May 2014 - Boost - lists.preview.boost.org

[modularization] proposal and poll
by Julian Gonggrijp 09 Jun '14

09 Jun '14

Dear Boost mailing list members, I would like to revive the topic of Boost inter-library dependencies, suggest two global changes to Boost and ask whether you agree with the course of action I propose. RATIONALE Boost recently transitioned from a monolithic SVN repository to a "modular" Git repository. One of the potential benefits of this new structure is that it may allow users to copy much less of Boost when they need only a small part of it. I believe that two additional, evolutionary changes are needed to truly unlock that benefit. The currently-standard way to use the new Git repository is to recursively clone the super-project and all its submodules, one for each of the >100 libraries. In order to use Boost in a truly modular way, a developer would rather just clone only the libraries she needs (including dependencies). Unfortunately, there is currently no feasible way to do that [see also 1]: - Manually: this is cumbersome and generally hard because there are many (undocumented) interdependencies between Boost libraries. - Automatically: as far as I know Ryppl is the only existing attempt at providing automated dependency handling for Boost, but it is a very ambitious project that attempts to do much more and it is currently neither finished nor being actively developed. I think there could be a less ambitious but faster way to provide automated dependency handling. It would take no more than some static configuration, to list for each library what other Boost libraries it depends on, combined with a simple script that automatically traverses the dependency graph and clones just the necessary submodules. This script may then be hooked into the bootstrap script. I think I could write the dependency handler, and I would be willing to, though I can't currently provide any indication of when I would be able to start. However, even if we had an automated way to handle dependencies, that would not be a sufficient solution to true modularity. As has been discussed before, the interdependencies between Boost libraries are currently so strong, that even using a single header file may likely require the user to copy a substantial portion of Boost [2]. This is especially the case because there are many cyclical dependencies [3, 4]. Stephen Kelly investigated how the interdependencies between Boost libraries could be reduced, especially the cyclical dependencies. He concluded that most cyclical dependencies could be eliminated and that the dependency graph could be simplified substantially overall. He also described concrete ways to achieve this. [3 and onwards] I believe it would still be worthwhile to reduce interdependencies between Boost libraries in the way Stephen proposed, or with a very similar approach. Stephen told me he is willing to re-analyze the dependencies as many times as needed as he did before and to make recommendations. Together with an automated way to handle dependencies, this should enable users to install only a small portion of Boost if that is all they need. It will also make the Boost super- project more flexible overall. PROPOSAL The following (evolutionary) global changes to Boost should be planned and given priority over any other proposals [e.g. 5], in the following order: 1. Reduction of dependencies between Boost libraries. 2. Simple but effective automation of dependency handling. POLL I think it may be easier for the steering committee to justify any decision on this proposal, if they have hard statistics about the amount of support it receives. Therefore, I invite every opinionated person reading this to reply with a vote, either in favour of the proposal or against it. I invite you to do this even if you do not feel the need or find the time to provide arguments for your position. You can later re-vote if you change your mind. I will count the votes, and when there are more than 30 I will start posting the statistics occasionally. Votes are public so everyone can verify my data when necessary. Anyone, including the steering committee, can then use the numbers however they choose. Thanks in advance, -Julian ____________ [1] http://lists.boost.org/Archives/boost/2014/05/213115.php [2] http://lists.boost.org/Archives/boost/2013/10/207379.php [3] http://lists.boost.org/Archives/boost/2013/10/207384.php [4] http://lists.boost.org/Archives/boost/2013/10/207422.php [5] http://lists.boost.org/Archives/boost/2014/05/213229.php

15 49

[optional] Is detail/none_t.hpp needed?
by Andrey Semashev 06 Jun '14

06 Jun '14

Hi, I've noticed there's boost/detail/none_t.hpp header, which basically reflects boost/none_t.hpp. Worse even, it would cause a conflict if both these headers are included. A search didn't reveal any users of this header. Since it's a detail, could this header be removed?

2 5

[multiprecision] General/design query about which conversions should be implicit/explicit
by John Maddock 04 Jun '14

04 Jun '14

Folks, I have an open bug report https://svn.boost.org/trac/boost/ticket/10082 that requests that conversions from floating point to rational multiprecision-types be made implicit (currently they're explicit). Now on the one hand the bug report is correct: these are non-lossy conversions, so there's no harm in them being implicit. However, it still sort of feels wrong to me, the only arguments against I can come up with are: 1) An implicit conversion lets you assign values such as 0.1 to a rational (which actually leads to 3602879701896397/36028797018963968 not 1/10), where as making the conversion explicit at least forces you to use a cast (or an explicit construction). 2) Floating point values can result in arbitrarily large integer parts to the rational, effectively running the machine out of memory. Arguably the converting constructor should guard against that, though frankly exactly how is less clear :-( Does anyone else have any strong views or insights into this? Thanks in advance, John.

8 8

[GSoC][MPL11] Post C++Now update
by Louis Dionne 04 Jun '14

04 Jun '14

Dear Boost, A lot happened during the past week in Aspen; the goal of this message is to summarize insights, changes and feedback received wrt MPL11. The slides of the presentation are available at [1]. The second part of the presentation is especially relevant to get an overview of the designs considered so far. Merging the MPL and Fusion -------------------------- After discussing the issue several times during the week, I (and others) think it might be possible to merge Fusion and the MPL into a single library. I am currently trying to write a library that does that. Since this constitutes a large reorientation, I created a new repository which is available at [2]. Those with interest should consider subscribing to the repository to be updated as I make progress. Iterator-free design -------------------- I was a bit anxious that people would complain about this, but the contrary happened. Nobody complained and some people expressed support. This design aspect will stay as-is. Request for guinea pigs ----------------------- I am looking for people with hardcore C++1y metaprogramming needs who will be willing to test the library I come up with (MPL11 or MPL + Fusion merge). I also need to see use cases for the library, so please reply even if you are not willing to test. Roadmap ------- Finish exploring the design space. If it turns out that we can merge the MPL and Fusion, then I'll be working on that for a while. If I find that this is not possible or not a good idea, I'll go back to my initial GSoC planning. In both cases, I hope to have a library in the incubator by the end of the summer. Also, this new roadmap is approved by my mentor. Feel free to reply with questions and/or comments. Regards, Louis [1]: http://ldionne.com/mpl11-cppnow-2014 [2]: http://github.com/ldionne/hana

15 34

Proposal: Boost.Core
by Peter Dimov 04 Jun '14

04 Jun '14

I propose the creation of a new module, Boost.Core, a repository of Boost core utilities. The criteria for inclusion into this module would be that the utility component is (a) simple, (b) used by other Boost libraries, (c) does not depend on any headers outside of Core and Config (and perhaps Assert). Its primary purpose would be to break dependencies. It will not serve as a repository of "this might be useful". Based on the interest so far, I volunteer myself, Glen Fernandes and Andrey Semashev as maintainers. The following utilities are immediate candidates for migration into the new module: - addressof - enable_if - boost/pointee.hpp - boost/indirect_reference.hpp - boost/type.hpp In addition, the following can be duplicated inside to reduce dependencies: - mpl::if_, if_c, eval_if - detail/is_incrementable.hpp - remove_cv and so on.

10 51

Boost 1.56 timeframe
by Jonathan Jones 03 Jun '14

03 Jun '14

Hi folks, Is there a timeframe for the release of Boost 1.56? The current 1.56 page simply states that it's at an early stage in its release cycle (but it has said this for months). http://www.boost.org/users/history/version_1_56_0.html I'm asking in the context of Visual Studio 2013. Boost 1.55 is known to have issues with VS2013, which I'm hoping the 1.56 release would address. Sincerely, Jonathan Jones

7 10

github issue
by Christian Henning 02 Jun '14

02 Jun '14

Hi all, I'm cloning boost but something went wrong when trying to get the python submodule. I'm not sure if the problem is related to boost, github, my git setup, or maybe my internet connectivity. Here is a snippet from the console output: ... Cloning into 'libs/proto'... remote: Reusing existing pack: 4233, done. remote: Total 4233 (delta 0), reused 0 (delta 0) Receiving objects: 100% (4233/4233), 1.89 MiB | 893.00 KiB/s, done. Resolving deltas: 100% (2062/2062), done. Checking connectivity... done. Cloning into 'libs/ptr_container'... remote: Reusing existing pack: 2887, done. remote: Total 2887 (delta 0), reused 0 (delta 0)Receiving objects: 98% (2830/28 87), 2.36 MiB | 924.00 KiB/s Receiving objects: 100% (2887/2887), 2.78 MiB | 924.00 KiB/s, done. Resolving deltas: 100% (1839/1839), done. Checking connectivity... done. Cloning into 'libs/python'... ssh: github.com: no address associated with name fatal: Could not read from remote repository. Please make sure you have the correct access rights and the repository exists. Clone of 'git@github.com:boostorg/python.git' into submodule path 'libs/python' failed I'm using git version 1.9.0.msysgit.0 on Windows 7. Anyone know what's going? Thanks, Christian

2 2

Re: [boost] [modularization] proposal and poll
by Niall Douglas 01 Jun '14

01 Jun '14

I'll top post my reply as Robert's valuable post didn't make it onto the list again. Please do read his reply below, it's worth reading as Robert is one of the lead thinkers on the steering committee about this. Robert, I think you are under a misapprehension about what I mean by "fork". By a fork, I mean forking just a very small part of existing Boost, namely whatever minimal support is necessary to start from scratch with a new *empty* set of C++ 14 only Boost libraries. I don't propose to port *any* of the legacy Boost libraries whatsoever, they stay where there are. If their maintainers - or a sponsor - are willing to do: 1. Make their library entirely C++ 11/14 STL based, porting only the absolute minimum of old Boost. 2. C++ Modules ready i.e. restructured and reorganised to properly fit C++ Modules. 3. Contribute any general utility code to the core utilities library, not duplicating any functionality there, and in a form suitable for entering the C++ 17 STL. 4. Meet the raised unit test and CI soak test and sanitiser/valgrind requirements. Then they have a good chance of their library appearing highly ranked in the list of C++ 14 Boost libraries, where that ranking is generated by a set of clang AST analysers. Does this make sense? I am specifically advocating a structural break, a clean start. I think working on a fresh, empty Boost will get people to volunteer to give up family time to write Boost code again. Right now with the present situation there are few incentives to sacrifice family time. That needs to change for Boost to survive, else the continuing slow decline will continue. Niall On 29 May 2014 at 17:52, Robert Ramey wrote: > > > On Thursday, May 29, 2014 2:17:11 PM UTC-7, Niall Douglas wrote: > > > > On 29 May 2014 at 22:22, Julian Gonggrijp wrote: > > > > If a fork of Boost becomes necessary later this year, > > > LOL - this is not going to be necessary. > > > > The hard part isn't what to do. > > > LOL - not for you! > > > > The hard part is gaining community > > buy in when "their" code gets broken by the necessary changes. I have > > concluded it isn't possible, > > > LOL - you've underestimated the amount of work to convert > everything to C++11 by a factor of .... 1000 ? Basically it's > a whole rewrite of 100 modules. And the gain is ....? > > hence me proposing a fork where such > > "radical" changes won't upset anyone. > > > Your free to do this. And you should if you really believe that > it would benefit anyone or anything. > > > I think it may be easier for the steering committee to justify any > > > decision on this proposal, > > > The steering committee doesn't have anything to decide here. The > "Proposal" is: > > 1. Reduction of dependencies between Boost libraries. > > It's not clear what action is being proposed - but I'm pretty sure > the steering committee won't be doing it. > > 2. Simple but effective automation of dependency handling. > > I'm pretty sure the steering committee won't be doing this either. > If anyone believes that they can implement "simple and > effective" automation of dependency handling - just go ahead > and do it and let us see it. > > > > Decisive and immediate action on this is vital if Boost is to > > survive. I vote yes to both, ASAP. > > > > LOL - There is wide agreement that boost has growing pains and > will have to evolve. But I haven't seen any proposals for > "Decisive and Immediate" which would do anything but make > things worse. > > And since you've got me started - Here is what boost needs go to the next > step. > > a) Make the review process more effective and efficient - we're taking real > practical action to address this now. The jury is still out regarding the > effectiveness of our efforts - but we ARE actually doing something. > > b) Recognize that we now have 4 variants of C++ - C++98, C++03, C++11 > and C++14. Our infrastructure doesn't explicitly address this. The > following > needs to change > 1) for each library we need to explicitly state somewhere what > version(s) of C++ the > library is meant to support. > 2) support running of tests and builds for only those platforms. > > I've suggested that we enhance B2 to address this. So far > no one has responded to this suggestion > > Note that this would have he effect of encouraging libraries to evolve > to later versions of C++ with minimal disruption to both library > authors and users. > > c) Decouple Deployment of Boost Libraries from the Certification > of boost libraries. Beman has noted that what he would like to > see is something like CYGWIN whereby one specifies what one > want's and somehow automagically you get that plush what you > need. I realize that this is more or less what is being suggested here. > But no one has presented credible proposal for making it it happen. > > This would also address the issue of "old" libraries. As they > fall into disuse are become superseded by newer ones, the can > still be in boost even though their frequency of deployment drops off. > > The above constitutes a real path forward. This is what Boost > needs for the next 15 years. > > Robert Ramey > -- ned Productions Limited Consulting http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/

4 4

throw or boost::throw_exception?
by Andrzej Krzemienski 01 Jun '14

01 Jun '14

Hi, I am not sure where, but I am convinced that I have read that it is recommended for Boost libraries that they not use naked throw expressions but instead call boost::throw_exception. Is that true? If so, it enforces a dependency of more than half Boost libraries on Boost.Exception. Am I right? If so, is this the right recommendation in the context of the discussion about reducing dependencies between the libraries? Regards, &rzej

8 22

[next_generation][modularization] Is the community interested in creating a smooth path to a modularized Boost
by Vicente J. Botet Escriba 31 May '14

31 May '14

Hi, I've started a new thread in response to Julian Gonggrijp post [modularization] proposal and poll, as I would not answer directly to his questions. I'm sure that almost all of us want a modularized Boost. What does this means? It is up to us to define it: * what the user can request as a delivered package for its own needs? * what the authors must state explicitly? * what must be tested/checked? The main problem is how to achieve it ensuring backward compatibility . I'm persuaded that we need a new Boost repository (or whatever we want to name it) that would include libraries that respect the modularization constraints we could define. My question is who would be interested in starting the next generation of a modularized Boost that starts from a repository of ZERO modularized libraries. The idea is to add to the current Boost libraries new ones that respect the modular rules and adapt the current ones to be defined on top of the modularized ones. At the end we could deprecate the old ones. In order to do that we need of course to add new interfaces (a new namespace mod_boost or whatever is preferred) and so the user would need to move to these interfaces smoothly. Before starting we need to * Define the constraints of a modularized system - no cyclic dependencies, .... * Define the tools that can be used to support the modular system - which build system(s), - which test system(s), - which regression system(s), - which doc system(s), ... * Define the rules - who could be able allowed to do the refactoring? - Should the refactoring be reviewed to ensure a improvement on the quality of Boost? - who maintains what?, ... * Setup the infrastructure ( build system, test system, regression system, doc system, ... ) Only then we could * Start from libraries that don't depend on any other library (with possible refactoring) and adaptation of the refactored libraries from which these modular libraries have been extracted. * Deliver them * Build a second layer of libraries that depend only on first layer * Deliver them * ... Of course, in order to check that the rules and tools are the appropriated ones we will need to experiment with some examples. I recognize that this would mean a lot of work, however I hope that the final result would satisfy better the community expectations and the quality of Boost libraries. Best, Vicente

2 2