- Boost - lists.preview.boost.org

[1.32 release] TODO list synchronization
by Aleksey Gurtovoy 16 Jul '04

16 Jul '04

As we are coming closer to the branch-for-release date, I'd like to inquire developers working towards the release about the current status of their corresponding TODO items below. Simply replying to this message and saying "Done", "Not done, will be done by date X", "Abandoned" or whatever you wish to say on each of the items would be more than fine. > Circular Buffer > /////////////// > > Pavel Vozenilek writes: > > I asked Jan Galdun and he will try to make circular_buffer available > for 1.32. > > > Date-Time > ///////// > > Jeff Garland writes: > > And while I'm on the subject date-time plans for release, the only 'major > issue' for this release is whether we are switching to boost-book for this > release. We would like to, but if we don't make it we will do it for the next > release. > > > Iterators > ///////// > > David Abrahams writes: > > There are a couple of things that ideally I would like to handle, but > which I am not promising (yet): > > 1. Minor adjustments to code in the iterators library > 2. Iterators library doc fixes > > > MPL > /// > > New version of the library will be checked in the main trunk before July 11. Not done yet; July 15 is the new target. > > > Multi-Index > /////////// > > Joaquín Mª López Muñoz writes: > > Current status: commited in the CVS since 6 weeks or > so. Regression tests are OK for all the compilers that > I intend to support. If someone is interested in some > other compiler being supported, please contact me. > > Pending issues: > #1 Minor updates to the docs. No big deal. > #2 Some changes related to the new defect macro > BOOST_NO_POINTER_TO_MEMBER_TEMPLATE_PARAMETERS > > These issues will be resolved way before the branching > date, I hope. > > > Numeric Conversions > /////////////////// > > Fernando Cacciola writes: > > In preparation for the upcoming 1.32 release I added this library to CVS. > The code and docs have been changed according to the extensive review > results. > I still need to extend the documentation adding examples and motivating > cases though, which I plan to do before the release. > > I haven't yet changed "boost/cast.hpp" to use the new stuff, I will as soon > as I see the code stablilized among different compilers. > > In the meantime, new users can use "boost/numeric/conversion/cast.hpp" which > contains the new version of numeric_cast<> > > > Program Options > /////////////// > > Vladimir Prus writes: > > 1. Add markup for the current failures on Windows. I've done all workarounds I > planned. > 2. Finally get some results on gcc2.95-stlport. I expect that the library > should work there, but due to Boost.Test failure I don't have any results. > 3. Try to make Sun compiler happy. I'm comitted a workaround some time ago, > but tests on Solaris were not run since then. > 4. Check dynamic linking on Windows. > > > Serialization > ///////////// > > Robert Ramey writes: > > I would very much like to see the serialization library in the 1.32 Release. > However, I'm not sure it will be possible to make the posted schedule. > > I have made all changes to address issues raised in the review. On my > personal machine I have passed tests for the compilers VC 7.1 VC 6, gcc 3.3. > Borland compiles (almost everything but requires some debug - it will pass > most tests) Comeau- compiles everthing - accept for a show stopping problem > I've been unable to resolve. My intention has been: > > a) post version compatible with 1.31 next week. > > b) build and test against the CVS tree on my computer > > c) check in > > In order to do this I need: > > a) authority to check in > > b) information on how to mark tests/compilers known to fail. > > c) I have a couple of questions that I need an expert opinion on. > > i) the first one is that in order to work, versions for vc 6.0 and > borland need to be compiled with spirit 1.61. So the a special > environmental variable indicating the directory path where this version of > spirit is located on the testing machine. SPIRIT_ROOT. This variable has > to be set when serialization libraries are built for these compilers. > > I should say I'm more than a little bit concerned about changes to mpl and > iterators. I'm very much aware of the delays and difficulties associated > with the 1.31 release - and it didn't even include this package. > > So I would like to continue on the path I've established above and we'll see > where things stand July 16 when/if you are ready to branch for release. > > > String Algorithms > ///////////////// > > Pavol Droba writes: > > I have few outstanding issues to fix, but they are mostly related to > the documentation. I think, I can manage it. > > As it seems, that the Range library will not make it to this release, > I will use the internal versions of facilities that are currently in the string_algo lib. > > So my todo list is as follows: > > 1) Finish the documentation > - Add functions list > - Add exception specifications > - Cleanup > > 2) Fix examples > 3) If there will be time, improve find_iterator > 4) Fix possible regressions. > > > Questions: > - I have posted some requestes/bugs on the boost-doc list. What is the status > of this. Will they be fixed before release? > > - According to metacom's tables regex is failing on comeau a so string_algo regex > test. Is this a setup or regex problem? > > - There is an problem with Intel compiler, split tests are failing and I cannot figureout > why. The message is very vague. Could somebody with an access to this platform, > try to debug the cause of the problem? > > - In regards to the older compiler support. I cannot resonably support compilers, that I > don't have an access to. Therefor if somebody is interested to use string_algo library > with one of such compilers, I will need an active help with porting. > > > uBLAS > ///// > > Michael Stevens writes: > > The current status of uBLAS is that it has been relatively untouched in the > Boost CVS single 1.30.0. The active development takes place on a separate > CVS. > > Therefore now would be a good time to bring in the latest code which as some > quite significant improvements. The code is relatively stable and tested on > core compilers. > > Issues: > On exotic (old) compiler the code is relatively untested and expected to fail. > Therefore the real need with regard to 1.32 is to get feedback from extensive > regression testing results in good time. > > As ever the documentation could do with further work to bring it up to date > with the to be release version. Thanks in advance, -- Aleksey Gurtovoy MetaCommunications Engineering

10 26

lexical_cast reviewed
by Jason 16 Jul '04

16 Jul '04

after fidling around with template functions i discovered you could write lexical_cast without any template arguments _required_. Also I neat little C-Style cast although no one is probably interested in using that. Anyways the syntax is as followed for both: // lexical cast improved: string str = lexical_cast( 3.1415f ); int i = lexical_cast( "12345" ); // C-style cast fun! string str = (cstyle_cast)3.1415f; int i = (cstyle_cast)"12345"; I would say that anyone already using lexical_cast would'nt have a problem with the improved one. Let me know what you guys think and if we should implement that. -Jason SOURCE CODE: #include <sstream> using namespace std; class cstyle_cast { public: template <typename Source> cstyle_cast( const Source & source ) { ss << source; } template <typename Target> operator Target() { Target target; ss >> target; return target; } private: stringstream ss; }; class lexical_cast { public: template <typename Source> lexical_cast( const Source & source ) { ss << source; } template <typename Target> operator Target() { Target target; ss >> target; return target; } private: stringstream ss; }; int main() { string str1 = (cstyle_cast)3.141592f; string str2 = lexical_cast(3.141592f); }

2 1

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