- Boost - lists.boost.org

[config/multiprecision/units/general] Do we have a policy for user-defined-literals?
by John Maddock 06 May '13

06 May '13

Folks, I've been experimenting with the new C++11 feature "user-defined-literals", and they're pretty impressive ;-) So far I've been able to write code such as: auto i = 0x1234567890abcdef1234567890abcdef_cppi; and generate value i as a signed 128-bit integer. It's particulary remarkable because: * It actually works ;-) * The value is generated as a constexpr - all evaluation and initialization of the multiprecision integer is done at compile time. * The actual meta-code is remarkably short and concise once you've figured out how on earth to get started! Note however my code is limited to hexadecimal constants, because it can't do compile time radix conversion - that would require a huge meta-program for each constant :-( This is obviously useful to the multiprecision library, and I'm sure for Units as well, but that leaves a couple of questions: 1) We have no config macro for this new feature (I've only tested with GCC, and suspect Clang is the only other compiler with support at present). What should it be called? Would everyone be happy with BOOST_NO_CXX11_USER_DEFINED_LITERALS ? 2) How should libraries handle these user defined suffixes? The essencial problem is that they have to be in current scope at point of use, you can never explicitly qualify them. So I suggest we use: namespace boost{ namespace mylib{ namespace literals{ mytype operator "" _mysuffix(args...); }}} Then users can import the whole namespace easily into current scope right at point of use: int main() { using namespace boost::mylib::literals; boost::mylib::mytype t = 1234_mysuffix; } 3) How should the suffixes be named? There is an obvious possibility for clashes here - for example the units lib would probably want to use _s for seconds, but no doubt other users might use it for strings and such like. We could insist that all such names added to a boost lib are suitably mangled, so "_bu_s" for boost.units.seconds, but I'm not convinced by that. Seems to make the feature much less useful? Many thanks in advance for your thoughts, John.

6 12

[chrono/date] Performance goals and design summary
by Vicente J. Botet Escriba 05 May '13

05 May '13

Hi, After the discussion on other threads I would like to summarize some of my performances goals while refactoring the H.H chrono/date library, just to see if I was wrong and going on the bad direction: 1. *Be able to build dates without validating them.* 2. *I wanted the library to be as efficient as possible when the user uses constant objects or literals, no need to validate at runtime what has already been validated by the compiler*. 3. *Possibility to reassign a date if the parameters represent a valid date.* 4. *No a single date class is better than the others on all the algorithms*. 5. *No additional meta-data in absolute dates so that you don't pay for the relative dates overhead when using absolute dates.* 6. *The user must be able to get all the date related informations at once.* Do you have other? The way I have managed with these goals is 1. *P*rovide an unchecked date construction interface with a no_check_t parameter and an is_valid() function. 2. *T*he date parts day/month/year need to be validated or constructed with no_check_t. 3. dt.set_if_valid(year, month, day) 4. Define the Date requirements all the dates must satisfy, implement ymd_date, days_date, ... 5. Separate the absolute and the relative Date requirements into two Date concepts AbsoluteDate and RelativeDate. Provide at least a relative_date class based on the ymd representation. 6. Conversion between different concrete date classe. H.H approach respond to these goals in a different way 1. date and its parts provide constructors that are not validated, the validation is provided by factories as operator/. 2. I don't think this can be achieved when the date must be validated. 3. No problem, this could be added to H.H approach (at least for its YMD date). 4. Same approach. H.H proposes two classes chrono::day_point (only unchecked) and YMD date. 5. I'm not sure the date YMD class doesn't include meta data (Howard could you clarify your point here) 6. same approach. I have some questions for followed by my personal answer; * *Do you see some unnecessary goals, in particular, is the goal 2**necessary?* IMO, yes. Not providing an interface that allows compiler optimizations is against the C++ philosophy. The question is as always, what would be the gain? In the best case we could avoid 6 comparisons. * *Does the introduction of no_check to achieve goal 1 and 2 results in a uglier date interface?* IMO, yes, but I don't know how to achieve goal 2 without. * *Should a concrete date class provide just the functionalities it is able to do efficiently?* IMO, yes. This would prevent the user of non efficient services (Note that my current design doesn't follows this) * *Do we need a Date concept that a concrete date must model**?* IMO, yes. SO we need two levels, level one provides the most efficient service, level two provides an homogeneous interface. * *Do we need to make the separation between absolute and relative dates?* IMO, yes. While relative dates are powerful they incur on unwanted overhead when absolute dates are needed. Best, Vicente

4 14

[build] Conditional sources and compiler options
by Andrey Semashev 05 May '13

05 May '13

Hi, I'm trying to add some SIMD support to my library and I'm having problems with setting up my Jamfile properly. Basically, I have a number of source files that only need to be compiled when the compiler is of a specific version or later (say, gcc >= 4.7 or msvc >= 12 or intel >= X.Y on x86 but not for other compilers/architectures and not for older versions of the mentioned compilers). Next, I need to add special compiler options (specific for each compiler) just for these files. Natuarlly, I also have other source files that need not these tricks and should be compiled normally, without any special options. All source files have to be compiled into a single library (static or dynamic). Is there any way to do it in Boost.Build? I've read about alternatives in the docs, but they require duplicating the library target, which I'd like to avoid. I also read about conditional requirements, but as I understand they don't allow to specify version conditions (like <toolset>gcc >= 4.7). Also I didn't find any way to specify special build options for particular files. Thanks in advance.

2 2

[boost.preprocessor.iteration] relative file iteration not supported
by Adrian H 05 May '13

05 May '13

Hello, I have used Boost in the past and do like the use of the preprocessor as a way of making repetitious code. However, I found that in a project that I was working on, I needed the ability to iterate over a file relative to the file I was iterating from, and that file wasn't relative to any of the include directories. I thought about this and it makes sense to why it didn't work, but I can see that it can prove to be inconvenient sometimes when someone wishes to make library header files relative to each other to allow for flexible positioning of the related files in the source tree or not have to add new items to the include directory set. I have written a preprocessor library to fill this need, but was wondering if such a library would be of interest to the Boost community. I deliberately tweaked it to be similar to the Boost Preprocessor Iterator library, and can be acquired at https://sourceforge.net/projects/relativeiterate/, including documentation and an example located in the README.txt file. I would appreciate any feedback you may have on this and would be more than willing to make changes to accommodate what would be needed for it to be included into the Boost set. Thank you for your time. Adrian -- ======================================== Adrian Hawryluk BSc. Computer Science ---------------------------------------- Specialising in: OOD Methodologies in UML OOP Methodologies in C, C++ and more RT Embedded Programming GUI Development ========================================

3 4

[chrono/date] constructor, factories and validation
by Vicente J. Botet Escriba 05 May '13

05 May '13

Hi all, Two approaches to checking date validation have been proposed Howard approach: * date constructors build unchecked dates (provide the same order of parameters that the date factories. * date factories (expression using operator/()) build checked dates. * could be completed with make_date that will build checked dates. Advantages: * avoids the ugly no_check (that could be replaced by no_throw) Drawbacks * The 3 sets of date constructor factories must provide all the same order of parameters. * Creating a new factory will mean creating a new constructor on each concrete date class. * The factories must be friend of the concrete dates. My original approach complemented with make_date factory: * constructors without no_check build checked dates, and with no_check parameter build well unchecked ones. Provides only decreasing lexicographic order of arguments e.g. date(year,month,day), date(year,month,day,no_check_t). * make_date provides the usual combinations as provided by factory using operator/() and no_check. * date factories (expression using operator/()) build checked dates Advantages: * There are less date constructors. * We can add as many factories as we are able to construct with checking or not. Drawbacks * the no_throw/no_check/unchecked parameter is ugly and a source of conflict. Alternative approach (reverse the defaults): * constructors without check build unchecked dates, and with check parameter build well checked ones. Provides only decreasing lexicographic order of arguments e.g. date(year,month,day), date(year,month,day,check_t). * make_date provides the usual combinations as provided by factory using operator/() and check. * a make_checked_date could have also a sens. * date factories (expression using operator/()) build checked dates Best, Vicente

1 0

[operators] The Dangling Reference Polarization
by Daniel Frey 05 May '13

05 May '13

I see some good discussion going on, thanks to everyone participating! Rest assured that I value both side's contributions, it is important for me (and those agreeing with my perspective) to be able to defend this position and also to keep looking for improvements. If our side can not be defended, it's probably not good enough :) That said, here's what I hope can serve as a first summary, let's first look at it from a technical perspective: If an operator returns an rvalue reference (T&&), it is safe for all use-cases except binding the result to an rvalue- or const-lvalue-reference: T r = a + b + c; // no problem const T r = a + b + c; // no problem T& r = a + b + c; // compile-time error const T& r = a + b + c; // problem (P1) T&& r = a + b + c; // problem (P2) It seems that this is consistent behavior across at least GCC, Clang and VC++. If a compiler shows a different behavior, its likely a compiler-bug. Technically, there is no doubt that the above is a bug, but that doesn't answer the question of whether this is a bug of the user who wrote the above code or if it's a bug of the operator to return an rvalue reference instead of an rvalue. This is where we leave the technical-aspects-land and enter the minefield of opinions, expectations and decisions. I am of the opinion that the above problematic use-cases are a bug on the users side, my line-of-thoughts and my arguments are: a) The above code is explicit, you can spot it in the users code. It's not something that happens silently. b) The above code always looked suspicious to me anyways, it has a code-smell that I just don't like. c) It rely's on the fact that it extends the lifetime of a returned temporary value. The rules of the C++ standard that allow this are quite clear, and hence it's the users responsibility to make sure that there actually is a temporary value whose lifetime can be extended. It has always been an error if the expression returned a reference and not a value. Consider const std::string& oops = std::vector<std::string>(1, "hello, world" ).front(); This compiles are creates a dangling reference, as the standard says that front() returns a reference and not a value. It's not 100% comparable as in this case the standard guarantees a bug, but the main point for C++ has always been which guarantees are given and not about a certain expectation one might have. As an example: static_assert( std::is_empty< std::tuple<> >::value, "Expectation 1 broken" ); static_assert( std::is_empty< std::array< T, 0 > >::value, "Expectation 2 broken" ); // with any valid T The first case is certainly a reasonable expectation and it currently works for all implementations of std::tuple that I know of, but it is not guaranteed! If your code relies on it and it breaks in the future, you can not blame the person who modified std::tuple. The second case is also a reasonable expectation, but it breaks everywhere and AFAIK there is no standard-conforming way to implement std::array so that std::array<T,0> is empty. For the curious, here's the StackOverflow question I asked about it: <http://stackoverflow.com/q/15512827/2073257>. Another example about how to properly handle expectations vs. guarantees: <http://stackoverflow.com/q/14882588/2073257>. d) I fail to see any valid use-case for binding the result of the expression to a reference. Can someone please provide a convincing example of why (P1) or (P2) are needed/useful? Keep in mind that in the context of operator+ (or any other operator in question), we already require that the type T is copy-/moveable and that copy-elision most likely takes place if you use "T r = …" instead of "const T& r = …". e) The alternative needs to prove its claimed/theoretical efficiency in practice. Too often I have heard about "sufficiently smart compilers", but I'm more the prove-it-in-the-real-world type. So summarizing it once again: For a) and b), this is just an observation and my personal opinion. c) means it's important to talk about guarantees, not just about expectations (no matter how reasonable they are). d) poses the question if we loose something if we don't guarantee that (P1) and (P2) are allowed. I currently have several ideas/option of how we could improve Boost.Operators, some of which I already mentioned, some which only recently occurred to me. Before I will discuss these, I'd like to resolve (if possible) the Dangling Reference Polarization and gather more opinions/feedback. So far, I found the discussion both challenging and very helpful, again thanks to everyone participating! Best regards, Daniel

5 13

[gsoc] Boost.Multiprecision Proposal
by Saurav 04 May '13

04 May '13

Hi, I have attached my proposal for the Boost.Multiprecision project. It has been provided in PDF format for better readability. Please have a look and give your feedback. Best regards, Saurav

3 4

[Boost.ThreadPool] GSOC 2013 Proposal
by Ian Forbes 04 May '13

04 May '13

Hi, I am submitting my Google Summer of Code proposal for review. My proposal is for the Boost.ThreadPool library. It will be based on work done here: https://svn.boost.org/trac/boost/wiki/SoC2013#Boost.ThreadThreadPool The current proposal can be found at: http://cs.mcgill.ca/~iforbe/gsoc2013.html The original proposal can be found here: http://www.google-melange.com/gsoc/proposal/review/google/gsoc2013/en4bz/1 , although it has broken HTML tags and is ill formatted. Thanks goes to Vicente J. Botet Escriba for pointing out that Boost does in fact has LockFree concurrent data structures. Please feel free to respond with any questions or criticism. I will try to respond to the best of my ability. Thanks, Ian Forbes

1 0

[chrono/date] year/day/week literals
by Vicente J. Botet Escriba 04 May '13

04 May '13

Hi, on the GSoC discussion about Boost.Chrono/Date proposal we were discussing about date construction. Some of us think that we need to use named types for day, month, year and week so that the date constructors are not ambiguous. Everyone agree with the constant object for month. date dt(year(2013), may, day(3)); But having to use day(3) or year(2013) seems to wordy. I was wondering if we can not add some literals for day, year and week so that we can just write date dt(2013y, may, 3d); The advantage I see in addition to been less wordy, is that we will have a compile error when the year, day or week is out of range. What do you think? Vicente

5 11

Live read-only GIT mirrors of Boost trunk SVN
by Niall Douglas 04 May '13

04 May '13

The live GIT mirrors of Boost trunk SVN have become a bit stale in recent months: https://github.com/ryppl/boost-svn, long time favourite of many, now 4 months stale. https://github.com/Beman/boost-trunk-git-svn is now 2 months stale. https://github.com/alexanderkjeldaas/boost-mirror is fresh, but has no branches nor tags. They all generally use git-svn, which tends to be brittle and not translate commits perfectly unfortunately e.g. SVN eol properties and other such metadata. So I decided to try out a new Subversion <=> Git mirroring tool, SubGit v2.0 EAP (http://subgit.com/eap/) whose author very kindly gave me a free open source licence and the result is at: https://github.com/ned14/boost-trunk All branches and tags, as far as I can tell, are mirrored correctly. I've set a cron job every 24 hours. Theoretically if we had an authors mapping file it would work in the opposite direction too i.e. be read-write. My next task is to live mirror a github repo into Boost Sandbox. Currently SubGit segfaults on this due to the complex SVN properties auto-translation which requires a specially marked up .gitattributes, but the author is very keen to fix the problem. In theory then it ought to just work. Any mistakes in my Boost mirror, or any problems, or even thoughts are gratefully received. Niall

4 6