_traits); MAKE_BINOP_METAFUNCTIONS(*, is_multipliable, multiplication_traits); Obviously you can define other operators similarly, as well as unary operators. Zach

2 4

Re: [boost] [explore] Library Proposal: Container Streaming
by Brian Wood 06 Dec '09

06 Dec '09

Jeff Flinn writes: > Regardless of the approach taken, I think this would be a useful > facility. I've created a similar, albeit much simplified facility for > two separate projects in the last 5 years. There've been several more > projects that could have used the more general facilities that this lib > offers. I think an on line approach is helpful and we have a service that writes Print functions based on user input -- http://webebenezer.net/cgi-bin/enrobo.cgi. This input: (list<int>) results in this output: void Print(std::ostream& so, const list<int>& abt1) { int indentation = 0; so.fill(' '); so << "\nabt1: "; so << "list has " << abt1.size() << "\n"; indentation += 2; list<int >::const_iterator mediator1 = abt1.begin(); list<int >::const_iterator omega1 = abt1.end(); for (; mediator1 != omega1; ++mediator1) { so.width(indentation); so << " "; so << (*mediator1); so << "\n"; } indentation -= 2; so << std::flush; } Brian Wood http://www.webEbenezer.net <http://www.webebenezer.net/> "The Kingdom of Heaven is like a treasure hidden in the field, which a man found and hid again; and from joy over it he goes and sells all that he has and buys that field."

1 0

MSM Review period extended to Dec 13th
by David Abrahams 06 Dec '09

06 Dec '09

Hi All, We thought this might happen because of the Thanksgiving holiday, and it did: we are extending the review period for the Meta State Machine library until Dec 13th. However, please don't let that delay your review; we'd like to see them all as soon as possible. Msm was written by Christophe Henry, inspired in part by the example in the book "C++ Template Metaprogramming" (http://boostpro.com/mplbook) About the library: Msm (Meta State Machine) is a library allowing the user to easily define high-performance, straightforward and readable UML2 finite state machines, while respecting the principles of efficiency and declarativeness of the original example. To that end, Msm offers several Domain-Specific Embedded Languages that strive to closely emulate a standard UML diagram. Msm supports a broad range of UML features: - state entry/exit - transition actions/guards - composite states - history - orthogonal regions - terminate states - deferred events - explicit entry/fork - entry/exit pseudostates - anonymous transitions - transition conflicts And several non-standard features to make the developer's life easier (interrupt states, query about the current state(s)). Msm (in its current version, 2.0) can be found: * in the sandbox (http://svn.boost.org/svn/boost/sandbox/msm/) * or in the vault: (http://www.boostpro.com/vault/index.php?directory=Msm) Everybody on this list is invited to participate in this formal review. I hope to see your review of Msm, including your vote on acceptance, and I welcome your participation in the ensuing discussions on the Boost developers' mailing list. Please always include in your review a vote as to whether the library should be accepted into Boost. Additionally please consider giving feedback on the following general topics: - What is your evaluation of the design? - What is your evaluation of the implementation? - What is your evaluation of the documentation? - What is your evaluation of the potential usefulness of the library? - Did you try to use the library? - With what compiler? - Did you have any problems? - How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? - Are you knowledgeable about the problem domain? -- David Abrahams BoostPro Computing http://boostpro.com

2 1

Regex allocator support
by Fred Sundvik 06 Dec '09

06 Dec '09

Hi, We recently had to use the regular expression library in conditions where there was no available physical memory, for reporting resource consumption when our console game runs out of memory. This wasn't possible however due to the fact that the basic_regex does some dynamic allocations and doesn't provide any allocator specialization, allthough match_results for example does. So we modified the boost code a little bit and added a new template paramter for the allocator, and made constructors that takes allocators as parameters. This was quite straightforward to do, the allocators just had to be passed forward to the all internal classes it uses. The shared_ptr's had to be changed to use the three parameter constructor, that takes a a deleter and allocator, to prevent them from allocating. The mem_block_cache system was a bit more complicated. Initially we disabled it completely, but that obviously wasn't the best solution. So we figured out, that we can use a custom allocator as default, instead of std::allocator. That custom allocator then had to be specialized for the mem_block_nodes, and the funcion calls to get_mem_block replaced by allocator calls. Our implementation isn't ideal, though as we used mem_block_node as the specialization, and to pass the correct size to the allocator, we allocate BOOST_REGEX_BLOCKSIZE / sizeof(mem_block_node) of them, this way it works for allocators that don't specialize for mem_block_node(but only if the size is dividable by sizeof(mem_block_node) ). Better would probably be to make a new type that is BOOST_REGEX_BLOCKSIZE big. We also found another problem with allocation, the match_result::format functions only takes string_type, that isn't specialized by any allocator as they input. This made it impossible to call them with strings using different allocators. So we added specializations of those functions for different allocators and traits. There are probably more functions like that, but since our code doesn't use them there was no need to change anything. So my question is, what are the chances that theese kinds of modificiations gets implemented in the actual library, and possibly in the next c++ standard library as well? I consider it a quite big design flaw to not have any control of the internal allocations, so for me this a quite big thing. Obviously we are doing well with the modifications we made here internally, but there are other applications too that would benefit, many embedded applications for example. Would our code changes be of any use? The changes are only the minimal needed to make it work in our case, and there's many compiler and option permutations not taken care of, so obviously a lot of work would still be needed. Also there's no additional unit tests made. Fred Sundvik Bugbear Entertainment Ltd.

5 7

Notice: Boost.Atomic (atomic operations library)
by Helge Bahmann 05 Dec '09

05 Dec '09

Hello, as promised I have started extracting an atomic operations library. Current state is available at: http://www.chaoticmind.net/~hcb/projcets/boost.atomic It implements boost::atomic<TYPE> which faithfully mimics std::atomic<TYPE> as specified in the C++0x draft standard. As allowed by the standard, operations transparently fall back to locking when the underlying architecture does not support the requested operation, so the library already contains a "fallback" implementation that works an all platforms (using mutex from boost::thread). It currently natively supports gcc/x86, gcc/powerpc and gcc/alpha (I can vouch for the correctness of the implementations on these targets). It contains some entirely untested support for building implementations from CAS operations on other systems (e.g. _InterlockedCompareExchange on win), so I would greatly appreciate any feedback if it works/doesn't on any particular platform. There is some preliminary documentation, but not in boostdoc format -- after unsuccessfully struggling with bjam/boostbook & friends for a few hours I simply gave up and reverted to trusty old doxygen :( Is there any step-by-step guide on how to create, build and document a new library? I could really use that as the boost build and documentation system is pretty alien to an autotools-accustomed guy like me. Best regards Helge

10 33

Re: [boost] [msm] Review
by Christophe Henry 05 Dec '09

05 Dec '09

Hi David, > It is sad that my comments are construed as criticism of MSM, since they merely point out the (I thought obvious...) >similarities with another library already in Boost. >After having tried MSM for a few days, most of it rocks, and I will provide a review later today. Oh I didn't take it as a criticism of MSM at all, merely of a valid issue which will have to be solved in Boost at some point because, as Boost grows, it is bound to happen again. I'll look at your review with great interest and thank you for taking the time to write it. To the core of your point, which is actually a valid concern. I agree that MSM and Statechart are solving the same problem. To explain what I (and probably Andreas) mean, let's look at the problem from a different angle. Let's suppose we'd build a Statechart 2.0. How would it look like? Would it go for the MSM way, putting as much as possible on the compiler, at the risk of breaking it, or the Statechart way, less metaprogramming and easier on compilers but slowlier and with a less descriptive interface (in my opinion)? Would it go for a transition table way or a state::react way? What kind of state storage would it take? And so on and so on. What is likely is that we'd probably not even manage to agree on the goals, much less on the design or the implementation. MSM comes from my frustration with not managing to make Satechart do what I needed. This shows that the philosophy of both authors is quite different and likely incompatible. Christophe

1 0