- Boost - lists.boost.org

[gsoc18] Boost.Intrusive mentor request
by BogdanCodreanu 08 Mar '18

08 Mar '18

Hello, I've read about the Boost.Intrusive library and it matches my desires of programming. Is there a mentor available for this project? How should I prove myself worthy of this project? -- Sent from: http://boost.2283326.n4.nabble.com/Boost-Dev-f2600599.html

2 1

Re: [boost] [system] [teeks99] Mass test failures on MSVC-14.0
by Andrey Semashev 08 Mar '18

08 Mar '18

On 02/26/18 17:20, Tom Kent wrote: > On Mon, Feb 26, 2018 at 6:56 AM, Andrey Semashev via Boost > <boost(a)lists.boost.org <mailto:boost@lists.boost.org>> wrote: > > On 02/26/18 15:43, Tom Kent via Boost wrote: > > On Sat, Feb 24, 2018 at 3:45 PM, Andrey Semashev via Boost < > boost(a)lists.boost.org <mailto:boost@lists.boost.org>> wrote: > > Hi, > > Most of Boost.System tests are failing on the > teeks99-09-f-win2016-64on64 > tester, MSVC-14.0[1]. The test logs don't show any errors > and I can't > reproduce it locally. I'm seeing similar test failures in > other libraries, > like Boost.Atomic for example, and the links point to > Boost.System. > > Is it possible to get the error log from the tester? > > >> > > Here's all 82MB of it: > > http://boost.teeks99.com/temp/teeks99-09-f-results-bjam-2018-02-26.log > <http://boost.teeks99.com/temp/teeks99-09-f-results-bjam-2018-02-26.log> > > > > But be warned, I just moved the -09 runner onto a new (Win 2016) VM about a > > week ago, and there have been some teething issues. If all goes well, the > > -f run should hit again in the next 24-48hrs. > > The link returns 403: Forbidden. > > Try again now. Sorry about that. I got the log now, thanks. The relevant part seems to be this: msvc.link.dll D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.dll call "D:\teeks99-09\run\results\boost\bin.v2\standalone\msvc\msvc-14.0\adrs-mdl-64\archt-x86\msvc-setup.bat" amd64 >nul link /NOLOGO /INCREMENTAL:NO /DLL /DEBUG /MACHINE:X64 /MANIFEST /subsystem:console /out:"D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.dll" /IMPLIB:"D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.lib" @"D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.dll.rsp" if %ERRORLEVEL% NEQ 0 EXIT %ERRORLEVEL% Creating library D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.lib and object D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.exp msvc.manifest.dll D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.dll if exist "D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.dll.manifest" ( call "D:\teeks99-09\run\results\boost\bin.v2\standalone\msvc\msvc-14.0\adrs-mdl-64\archt-x86\msvc-setup.bat" amd64 >nul mt -nologo -manifest "D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.dll.manifest" "-outputresource:D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.dll;2" ) 'mt' is not recognized as an internal or external command, operable program or batch file. ...failed msvc.manifest.dll D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.dll D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.lib D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.pdb D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.pdb D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.pdb D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.pdb D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.pdb... ...removing D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.dll ...removing D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.lib ...removing D:\teeks99-09\run\results\boost\bin.v2\libs\system\build\msvc-14.0\dbg\adrs-mdl-64\thrd-mlt\boost_system-vc140-mt-gd-x64-1_67.pdb I think, this problem has been discussed on the list not long ago - `mt` being not found on the system after some MSVC update: http://boost.2283326.n4.nabble.com/build-msvc-14-linker-step-failing-tc4701… Looks like MSVC needs to be reinstalled or installed on a separate machine. Will you be able to fix this?

3 7

[release] 1.67.0 closed for Beta
by Daniel James 07 Mar '18

07 Mar '18

The master branch is now closed for all changes except by permission for the Beta release. Also, remember to add your libraries' changes to the release notes at: <https://github.com/boostorg/website/blob/master/feed/history/boost_1_67_0.q…> You can preview the release notes at: <http://www.boost.org/users/history/in_progress.html> Thanks, -- The Boost Release Team

1 0

About authorizing CircleCI
by Vicente J. Botet Escriba 07 Mar '18

07 Mar '18

Hi, recently I heard of CircleCI (https://circleci.com/) and it seems much faster than Travis. Is someone using it? What is your feedback? If the feedback is positive, could someone authorize CircleCI to access boostorg? Best, Vicente

6 11

[gsoc18] Mentor Request for uBLAS (prefer support for tensors)
by HiImKarl 07 Mar '18

07 Mar '18

Hi, I am Karl, a financial engineering student that is very passionate about C++. I think my background in applied math and stats would make me a great candidate for writing a tensor library. I would also like to point to this header-only multi-dimensional array template library with compile time tensor rank that I've written in the past: https://github.com/HiImKarl/Tensors, which may be good groundwork. It is a little late into March, so I was wondering if there were any mentors still looking for students for this project. I have already completed the competency test. -- Sent from: http://boost.2283326.n4.nabble.com/Boost-Dev-f2600599.html

2 1

[gsoc2018] Project 1.1: Add statistics and machine learning functions
by Dattatreya Mohapatra 07 Mar '18

07 Mar '18

Hi, My name is Dattatreya Mohapatra. I am an undergrad at IIIT Delhi, majoring in CS. I am interested in working on extending the Boost.uBLAS library as a project. I have taken courses on Linear Algebra and Machine Learning in the past, so I have a decent basic understanding of the subject. I have used C++11 in several of my course projects, so I have some familiarity with advanced features like lambdas and template programming. I have gone through the earlier threads on the same, and have started working on the core-competency test (CCT). Can the mentors of this project please guide me as to how should I get the CCT reviewed when its ready? Also, is there anything else I should look at apart from the documentation which would be helpful? Thanks and regards, Dattatreya Mohapatra

2 2

VS 2017 15.6 released, _MSC_VER=1913...
by Peter Dimov 07 Mar '18

07 Mar '18

... and Boost.Config once again warns: 1>c:\boost-git\develop\boost\config\compiler\visualc.hpp(349): note: Info: Boost.Config is older than your compiler version - probably nothing bad will happen - but you may wish to look for an updated Boost version. Define BOOST_CONFIG_SUPPRESS_OUTDATED_MESSAGE to suppress this message. Maybe we should display this only on 1920 and above, as I suggested last time? :-)

5 5

[gsoc18] [Boost.Intrusive] Mini version of project
by Rishija Mangla 07 Mar '18

07 Mar '18

Hello, My name is Rishija Mangla. I’m a mathematics graduate and pursuing my MCA (2nd year) from University of Delhi, India. I want to contribute in enhancing Boost.Intrusive library and have implemented a mini library for the same ( as asked for competency test ). Library includes: - Generic segment tree ( support range queries where operation is given by the user over any general class ) - Expression templates for lazy evaluation - Written test programs to check the accuracy ( using Boost.test unit test framework ) - Test cases are generated randomly Link to library implemented https://github.com/Rishija/boost_competency_test <https://github.com/Rishija/boost_competency_test> I plan to implement more complex data structures that have been mentioned in the list. I have already implemented prefix tree, both dynamic and via filesystem ( will provide link to the project in case you want to have a look ). So I’ll be pretty comfortable in developing other similar data structures and algorithms for them. Thanks

1 0

[Boost.Yap] Review results
by Louis Dionne 06 Mar '18

06 Mar '18

Dear Boost, I am pleased to announce that the Yap library is CONDITIONALLY ACCEPTED into Boost. I counted the following formal reviews: - Alex Hagen-Zanker: REJECT / CONDITIONAL ACCEPT - Barrett Adair: ACCEPT - Brook Milligan: CONDITIONAL ACCEPT - John Fletcher: ACCEPT - Paul Fultz II: ACCEPT - Steven Watanabe: ACCEPT I'd like to thank everyone who participated to the review for their feedback, and also Zach Laine for his submission. The following is a summary of the main issues that were raised during the review: - Contextual evaluation of terminals The concern was that there was no way of evaluating an expression tree with a custom context other than to first transform it with a custom transform, and then evaluate it. This, however, requires encoding the recursion into the transform even when we only want to modify terminals. After a lot of discussion, everybody seemed to agree that the following would be desirable (paraphrasing Steven): `transform_evaluate` would behave exactly like `transform`, except that the default behavior for nodes that are not handled by the context is to evaluate the operators instead of building a new expression. Zach fixed that in https://github.com/tzlaine/yap/issues/63 by introducing a new transform called `evaluation()` which evaluates terminals, and the ability to pass multiple transforms `t1`, ..., `tN` to `transform`. The first matching transform is executed, and `transform` will copy-and-recurse if no transforms match. `transform_evaluate(expr, tform)` can then be implemented as `yap::transform(expr, tform, yap::evaluation())`. - Auto-unwrapping of terminals in transform and automatic application of terminal transforms before unwrapping There was a concern that automatically unwrapping terminals in transforms and applying terminal transforms before unwrapping would result in counter intuitive behavior. After some discussion, it was determined that automatic application of transforms before unwrapping could indeed lead to surprising results and was removed. However, unwrapping of terminals is necessary to keep the matching of terminals reasonable, so it was kept. I am comfortable with that decision, as the unwrapping of terminals, which leads to a very intuitive way of matching terminals, was one of the features that sprung to me as being a great usability boost when I first looked at the library. - Positional placeholders should stay or go? Some of the reviewers felt like it was not desirable to provide positional placeholders as part of Yap, because those aren't the primary purpose of the library and C++14 obsoletes them with lambdas. Other reviewers felt like they were in fact very useful and shouldn't be removed. Since no clear consensus could be reached, no acceptance condition is attached to this issue. However, I personally favor the removal of placeholders from the core API on the basis of minimality. I would like to encourage Zach to check whether their removal would result in API simplifications; if so, I would encourage (but not require) their removal, at Zach's discretion. - Documentation The documentation is not sufficiently extensive, and takes too much for granted from the reader. This was raised by a couple of reviewers and it has been my largest problem with the library so far. More on this below. - The eval_xxx customization points. Nobody liked them, and they are already gone. - BOOST_YAP_CONVERSION_OPERATOR_TEMPLATE Nobody liked it, and it is already gone. - Short-circuiting of && and || in evaluate Steven brought up the fact that `yap::evaluate` would not short-circuit the evaluation of && and ||. This issue was mooted when customization points were removed. Also, the new `evaluation()` transform, which could suffer from the same short-circuiting problems, doesn't because it was implemented properly. - Support for Boost.Build is missing. - Hana vs not Hana Full disclosure: I am the author of Boost.Hana. A couple of reviewers raised concerns with Hana being part of the public interface. The main concern seemed to be about MSVC not being able to compile Hana, and hence Yap being unusable on MSVC. While I agree that MSVC is a big player, I believe that should not dictate what all Boost libraries do. Some libraries are meant to bridge gaps with older compilers, some libraries are cutting edge. Yap is the latter, and that's fine. The MSVC folks are working hard on their C++ conformance, and they will eventually get there. Also, it is likely that Hana wouldn't be the barrier anyway: Yap relies a lot on SFINAE and MSVC is known to be poor at it. Many of the features that one would likely want to use when using Yap are also not too well supported by MSVC, so users would probably hit limitations there anyway. Plus, as soon as MSVC is close enough, I'll start working on Hana's side to support it. Taking my review manager hat off for a second, I'd also like to point out that Yap is exactly the kind of use case that Hana was built for. Just like MPL was (and is) used in many libraries' implementation and APIs, I don't see a problem with using Hana in Yap's API, since that is the right tool for the job. That being said, if Zach wishes to explore using Hana's concepts (like hana::Sequence) to conceptify his API, I encourage him to do so. This may make it possible to use any tuple-like sequence for the children of an expression, thereby reducing the coupling between the two libraries. If Zach wants to explore using Fusion instead of Hana, that is also an option that would make MSVC support achievable. - A single macro for symmetric binary operators It was brought up that there's no macro to define both `arg <op> expr` and `expr <op> arg` at once. I believe this was initially done to avoid potential conflicts with the member variants, which already provide `expr <op> arg`. I believe it should still be possible to do it, since, as Steven said, most of the time a user will need symmetric behavior anyway. - External dependencies checked-into the repository The library includes internal copies of some of its dependencies, which is unusual for a Boost library. Zach said it wasn't his intent to put the library into Boost that way if it were accepted. To see all the issues that were raised during the review, you can search for tickets labeled with `Boost Review` on GitHub: https://github.com/tzlaine/yap/issues?q=label%3A%22Boost+Review%22 Based on the concerns above, the following lists the conditions for acceptance: - Documentation In my opinion, this was the biggest concern to be addressed. A lot of progress was made during and after the review, but I still think the documentation can be improved and made more complete. I especially believe that integrating more examples inside the "tutorial" (not just at the end) and introducing new concepts more gradually would help readers that are not already expression template gurus. It is very difficult to set a clear action item for "improving the documentation". For this reason, Brook Milligan and I will work with Zach to define what it means for the documentation to meet this acceptance condition. Brook was kind enough to agree to this when I asked him in a private communication. - A single macro for symmetric binary operators The library must provide a way to define symmetric binary operators in one go. Exactly how this can be achieved and whether we should be able to SFINAE-constrain the arguments is left to Zach's discretion. - External dependencies The external dependencies must not be packaged with the library itself. Another mechanism should be used; I leave that mechanism unspecified, but it should be possible to test the library using the usual Boost.Build workflow (e.g. an ExternalPackage download that would only work in CMake is not acceptable). - Boost.Build Yap needs to support Boost.Build like all the other Boost libraries. I will post an update to this list when the conditions of the review are met. If they wish to do so, reviewers can then verify that the conditions are met in a satisfactory fashion and the library will automatically be accepted unless someone objects. I would like to thank everyone who participated in the review, and congratulations to Zach for his library. Regards, Louis Dionne

5 7

[graph] r_c_shortest_paths writes/reads outside allocated memory
by Mateusz Polnik 06 Mar '18

06 Mar '18

Dear Colleagues, The algorithm for computing resource constrained shortest path writes/reads outside allocated memory which may result in segmentation faults. The finding is confirmed using Valgrind. I fixed the bug and opened a pull request. Could you review the pull request and give feedback if there is something more that should be done to meet quality requirements for the merge? The pull request https://github.com/boostorg/graph/pull/93 explains the issue, fixes it and contains a test for preventing regressions. Thank you in advance for your help. Regards, Mateusz Polnik

1 0