- Boost - lists.boost.org

BCM - Boost CMake Modules request for endorsement
by paul 22 Sep '17

22 Sep '17

Hi, So the Daniel Pfeifer has agreed to be the review manager for the Boost Cmake Modules. I would like ask for an endorsement so we can schedule a formal review for BCM. Thanks, Paul .

5 7

Travis/Appveyor improvements for the superproject
by Peter Dimov 22 Sep '17

22 Sep '17

Travis and Appveyor tests for the Boost superproject have been improved to automatically run the tests for the library mentioned in the commit message. For example, if the message is "Update iterator from develop", the corresponding CI job runs the tests of the Iterator library (among other things). This is a bit redundant if the library itself has Travis/Appveyor set up, but keeping an eye on https://ci.appveyor.com/project/boostorg/boost/history and https://travis-ci.org/boostorg/boost/builds can't hurt. For example, Appveyor says that Iterator fails a test on msvc-12.0 (Visual Studio 2013): https://ci.appveyor.com/project/boostorg/boost/build/1.0.3744-develop and Utility fails a test on msvc-14.0/msvc-14.1: https://ci.appveyor.com/project/boostorg/boost/build/1.0.3765-develop

9 16

[review][Fit] Review of Fit starts today : September 8 - September 17
by Matt Calabrese 22 Sep '17

22 Sep '17

A formal review of the Fit library developed by Paul Fultz II starts today, September 8, and runs through September 17. A branch has been made for the review. You can find Fit for review at these links: Source: https://github.com/pfultz2/Fit/tree/boost Docs: http://pfultz2.github.io/Fit/doc/html/doc/ Anyone who is able to provide an assessment of the design, implementation, and/or docs can participate in this review. Usage experience is not absolutely necessary to give feedback, but is highly recommended if you are writing a review. Discussions about aspects of the library are welcome even if a complete review is not presented, though participants are asked to please state whether they would like to accept or reject the library before the deadline of September 17, paying close attention to the questions raised at the end of this email. The success of Boost is partially a result of the quality review process which is conducted by the community. You are part of the Boost community. We will be grateful to receive a review based on whatever level of effort or time you can devote. This is the second time that Fit will go through the review process, with the previous time being in March of 2016. The result of the initial review can be found here: https://lists.boost.org/Archives/boost/2016/04/228770.php In this second review, those who choose to participate should be as rigorous as they'd be in any other review, but are also encouraged to voice whether or not the concerns from March of 2016 were addressed. The following is a brief description of the library from its author: ==================== Fit is a header-only C++11/C++14 library that provides utilities for functions and function objects, which can solve many problems with much simpler constructs than whats traditionally been done with metaprogramming. Fit is: Modern: Fit takes advantages of modern C++11/C++14 features. It support both constexpr initialization and constexpr evaluation of functions. It takes advantage of type deduction, varidiac templates, and perfect forwarding to provide a simple and modern interface. Relevant: Fit provides utilities for functions and does not try to implement a functional language in C++. As such, Fit solves many problems relevant to C++ programmers, including initialization of function objects and lambdas, overloading with ordering, improved return type deduction, and much more. Lightweight: Fit builds simple lightweight abstraction on top of function objects. It does not require subscribing to an entire framework. Just use the parts you need. Fit is divided into three components: Function Adaptors and Decorators: These enhance functions with additional capability. Functions: These return functions that achieve a specific purpose. Utilities: These are general utilities that are useful when defining or using functions. Requirements This requires a C++11 compiler. There are no third-party dependencies. This has been tested on clang 3.5-3.8, gcc 4.6-6.2, and Visual Studio 2015. Gcc 5.1 is not supported at all, however, gcc 5.4 is supported. Contexpr support: Both MSVC and gcc 4.6 have limited constexpr support due to many bugs in the implementation of constexpr. However, constexpr initialization of functions is supported when using the FIT_STATIC_FUNCTION and FIT_STATIC_LAMBDA_FUNCTION constructs. Noexcept support: On older compilers such as gcc 4.6 and gcc 4.7, noexcept is not used due to many bugs in the implementation. Also, most compilers don’t support deducing noexcept with member function pointers. Only newer versions of gcc(4.9 and later) support this. ==================== Please provide in your review whatever information you think is valuable to understand your final choice of ACCEPT or REJECT including Fit as a Boost library. Please be explicit about your decision. Some other questions you might want to consider answering: - 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 which compiler(s)? 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? - Were the concerns from the March 2016 review of Fit addressed? More information about the Boost Formal Review Process can be found here: http://www.boost.org/community/reviews.html Thank you for participating! -- -Matt Calabrese

13 39

Crypto/hashes library?
by James E. King, III 21 Sep '17

21 Sep '17

Hello, The boost uuid module contains an implementation of the sha1 hash algorithm for purposes of generating a uuid from arbitrary input. I've opened an issue on github to make the hash algorithm configurable: https://github.com/boostorg/uuid/issues/26 I was hoping to find a concept and implementation of a few hashes somewhere in boost that could be used instead, however I haven't found any. Back in 2010 someone posted a link to a "boost::hashes" repository but I was unable to see what the outcome of the discussion was other than that it was not included. A collection of hashing algorithms that are well tested would be quite useful in a number of ways; and I like the hash size template style used in the repo from 2010: https://svn.boost.org/svn/boost/sandbox/hash/ Indeed there's mention of a discussion about the proper placement of sha1 inside boost::uuid in there: https://svn.boost.org/svn/boost/sandbox/hash/libs/hash/doc/html/rationale.h… So I was curious where that effort ended up, and if there was a resolution on whether an independent library for hashing was ultimately something that folks wanted to pursue or avoid. Thanks, Jim

8 16

[move] [container] workaround for incorrect [-Wignored-attributes]
by Groke, Paul 21 Sep '17

21 Sep '17

Hi, we've stumbled over a supposed bug in GCC 6 that generates incorrect [-Wignored-attributes] warnings. Using anything in Boost that relies on boost::move_detail::aligned_storage (like boost::container::small_vector), with a small enough type, will trigger this warning when compiling for certain architectures with at least GCC 6.2~6.4. Affected architectures are e.g. SPARC and ARM - not affected are x86 and ARM64. GCC bug ticket: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82270 It would be possible to work around this issue by changing the implementation of boost::move_detail::aligned_storage_impl so that the aligned type no longer carries the attribute(aligned) itself, but just contains a member which carries that attribute. If I understand this correctly this should have the same effect (barring compiler bugs). Question: Would a pull request for such a change be welcome? Regards, Paul Groke

2 1

Mini review of the Fit library
by Glen Fernandes 21 Sep '17

21 Sep '17

What follows is my mini review of the Fit library. I had no time to write a full review, but feel that most of my points have been covered by others in both the first and this second review. A. Should the library be accepted into Boost? Yes; I vote to accept Fit - with no conditions for acceptance. B. What is your evaluation of the design? I found it: 1. Practical: For facilities like BOOST_FIT_LIFT. 2. Innovative: For facilities like infix(). And overall useful. C. What is your evaluation of the implementation? Good. Appears to be correct C++ code, and is definitely written in a very maintainable way. Ordinarily I'm in favor of not seeing macros used for what is accomplished by FIT_INHERIT_DEFAULT, FIT_INHERIT_CONSTRUCTOR, and similar, but this is a minor thing, and it does serve to make the code smaller and easier to review. I also like that the lowest version of C++ standard support required is C++11 (and not higher), as I am still aware of a large enough pre-C++14 user base that do have access to and use Boost libraries. D. What is your evaluation of the documentation? Excellent. Better than what I generally write. Useful enough for me, and in my opinion, sufficient for understanding the library and learning how to use it - for any user that takes the time to read through it all. E. What is your evaluation of the potential usefulness of the library? I believe it to be useful, and I know of at one developer that I have suggested Fit to, that has started using it - and another that would like to start using it if it is included in a Boost release (for which their organization has blanket approval to use). F. Did you try to use the library? With which compiler(s)? Did you have any problems? Not recently, but Yes. With g++ 6.3. I encountered no problems. G. How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? I had spent at least 3+ hours looking into Fit with the intention of reviewing it for my own use, and learning how one part of it was implemented. I did not have a chance to spend as much time as I would normally have to write a full review of the library. H. Are you knowledgeable about the problem domain? Familiar with enough functional programming design and generic programming to appreciate the library. Other notes: No concern about the name. We have Beast, Hana, Spirit; Fit is a fine enough name, and boost::fit:: is a convenient enough prefix. Glen

2 1

docs to release a new lib
by Lorenzo Caminiti 21 Sep '17

21 Sep '17

Hello all, Where are the latest docs to add a new lib to Boost (including regression tests, PDF docs, etc.)? Poking around a bit I found the following links... https://svn.boost.org/trac10/wiki/StartModMaint https://svn.boost.org/trac10/wiki/DeveloperChecklist I'm asking because I'm considering adding Boost.Contract to the Boost 1.66.0 release. Thanks a lot! --Lorenzo

1 0

[review] Final Notice! Fit Review Ends Today!
by Matt Calabrese 20 Sep '17

20 Sep '17

Hi everyone, The formal review of the Fit library is ending today, September 20th. For anyone who was waiting until the last minute to submit a review, this would be it. Anything that you can provide is helpful in making a decision, even if you do not write a complete review of the entire library. You can find Fit for review at these links: Source: https://github.com/pfultz2/Fit/tree/boost Docs: http://pfultz2.github.io/Fit/doc/html/doc/ Please provide in your review whatever information you think is valuable to understand your final choice of ACCEPT or REJECT including Fit as a Boost library. Please be explicit about your decision. Some other questions you might want to consider answering: - 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 which compiler(s)? 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? - Were the concerns from the March 2016 review of Fit addressed? More information about the Boost Formal Review Process can be found here: http://www.boost.org/community/reviews.html Thank you for participating! -- -Matt Calabrese

1 0

Re: [boost] [review][Fit] Review of Fit starts today
by Jason Rice 19 Sep '17

19 Sep '17

Please consider this review for the proposed Boost.Fit library. I recommend the review manager ACCEPT this library for inclusion in Boost without condition. # What is your evaluation of the design? - Making high order functions easier and more concise is a good idea. - I like the point-free programming features. - Some of the names, conventions, and errors would take any kind of user some getting used to. - A version of `alias_value` that works in a generic context where the type might not be an alias would be more useful I think. # What is your evaluation of the implementation? - The code is consistent across implementations and the files are pretty flat - Each function can be included independently which is nice. - It appears that there is some effort towards making errors easier to read. For instance, I called `fit::partial` with too many arguments and the error was "error: no member named 'fit_function_param_limit'" - There is some readability issues due to compiler compatibility support and specific compiler error workarounds. - I noticed `fit::decay` is not used in the tests at all. It gets included in `example.h` but is never used. # What is your evaluation of the documentation? - I like that the documentation is easy to read in the comments of each file. - Some of the wording in the descriptions weren't very helpful for me personally, but the "Semantics" section seemed to always get the point of the function across. # What is your evaluation of the potential usefulness of the library? - It seems parts of the library might not be that useful if you have constexpr lambdas, but I believe there are still features that enable more concise coding even against latest C++17 features. - I will probably use this library now that I am familiar with it. # Did you try to use the library? With which compiler(s)? Did you have any problems? - I made a couple of my own examples. - I used a fairly recent fork of clang trunk with c++1z. - BOOST_FIT_RETURNS *crashed the compiler* when I was trying to construct an std::array with some pack expansion. It looked like it crashed the parser. # How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? - I spent a couple of hours looking at the docs in the source and making my own test programs to see how it could be used. # Are you knowledgeable about the problem domain? - I've used Boost.Hana a lot with its "functional" features. - I've done a lot of point free programming with RamdaJS in Javascript # Were the concerns from the March 2016 review of Fit addressed? N/A For what it's worth here is a link to some of my attempts to sample the library: https://gist.github.com/ricejasonf/9bfd233b3bc282eae9ecb8bc3198d948 Thank You, Jason Rice

2 1

[Container] small flat set ?
by Phil Endecott 18 Sep '17

18 Sep '17

Dear all, boost::container::small_vector and boost::container::static_vector are great. boost::container::flat_set is great too. But today I need a small_flat_set! And I don't think it's possible to compose flat_set and small_vector. (flat_set can use a custom allocator, but it can't take a custom implementation type in the way that, for example, std::priority_queue can.) Having said that, maybe composing flat_set and small_vector (or static_vector) isn't the optimum solution for the case where I will store something like 0...4 ints; I suspect that linear search in an unordered sequence will be quicker. But that gets complicated if you need iterators with the normal behaviour. Any thoughts? Regards, Phil.

5 14