- Boost - lists.boost.org

RE: Boost.Threads: Do we need all those mutexes?
by Batov, Vladimir 16 Jul '04

16 Jul '04

Guys, Could we please resist complicating interfaces when we do not have to? I personally consider an additional blocking parameter for something like try_lock excessive because it distorts try_lock into becoming scoped_lock. Let's keep different concepts different. Blocking in try_lock is an oxymoron -- try_lock cannot block because it is "try". The basic usages (covering about 80%) are: scoped_lock l(m); // Block until it locks. scoped_try_lock l(m); // Try (!) to lock. I'll check the result. scoped_timed_lock l(m, time); // Block until it locks or time runs out. I'll check the result. Simple and clean. The only functionality that I can see beyond that is when I have to only declare a lock but apply it later. Like: Void Foo() { scoped_lock l(m, do_not_lock); if (something) { l.lock(); // Can't declare here. ... } } For that purpose we'll need something like scoped_lock(mutex&, nolock_t); That's it. Isn't it? Alternatively we might consider making timed_lock our basic lock as scoped_lock and try_lock are merely border-cases (time=infinite and time=0 respectively). Then, scoped_lock and try_lock would be real refinements (convenience interfaces) of that general concept. V. Christopher Currie wrote: > Michael Glassford wrote: > > > To address these problems, I would like to redefine the lock concept > > constructors as follows (changes indicated with "^"): > > > > Definitions > > ----------- > > L: lock type > > l: lock variable (type L) > > m: mutex variable (appropriate Mutex type) > > s: initial lock state (enum type {NO_LOCK=0, LOCK=1}) > > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ > > b: block? (enum type blocking_state {NON_BLOCKING=0, BLOCKING=1}) > > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ > > > > Lock Concept > > ------------ > > L l(m,s) //Blocking lock if s > > L l(m) //Blocking lock > > > > > > TryLock Concept > > --------------- > > L l(m,s,b) //Lock if s; blocking if b > > ^ ^^^^^^^^^^^^^ > > L l(m,b) //Lock; blocking if b > > ^ ^^^^^^^^^^^^^ > > > > TimedLock Concept > > ----------------- > > L l(m,s,b) //Lock if s; blocking if b > > ^ ^^^^^^^^^^^^^ > > L l(m,t) //Blocking lock, failing if not obtained by time t > > > > I like this proposal, if for no other reason than the move to enums > makes the arguments clearer. A couple of points: Yes, I like LOCK and NO_LOCK better than true and false for the same reason. > TryLock: What would be the semantics of l(m, NO_LOCK, b)? In other > words, if you're not going to lock on construction, why specify a > blocking policy? The only reason is so that if you write l(m, LOCK, ...) you could also specify the blocking parameter. An expansion of Vladimir Batov's of using structs rather than enums could help here: struct nolock_t {}; nolock_t nolock; struct lock_t {}; lock_t lock; class TryLock { TryLock(TryMutex m, no_lock_t s) {...} TryLock(TryMutex m, lock_t s, blocking_t b) {...} } > L l(m, b) and L l(m, s) seem to cover the two > scenarios: either I don't want to lock, or I do and I need to specify > the blocking policy. > > TimedLock: I'm not sure I follow the reasons why TimedMutex refines > TryMutex. (This is a comment on the original design, not your specific > changed, but bear with me.) I think a timeout of zero seconds achieves > the same goal, even if TimedLock/TimedMutex don't implement the same > operations. > So I don't think that TimedLock needs a constructor that > takes a blocking argument, and ultimately I think TimedLock/TimedMutex > should refine Mutex/Lock, and not TryMutex/TryLock. > Also, the current implementation of TimedLock doesn't support a single > argument constructor, even though it is supposedly a refinement of > TryLock, which does. I'd like to see TimedLock implemented in such a way > that I could use it as if it were a simple Lock, which I currently cannot. > In summary, this is what I'd like to see: > > Lock Concept > --------------- > L l(m) //Blocking lock > L l(m,s) //Blocking lock if s > > TryLock Concept > --------------- > As Lock, *plus:* > L l(m,b) //Lock; blocking if b > > TimedLock Concept > ----------------- > As Lock (not TryLock), *plus:* > L l(m,t) //Blocking lock, failing if not obtained by time t > //If t == 0, acts as TryLock. > As I said in another post, I'll think about this some more. Thanks for taking the time to comment. Mike

15 169

ublas/traits.hpp not standalone
by Neal D. Becker 16 Jul '04

16 Jul '04

ublas/traits.hpp forgot to include some headers. Here is what happens: g++ -c -x c++ /usr/include/boost/numeric/ublas/traits.hpp /usr/include/boost/numeric/ublas/traits.hpp:1428: error: ISO C++ forbids declaration of `forward_iterator_base' with no type /usr/include/boost/numeric/ublas/traits.hpp:1428: error: template-id ` forward_iterator_base<std::forward_iterator_tag, I, T>' used as a declarator /usr/include/boost/numeric/ublas/traits.hpp:1428: error: syntax error before `; ' token /usr/include/boost/numeric/ublas/traits.hpp:1436: error: ISO C++ forbids declaration of `bidirectional_iterator_base' with no type /usr/include/boost/numeric/ublas/traits.hpp:1436: error: template-id ` bidirectional_iterator_base<std::bidirectional_iterator_tag, I, T>' used as a declarator /usr/include/boost/numeric/ublas/traits.hpp:1436: error: syntax error before `; ' token /usr/include/boost/numeric/ublas/traits.hpp:1443: error: ISO C++ forbids declaration of `bidirectional_iterator_base' with no type /usr/include/boost/numeric/ublas/traits.hpp:1443: error: template-id ` bidirectional_iterator_base<boost::numeric::ublas::sparse_bidirectional_iterator_tag, I, T>' used as a declarator /usr/include/boost/numeric/ublas/traits.hpp:1443: error: syntax error before `; ' token /usr/include/boost/numeric/ublas/traits.hpp:1451: error: ISO C++ forbids declaration of `random_access_iterator_base' with no type /usr/include/boost/numeric/ublas/traits.hpp:1451: error: template-id ` random_access_iterator_base<std::bidirectional_iterator_tag, I, T>' used as a declarator /usr/include/boost/numeric/ublas/traits.hpp:1451: error: syntax error before `; ' token /usr/include/boost/numeric/ublas/traits.hpp:1458: error: ISO C++ forbids declaration of `random_access_iterator_base' with no type /usr/include/boost/numeric/ublas/traits.hpp:1458: error: template-id ` random_access_iterator_base<boost::numeric::ublas::packed_random_access_iterator_tag, I, T>' used as a declarator /usr/include/boost/numeric/ublas/traits.hpp:1458: error: syntax error before `; ' token /usr/include/boost/numeric/ublas/traits.hpp:1465: error: ISO C++ forbids declaration of `random_access_iterator_base' with no type /usr/include/boost/numeric/ublas/traits.hpp:1465: error: template-id ` random_access_iterator_base<boost::numeric::ublas::dense_random_access_iterator_tag, I, T>' used as a declarator /usr/include/boost/numeric/ublas/traits.hpp:1465: error: syntax error before `; ' token

1 1

Re: [boost] Regression Tests, Boost.Serialization
by Reece Dunn 16 Jul '04

16 Jul '04

Martin Wille wrote: >Robert Ramey wrote: >>Jeff Garland wrote: >>>>Houston, we have a problem. >>[snip] >>It's a fact - running the bjam serialization test is going to consume 20 >>GB. > >The disk requirements can be changed. Stripping the test binaries >and removing the .o files already has been mentioned. Another >chance lies in merging similar (in terms of instantiated templates) >tests into one binary. It might also be beneficial to allow finer grained results within a test file. Let's say I have a test file stl_containers.cpp that tests std::vector, std::list, etc. Now, if a compiler has a problem with one of these (e.g. std::map) - if I understand the regression test system correctly - it will mark this test as failed in the regression results and a user would assume that all the stl_containers would be unusable! What I would like is to be able to group sets of tests (like in the unit tests), but to show a breakdown of those tests (e.g. for std::vector, std::deque, etc.) so that if one of these groups fail, the user can see that specific container doesn't work for a perticular compiler/platform. Regards, Reece _________________________________________________________________ Express yourself with cool new emoticons http://www.msn.co.uk/specials/myemo

1 0

Boost.Test problem
by Joel de Guzman 16 Jul '04

16 Jul '04

Hi, I'm having some minor problems with Boost.Test. Using g++3.2, MinGW, I get: In file included from ../../../boost/test/included/unit_test_framework.hpp:20, from ../../../libs/spirit/test/actor/unit_test.cpp:9: ../../../libs/test/src/test_tools.cpp: In function `bool boost::test_tools::tt_detail::equal_and_continue_impl(const wchar_t*, const wchar_t*, boost::wrap_stringstream&, boost::unit_test::basic_cstring<const char>, unsigned int, boost::unit_test::log_level)': ../../../libs/test/src/test_tools.cpp:188: `wcscmp' undeclared in namespace ` std' unit_test.cpp is just: #include <boost/test/included/unit_test_framework.hpp> Any hints? Regards, -- Joel de Guzman http://www.boost-consulting.com http://spirit.sf.net

6 24

Re:Re: [boost] Re: Serialization Menu
by Robert Ramey 16 Jul '04

16 Jul '04

Robert Ramey wrote: >Mattias Flodin wrote: >>> (ii) Requiring all documentation to support a no-script option. >>In think (ii) is an absolute requirement. I find myself using >>text-based browsers from time to time, especially when working across >>an SSH connection, and I get quite annoyed by some documentation sites >>that navigation exclusively though Javascript. I would hate for boost >>to go the same way. >>In addition to some text browsers not supporting Javascript at all, >>the CERT has on numerous occasions advised that Javascript be disabled >>in browsers such as IE for security reasons, and some companies make >>that a policy decision. > Actually I agree with this. I will change the serialization navigation > pane so that it works on browsers that don't support scripting. This > would leave the navigation pane but it would always be maximally expanded > for browsers that don't support it. I think that would fulfill the > requirement of making the documentation universally useable and still > permit us to use the facility we like most of the time. I've just modified contents.html slightly so that the expand/collapse functionality is only enabled all of the following are true: a) scripting is supported by the browser b) scripting is not turned off c) the browser is either IE or Netscape Under all other conditions, the navigator pane is displayed fully expanded with no graphic +/-/o buttons. The links to the actual doc pages work the same in all cases. I have tested this with Netscape with scripting enabled and disabled. I've tested with IE only with scripting enabled as I was unable to discover where to turn javascript off (or perhaps I did find it but it didn't work). If anyone wants to tweak this to enable this functionality for their favorite browser - be my guest. I would hope that almost everyone could find this solution satisfactory. Robert Ramey

2 1

Re:Re: [boost] Re: Serialization Menu
by Robert Ramey 16 Jul '04

16 Jul '04

Vladimir Prus wrote: >Jonathan Turkanis wrote: >> I think Robert's menu is very effective, and was hoping it represented >> a de facto relaxation of the no-script policy. >>I have some reservation >>1. It does not work very nice when image loading is turned off (as it is for me). >>2. It takes qutie a lot of screen space. Hmm - on IE and Navigator I can slide the frame all the way to the right and make the navigation window disappear. This also raise the question about browsers that don't support frames. It would be my understanding that usage of frames would also make documentation un-displayable on some browsers. >I think it's probably better to somehow make all new libraries use >Boost.Book. >If that's done, any presentation issues can be solved in one place. For >example, I think there were already discussion about expandable TOC on the >boost-doc failings list. I would like to see one package done in boost book before we impose such a requirement. When I go to the NightlyCVS link in the main boost page it shows what looks like boost book generated main page (whose aesthetics I'm not crazy about). When I click on one of the items it's a dead link. This gives me the impression that boost.book is still in development and that there are still too many pending issues to impose such a requirement. > At all events, I'd prefer constistent look of all Boost library docs (and > the ability to print them), to every author introducing something they > like. I think the boost documentation information www.boost.org is pretty good and helps promote this idea. I think it should used as a starting point. Robert Ramey

2 1

RE: [boost] Re: Boost.Threads: Do we need all those mutexes?
by Batov, Vladimir 16 Jul '04

16 Jul '04

Mike, > > And I feel that the interface (representing convenience, > > flexibility, usability, etc.) is the first thing that a library will be > > judged by. > > Perhaps, but the next thing it will be judged by is efficiency and > speed. Ultimately, many C++ programmers won't use even the cleanest and > most elegant designs if they result in code that isn't fast enough or > efficient enough--especially in something like a multi-threading library. > > Peter Dimov's argument, if I understand correctly, is that the "cleaner" > design could be made as fast/efficient as the current one, not that we > should choose a cleaner design at the expense of speed or efficiency. [Batov, Vladimir] I never advocated forgetting efficiency. Only that the first thing needs to be addressed/settled on first. Then, we move on to the second within the constraints of the first choice. > > I feel that for the user having different types of mutexes is > > inconvenient and unnecessary restricting. > > Inconvenient I can see, but how is having more choices restricting? [Batov, Vladimir] It is restricting because I presume that different types of mutexes will have different functionality. Therefore, the plain mutex will not have something I need for try_lock. If that assumption of mine is not correct and, say, try_mutex is complimentary -- an efficient implementation of the generic mutex (with efficiency vs. functionality trade-offs), then I am fine with it. > > Having additional (and maybe unused) data in a mutex does not seem like > > a huge problem > > Not for you, perhaps, but I can guarantee that there are others who > don't feel that way. [Batov, Vladimir] True. I am fine if we address the mainstream use first (that is, a generic mutex represents an access serialization to the relevant data with the *kind* of serialization (timed, try, etc.) refined by locks). Then, mutex refinements will be just that -- refinements for a particular purpose. Like generic mem. allocation mechanism (kinda OK for everyone) vs. optimized (extremely good for some). > > (as long as it does not impose performance penalties) --s > > memory is cheap and mutexes are not usually created in huge numbers. > > Memory isn't cheap for everybody: there are still some (many?) who are > forced to work in more constrained environments. For instance, > programmer's of handheld devices often deal with insufficient memory > problems. [Batov, Vladimir] True. That falls into the same specialization category. I simply do not want that specialization at the expense of the mainstream users/applications.

1 0

Re:[boost] Regression Tests, Boost.Serialization
by Robert Ramey 15 Jul '04

15 Jul '04

Martin Wille wrote: > Houston, we have a problem. I am sorry about this. A couple of weeks ago I did raise the issue about testing resources regarding Boost Serialization. I was mainly concerned more about execution time. After a short back and forth I was left with the impression that no one thought testing this would be a problem. I'm open to suggestions. Again I apologize for not giving a more specific warning on this - I guess having a 100MB on my home system just skewed my perspective. Robert Ramey

1 0

RE: [boost] Re: Lock unification [was: Boost.Threads: Do we need all those mutexes?]
by Batov, Vladimir 15 Jul '04

15 Jul '04

> Christopher Currie wrote: > > I can infer that the apparent justification for unification would be to > > allow a single lock to be used in different ways, but I have a hard time > > imagining a use case where this would be as clear and as error-resistant > > as separate scopes with separate locks. If anyone has one, or any other > > rationale, please respond. > > The main motivation behind the idea of unifying the lock types (and > mutex types) is simplification of concepts, I believe. > > Mike [Batov, Vladimir] I would not call it 'simplification' as it can send a wrong message but rather 'streamlining'. It's because conceptually (and in implementation) try_lock = timed_lock(time=0) scoped_lock = timed_lock(time=infinite) Although I have to agree that from user perspective those locks are distinct and it makes sense to me to provide thin timed_lock, try_lock, blocking_lock interfaces for the majority of mainstream uses.

1 0

RE: [boost] Re: Lock unification
by Batov, Vladimir 15 Jul '04

15 Jul '04

I like the suggestion of constructing an appropriate lock via the appropriate constructor rather than a function (after all, that's what constructors are for :-) ). Although I'd probably try being more explicit typedef scoped_lock Lock; Lock l(m); // block Lock l(m, Lock::try); // try Lock l(m, Lock::time(33)); // timed Lock l(m, Lock::deferred); // deferred BTW, I love 'deferred'. IMHO it is very good and descriptive name for a parameter. Much better than we had before (was it block/non-block?). > -----Original Message----- > From: boost-bounces(a)lists.boost.org [mailto:boost-bounces@lists.boost.org] > On Behalf Of David Abrahams > Sent: Friday, 16 July 2004 7:24 AM > To: boost(a)lists.boost.org > Subject: [boost] Re: Lock unification > > "Eric Niebler" <eric(a)boost-consulting.com> writes: > > > David Abrahams wrote: > >> "Eric Niebler" <eric(a)boost-consulting.com> writes: > >> > >>>I continue to believe that a better interface for a unified lock > >>>involves descriptively named helper functions: > >>> > >>>scoped_lock l1( m ); // lock unconditionally > >>>scoped_lock l1 = defer_lock( m ); // don't lock > >>>scoped_lock l2 = try_lock( m ); // try lock > >>>scoped_lock l3 = timed_lock( m, t ); // timed lock > >>> > >>> It's just as clear as you could hope for, > >> I'm not convinced it is. It potentially puts information about the > >> kind of locking being done very far from the actual call that does > >> locking. > > > > > > True, but the other design under consideration doesn't help with > > that. It requires two-step construction for some constructs which > > could result in code like: > > > > scoped_lock l( m, false ); > > maybe_takes_the_lock_i_dont_know( l ); > > I don't know what you're referring to. I don't see any two-step > construction going on there. Links or examples, please? > > >>>and it's extensible. I feel that the single lock c'tor with a bool > >>>arg is less readable and rather inflexible. > >> Inflexible how? > >> > > > > User's can't add more constructors if they want a different locking > > strategy, but they can add as many helper function as they like. For > > instance, they could define a function that tried to take a timed_lock > > and logs failures. Or throws an exception. Or they could write a > > lock_if helper that takes a predicate. And they can do all that with > > simple one-step construction syntax. It's more concise and it permits > > the lock declaration to appear in the condition of if statements. > > Sorry, I missed the "single lock ctor" part. I think I agree with > you there. > > BTW, it just occurred to me that > > timed_lock(m, 0) > > could potentially generate different code from > > timed_lock(m, some_int) > > If we can detect literal zero... which I think is possible. > > What about: > > scoped_lock l(m); // block > scoped_lock l(m, 0); // try > scoped_lock l(m, 33); // timed > > scoped_lock l(m, deferred); // deferred > scoped_lock l(deferred(m)); // alternate > > ?? > > -- > Dave Abrahams > Boost Consulting > http://www.boost-consulting.com > > _______________________________________________ > Unsubscribe & other changes: > http://lists.boost.org/mailman/listinfo.cgi/boost

1 0