- Boost - lists.preview.boost.org

[gsoc-18]
by Anurag Tangri 14 Feb '18

14 Feb '18

Hello Boost community ! My name is Anurag Tangri. I am a 3rd year undergrad at Thapar University. I am very much familiar with C++14 because of my competitive programming background.I have decent knowledge of Data structures and algorithms and eagerly want to contribute in open source. I am new to open source and stuff. So it would be very helpful if someone could guide me.

1 0

[release] 1.67.0 deadline for new libraries approaching
by Daniel James 14 Feb '18

14 Feb '18

One week from today master will closed for new libraries and breaking changes. Thanks, -- The Boost Release Team

2 1

[release notes] Linking to github issues/pull requests
by Daniel James 14 Feb '18

14 Feb '18

Hi, I've added templates for linking to github issues and pull requests from the release notes. There are two templates: [github module number] which expands to '#number' [github_pr module number] which expands to 'PR#number'. Both will be linked, they need the name of the module to link to the right place. We only really need the issue template as github redirects to the pull request if appropriate, but existing release notes have been using the PR#3 format, so I thought it should support that. You can see an example in the unordered release notes for 1.67.0: https://github.com/boostorg/website/blob/master/feed/history/boost_1_67_0.q… http://www.boost.org/users/history/in_progress.html Does that look okay? Daniel

5 17

[yap] Review
by P F 14 Feb '18

14 Feb '18

- Whether you believe the library should be accepted into Boost I think it should be Accepted. - Your name Paul Fultz II - What is your evaluation of the library's: The design is very nice and simple to follow. I have found Proto to be a little hard for me to grep, but this seems fairly simple to follow. I think the transform -> evalute workflow is great. Also, with transforms being just function objects, it makes transformations very composable. I think this library is very useful. I would love to use this for some projects at work, but the dependency on Hana prevents that(I need to support gcc 5 and MSVC). Looking at the codebase, it could easily be replaced with Boost.Fit(once its included in the release) in the future. Of course, I dont expect a reimplementation without Hana as a condition for acceptance. However, I think Hana at least can be removed from the "interface". The expression elements could be expressed as a `std::tuple` and the placeholders could be a `std::integral_constant`. This should make it easier to use another backend without breaking API compatibility. Some other notes: - Couldn't the operator macros be provided as base classes? So instead of writing: template <boost::yap::expr_kind Kind, typename Tuple> struct user_expr { static const boost::yap::expr_kind kind = Kind; Tuple elements; // Member operator overloads for operator!(). BOOST_YAP_USER_UNARY_OPERATOR_MEMBER(logical_not, ::user_expr) }; The user could write: template <boost::yap::expr_kind Kind, typename Tuple> struct user_expr : logical_not<::user_expr> { static const boost::yap::expr_kind kind = Kind; Tuple elements; }; You could also get rid of the need for `::user_expr` by just taking it as a type and using something similiar to `mp_transform` to change the parameters. Or even change `Kind` to be a type of `std::integral_constant<boost::yap::expr_kind, Kind>` so `mp_transform` could be used directly: template <typename Kind, typename Tuple> struct user_expr : logical_not<user_expr> { static const boost::yap::expr_kind kind = Kind{}; Tuple elements; }; - The customization points seem unnecessary. Why wouldn't the user just use `transform`? Using `transform` also keeps the customizations local. Also, implicit transformations seems like it would lead to suprises(and not good suprises). Are those on the cutting block? - It would be interesting to provide common transformations, like common subexpression elimination, or dead code elimination, but I guess that is beyond the scope of this library. - The `op_string` function is missing the case for `expr_ref` and `terminal`. Also, the tests only check for plus. - The `make_expr_from_tuple` does a move here: auto make_expr_from_tuple (ExprTemplate<Kind, OldTuple> const & expr, NewTuple && tuple) { return ExprTemplate<Kind, NewTuple>{std::move(tuple)}; } But maybe this should be using `std::forward`? - Did you attempt to use the library? If so: Yes, I ran it on clang 3.8 and gcc 7. - How much effort did you put into your evaluation of the review? Several hours

3 8

Regarding how to get selected for gsoc boost
by Syed Souban 13 Feb '18

13 Feb '18

Hi! I am Syed Souban. I am new to the world of open source and I want to contribute to it. I am a Computer Science And Engineering Undergraduate and I know C, C++, Java, Android App Development, HTML, CSS, JavaScript, PHP. I want some guidance on how should I start preparing to get selected for GSoC in your organisation.

2 1

[system] virtual ~error_category
by Peter Dimov 12 Feb '18

12 Feb '18

The virtual destructor of error_category creates problems for constexpr-ification because it makes the class non-literal. I can't think of a scenario in which we need the destructor to be virtual. It's probably virtual for "guideline" reasons - the class has virtual functions. But does it really need to be? Error categories aren't destroyed via a pointer to base. They are, in fact, static instances, not destroyed at all, except on program exit. Can anyone think of a case in which error_category needs the virtual destructor?

7 16

Boost C++ is a mentoring organization for GSoC'18 !
by David Bellot 12 Feb '18

12 Feb '18

We just received this message from Google. It says it all ! *Congratulations! Boost C++ Libraries has been selected as a Google Summer of Code 2018 mentor organization.* This year again, we will be a mentoring organization for the Google Summer of Code. If you want to be a mentor, please contact me right now and register on the Google Summer of Code website too. Best regards, David

1 0

[yap] Review
by Fletcher, John P 12 Feb '18

12 Feb '18

I have quite a lot to say and I may not get it all down in one go. > - Whether you believe the library should be accepted into Boost Accept. > - Your name John Fletcher > - Your knowledge of the problem domain I have been working on codes doing similar things for a long time. I started with FC++ and have worked with Phoenix (I did some maintenance) and a lot of other codes. I have also worked with Hana around the time of that review. I tend to work by trying things out and extending the examples given, which is what I have been doing with YAP. I suppose you can categorize me as someone working at an intermediate level, attempting to make use of libraries to make results available to users at a higher level in a way which is intuitive. As I have worked in the past with the old C++ I particularly appreciate new codes which take advantage of C++14 to make this work easier. > You are strongly encouraged to also provide additional information: > - What is your evaluation of the library's: > * Design I like the design and have found this grow on me as I worked with the examples and saw how the pieces work together. I have found some of the recent discussion about why things are how they are very helpful. > * Implementation I have encountered quite a lot of the implementation as I have worked on the examples. I appreciate that this library makes extensive use of Boost Hana. There are some places in the examples where the user needs to have an understanding of Boost Hana to work. For example the transform example for arity uses Boost Hana maximum and transform. I will except boost::hana::tuple from the comment as it is used so much and is similar to std::tuple. One issue I have had is with the namespace. I am not clear why some code needs the full name e.g. boost::yap::evaluate and in other places evaluate will do without using being defined. It would be good if that was consistent. > * Documentation I don't think that the documentation at present does full justice to the code. There are several things which I found very useful which are not much mentioned in the code. When I saw that the examples were taken from work with proto I set out to make one which was missing, the mini_lambda example. I have had a lot of success with that while having to dig quite hard to find information. make_expression_function I have found this very useful in moveing to a functional approach. It is not much mentioned. I think what is needed is more examples and some contributions from users as to what help is needed to understand how to get the best use out of YAP. > * Tests I have already mentioned working on a version of the proto mini lambda. I have also built and example of the member call operator as the only example was in a more complicated example. I have also extended the vector example to add some cases which are missing e.g. 3. * v3 and v3 * v . I have made it use v3*v3 to return a scalar product. > * Usefulness I have found this to be useful from the examples. > - Did you attempt to use the library? If so: Yes > * Which compiler(s)? Mainly clang 4.0 C++14 with corresponding libc++. I have run some examples with gcc 7.1 C++14. This work is on ubuntu linux. > * What was the experience? Any problems? Problems encountered have been with my own understanding and wanting to explore the boundaries. > - How much effort did you put into your evaluation of the review? I have spent many hours spread over several weeks. I started before the start of the review. I hope this is helpful. Zach, thank you for your work and please ask if anything is not clear. Best wishes John Fletcher

2 1

[build] msvc-14 linker step failing?
by John Maddock 12 Feb '18

12 Feb '18

I have, and have had for some time: using msvc : 14.0 : "C:/Program Files (x86)/Microsoft Visual Studio 14.0/VC/bin/cl.exe" ; In my user-config.jam However, with current develop this leads to the linker step always failing: msvc.manifest ..\..\..\bin.v2\libs\config\test\config_info.test\msvc-14.0s11\debug\address-model-64\link-static\runtime-link-static\config_info.exe 'mt' is not recognized as an internal or external command, operable program or batch file. if exist "..\..\..\bin.v2\libs\config\test\config_info.test\msvc-14.0s11\debug\address-model-64\link-static\runtime-link-static\config_info.exe.manifest" ( call "..\..\..\bin.v2\standalone\msvc\msvc-14.0s11\address-model-64\architecture-x86\msvc-setup.bat" amd64 >nul mt -nologo -manifest "..\..\..\bin.v2\libs\config\test\config_info.test\msvc-14.0s11\debug\address-model-64\link-static\runtime-link-static\config_info.exe.manifest" "-outputresource:..\..\..\bin.v2\libs\config\test\config_info.test\msvc-14.0s11\debug\address-model-64\link-static\runtime-link-static\config_info.exe;1" ) As you can see no tool name is set on the start of the command line. Anyone any ideas? Thanks! John. --- This email has been checked for viruses by AVG. http://www.avg.com

5 14

[Heads up] PR adding guidelines for breaking changes on the website
by Louis Dionne 11 Feb '18

11 Feb '18

Folks, In early December 2017, there was a thread asking for guidelines regarding breaking changes: [1]. The conclusion reached in that thread seemed to be agreed upon [2] and I suggested making it part of the library guidelines. I have finally submitted a PR [3] to add it to the library _guidelines_ (not _requirements_). Feel free to take a look at the PR and chime in as this concerns everyone. Cheers, Louis [1]: https://lists.boost.org/Archives/boost/2017/12/240071.php [2]: https://lists.boost.org/Archives/boost/2017/12/240108.php [3]: https://github.com/boostorg/website/pull/300

4 3