February 2015 - Boost - lists.boost.org

Re: [boost] [Boost-interest] Biicode dependendy manager with Boost support
by Vicente J. Botet Escriba 27 Feb '15

27 Feb '15

Le 20/02/15 20:54, JOAQUIN M. LOPEZ MUÑOZ a écrit : > Hi, > > The folks at biicode (biicode.com) a C/C++ dependency manager, have just > announced support for Boost: > > http://blog.biicode.com/dependency-management-boost-libraries/#more-2086 > > Biicode allows for easy multiplatform component tracking and builiding > thanks to their file-based dependency detection and CMake autogeneration > magic. Now, these capabilities can be used to make it easier to work with > Boost libraries. > Thanks Joaquin for the announce. I was wondering if we couldn't use it as a starting point of our modularization activity. Some questions: * Is the file-based dependency manager what Boost is looking for its own module dependencies detection? * Could an author state explicitly its dependencies on other libraries (versions)? * It Biicode open source? * How its CMake autogeneration and bjam interact? Best, Vicente

4 4

VMD library accepted
by Steven Watanabe 27 Feb '15

27 Feb '15

AMDG The VMD library submitted by Edward Diener is ACCEPTED into Boost. I apologize for the long delay for this announcement. Thanks to everyone who participated in the formal review * Nial Douglass * Bjorn Reese * Rodrigo Madera * Zhou Zhenghui * Matt Calabrese * Paul Mensonides Reviews for the library were generally positive. The library was considered to be a useful set of tools for macro programmers. Paul expressed concern about the way the the library handles v-sequences and v-identifiers. After some discussion, Edward has agreed that the interface could be significantly improved. Work Items: * Add examples and improved motivation in the documentation * Update the v-identifier/v-sequence interface based on Paul's comments * Fix inaccurate statements about Boost.Preprocessor per Paul In Christ, Steven Watanabe

1 0

[1.58] Any plans to merge the super-project to master?
by John Maddock 26 Feb '15

26 Feb '15

Is the Boost super-project going to be merged to master, Jamroot in particular has several important unmerged changed. Thanks, John.

1 0

Re: [boost] [mpl] multiset
by Ganesh Prasad 25 Feb '15

25 Feb '15

> Aside from that, I?ve been also in contact with Louis discussing the > idea of porting metalog to Hana as a use case. Perhaps logic > predicates could prove a useful add on to Hana as well. Porting metalog to hana as an add on might be a good idea, but I quite disagree from the user's point of view. Hana is a beautifully designed library and it does precisely what it's designed for. In short it has good modularity. And , indeed, libraries are supposed to be highly modular. So, it might be better to keep metalog as a separate entity and provide good interoperability between Hana and metalog. Best wishes Ganesh Prasad

1 0

Cannot add any file to GIT "libgit2 returned: Invalid data in index - invalid entry"
by Paul A. Bristow 25 Feb '15

25 Feb '15

Cannot add any new file to GIT repo :-( "libgit2 returned: Invalid data in index - invalid entry" Have recently locally committed changes to an existing Boost.Math sub-module (commit OK). (But realized that I had forgotten to add a few new files, so tried to add and recommit, but failed at the add stage). git status shows expected untracked files and ends: nothing added to commit but untracked files present (use "git add" to track) I:\modular-boost\libs\math> .gitignore is the standard modular-boost issue and looks OK. I:\modular-boost\libs\math>git version git version 1.9.5.msysgit.0 Suggestions? Thanks Paul --- Paul A. Bristow Prizet Farmhouse Kendal UK LA8 8AB +44 (0) 1539 561830

2 3

[date_time] Re: [Boost-docs] Pull request #22 on boostorg/date_time
by Rene Rivera 25 Feb '15

25 Feb '15

You emailed the wrong list. On Tue, Feb 24, 2015 at 9:41 PM, Aaron Brooks <brooksaar(a)gmail.com> wrote: > Hi, > > Could I get someone to pull in my changes to date_time? > https://github.com/boostorg/date_time/pull/22 > > Just a small fix - fixes references to date_from_tm in the ptime doc, but > the bug has been open for 10 months and it tripped me up. > > Thanks, > Aaron > > _______________________________________________ > Boost-docs mailing list > Boost-docs(a)lists.boost.org > http://lists.boost.org/mailman/listinfo.cgi/boost-docs > > -- -- Rene Rivera -- Grafik - Don't Assume Anything -- Robot Dreams - http://robot-dreams.net -- rrivera/acm.org (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail

1 0

Re: [boost] [mpl] multiset
by Ganesh Prasad 25 Feb '15

25 Feb '15

> It is not my intention to deprecate MPL. I'm just looking for some > meta-programming utilities that can be proposed to the C++ standard. > > Vicente MPL and 'something else which does exactly the same thing' can co-exist (?), though users are familiar with MPL. Metaprogramming can be made easier with an alternative syntax, like Hana does. But functional is not the only solution. We can try with a Logic oriented syntax. C++ supports almost every major paradigm out there, except for logic (like prolog). And wrapping a logic oriented syntax over metaprograms would actually help programmers to handle complex inheritance patterns conveniently. Well, it's just an idea. I had tried with functional syntax (from scratch) about one year ago. And I have been thinking about a logic oriented syntax (I had even considered something like JQuery, though I rejected that idea eventually). It would be great if the logic programming idea could be acceptable for GSoC, I would be more than happy to work on it. Here is the link to my attempt at functional syntax over metaprogramming. https://github.com/GnsP/GNS--META_LISCPP Also, Bruno C. O. Dutra has worked in a very similar field, in fact he has already implemented a compile time logic library. Best Wishes Ganesh Prasad

2 1

[Context] RFR: Add support for ARM aarch64 architecture
by Edward Nevill 25 Feb '15

25 Feb '15

Hi, I hope this is the right place to post this. The following patch adds support for ARMs aarch64 architecture on Linux (there is existing iOS support for aarch64). I have run the context regression test on this on an aarch64 board running Ubuntu 14.04 with gcc 4.9.1 and it passes all OK. Could someone please review this patch and if OK push it? Many thanks, Ed. --- CUT HERE --- commit bd34672217ff3bcee185f9c7383179342ec2fe9b Author: Edward Nevill <edward.nevill(a)linaro.org> Date: Mon Feb 23 14:37:45 2015 +0000 Add support for ARM aarch64 architecture diff --git a/build/Jamfile.v2 b/build/Jamfile.v2 index 80d1c46..edd4400 100644 --- a/build/Jamfile.v2 +++ b/build/Jamfile.v2 @@ -253,6 +253,14 @@ alias asm_context_sources <toolset>darwin ; +alias asm_context_sources + : [ make asm/make_arm64_elf_gas.o : asm/make_arm64_elf_gas.S : @gas ] + [ make asm/jump_arm64_elf_gas.o : asm/jump_arm64_elf_gas.S : @gas ] + : <address-model>64 + <architecture>arm + <binary-format>elf + ; + # MIPS # MIPS/O32/ELF alias asm_context_sources diff --git a/config/arm.cpp b/config/arm.cpp index 6a83737..6eece0b 100644 --- a/config/arm.cpp +++ b/config/arm.cpp @@ -6,7 +6,7 @@ // accompanying file LICENSE_1_0.txt or copy at // http://www.boost.org/LICENSE_1_0.txt) -#if !defined(__arm__) && !defined(__thumb__) && \ +#if !defined(__aarch64__) && !defined(__arm__) && !defined(__thumb__) && \ !defined(__TARGET_ARCH_ARM) && !defined(__TARGET_ARCH_THUMB) && \ !defined(_ARM) && !defined(_M_ARM) #error "Not ARM" diff --git a/src/asm/jump_arm64_elf_gas.S b/src/asm/jump_arm64_elf_gas.S new file mode 100644 index 0000000..3e0e2c2 --- /dev/null +++ b/src/asm/jump_arm64_elf_gas.S @@ -0,0 +1,128 @@ +/******************************************************* + * * + * ------------------------------------------------- * + * | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | * + * ------------------------------------------------- * + * | 0x0 | 0x4 | 0x8 | 0xc | 0x10| 0x14| 0x18| 0x1c| * + * ------------------------------------------------- * + * | d8 | d9 | d10 | d11 | * + * ------------------------------------------------- * + * ------------------------------------------------- * + * | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | * + * ------------------------------------------------- * + * | 0x20| 0x24| 0x28| 0x2c| 0x30| 0x34| 0x38| 0x3c| * + * ------------------------------------------------- * + * | d12 | d13 | d14 | d15 | * + * ------------------------------------------------- * + * ------------------------------------------------- * + * | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | * + * ------------------------------------------------- * + * | 0x40| 0x44| 0x48| 0x4c| 0x50| 0x54| 0x58| 0x5c| * + * ------------------------------------------------- * + * | x19 | x20 | x21 | x22 | * + * ------------------------------------------------- * + * ------------------------------------------------- * + * | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | * + * ------------------------------------------------- * + * | 0x60| 0x64| 0x68| 0x6c| 0x70| 0x74| 0x78| 0x7c| * + * ------------------------------------------------- * + * | x23 | x24 | x25 | x26 | * + * ------------------------------------------------- * + * ------------------------------------------------- * + * | 32 | 33 | 34 | 35 | 36 | 37 | 38 | 39 | * + * ------------------------------------------------- * + * | 0x80| 0x84| 0x88| 0x8c| 0x90| 0x94| 0x98| 0x9c| * + * ------------------------------------------------- * + * | x27 | x28 | FP | LR | * + * ------------------------------------------------- * + * ------------------------------------------------- * + * | 40 | 41 | 42 | 43 | | | * + * ------------------------------------------------- * + * | 0xa0| 0xa4| 0xa8| 0xac| | | * + * ------------------------------------------------- * + * | PC | align | | | * + * ------------------------------------------------- * + * * + *******************************************************/ + + .cpu generic+fp+simd + .text + .align 2 + .global jump_fcontext + .type jump_fcontext, %function +jump_fcontext: + # prepare stack for GP + FPU + sub sp, sp, #0xb0 + +# Because gcc may save integer registers in fp registers across a +# function call we cannot skip saving the fp registers. +# +# Do not reinstate this test unless you fully understand what you +# are doing. +# +# # test if fpu env should be preserved +# cmp w3, #0 +# b.eq 1f + + # save d8 - d15 + stp d8, d9, [sp, #0x00] + stp d10, d11, [sp, #0x10] + stp d12, d13, [sp, #0x20] + stp d14, d15, [sp, #0x30] + +1: + + # save x19-x30 + stp x19, x20, [sp, #0x40] + stp x21, x22, [sp, #0x50] + stp x23, x24, [sp, #0x60] + stp x25, x26, [sp, #0x70] + stp x27, x28, [sp, #0x80] + stp x29, x30, [sp, #0x90] + + # save LR as PC + str x30, [sp, #0xa0] + + # store RSP (pointing to context-data) in first argument (x0). + # STR cannot have sp as a target register + mov x4, sp + str x4, [x0] + + # restore RSP (pointing to context-data) from A2 (x1) + mov sp, x1 + +# # test if fpu env should be preserved +# cmp w3, #0 +# b.eq 2f + + # load d8 - d15 + ldp d8, d9, [x0, #0x00] + ldp d10, d11, [x0, #0x10] + ldp d12, d13, [x0, #0x20] + ldp d14, d15, [x0, #0x30] + +2: + + # load x19-x30 + ldp x19, x20, [sp, #0x40] + ldp x21, x22, [sp, #0x50] + ldp x23, x24, [sp, #0x60] + ldp x25, x26, [sp, #0x70] + ldp x27, x28, [sp, #0x80] + ldp x29, x30, [sp, #0x90] + + # use third arg as return value after jump + # and as first arg in context function + mov x0, x2 + + # load pc + ldr x4, [sp, #0xa0] + + # restore stack from GP + FPU + add sp, sp, #0xb0 + + ret x4 + .size jump_fcontext,.-jump_fcontext + +# Mark that we don't need executable stack. +.section .note.GNU-stack,"",%progbits diff --git a/src/asm/make_arm64_elf_gas.S b/src/asm/make_arm64_elf_gas.S new file mode 100644 index 0000000..5a6914f --- /dev/null +++ b/src/asm/make_arm64_elf_gas.S @@ -0,0 +1,80 @@ +/******************************************************* + * * + * ------------------------------------------------- * + * | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | * + * ------------------------------------------------- * + * | 0x0 | 0x4 | 0x8 | 0xc | 0x10| 0x14| 0x18| 0x1c| * + * ------------------------------------------------- * + * | d8 | d9 | d10 | d11 | * + * ------------------------------------------------- * + * ------------------------------------------------- * + * | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | * + * ------------------------------------------------- * + * | 0x20| 0x24| 0x28| 0x2c| 0x30| 0x34| 0x38| 0x3c| * + * ------------------------------------------------- * + * | d12 | d13 | d14 | d15 | * + * ------------------------------------------------- * + * ------------------------------------------------- * + * | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | * + * ------------------------------------------------- * + * | 0x40| 0x44| 0x48| 0x4c| 0x50| 0x54| 0x58| 0x5c| * + * ------------------------------------------------- * + * | x19 | x20 | x21 | x22 | * + * ------------------------------------------------- * + * ------------------------------------------------- * + * | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | * + * ------------------------------------------------- * + * | 0x60| 0x64| 0x68| 0x6c| 0x70| 0x74| 0x78| 0x7c| * + * ------------------------------------------------- * + * | x23 | x24 | x25 | x26 | * + * ------------------------------------------------- * + * ------------------------------------------------- * + * | 32 | 33 | 34 | 35 | 36 | 37 | 38 | 39 | * + * ------------------------------------------------- * + * | 0x80| 0x84| 0x88| 0x8c| 0x90| 0x94| 0x98| 0x9c| * + * ------------------------------------------------- * + * | x27 | x28 | FP | LR | * + * ------------------------------------------------- * + * ------------------------------------------------- * + * | 40 | 41 | 42 | 43 | | | * + * ------------------------------------------------- * + * | 0xa0| 0xa4| 0xa8| 0xac| | | * + * ------------------------------------------------- * + * | PC | align | | | * + * ------------------------------------------------- * + * * + *******************************************************/ + + .cpu generic+fp+simd + .text + .align 2 + .global make_fcontext + .type make_fcontext, %function +make_fcontext: + # shift address in x0 (allocated stack) to lower 16 byte boundary + and x0, x0, ~0xF + + # reserve space for context-data on context-stack + sub x0, x0, #0xb0 + + # third arg of make_fcontext() == address of context-function + # store address as a PC to jump in + str x2, [x0, #0xa0] + + # save address of finish as return-address for context-function + # will be entered after context-function returns (LR register) + adr x1, finish + str x1, [x0, #0x98] + + ret x30 // return pointer to context-data (x0) + +finish: + # exit code is zero + mov x0, #0 + # exit application + bl _exit + + .size make_fcontext,.-make_fcontext + +# Mark that we don't need executable stack. +.section .note.GNU-stack,"",%progbits

2 7

Friend class with namespace gives compiler error for C6000 TI
by Alin Popa 25 Feb '15

25 Feb '15

Hi, I compile my project with boost header only files, the compiler is C6000 Texas Instruments. The compile errors occurs for example in "boost/algorithm/string/find_iterator.hpp" I have extracted the cause of the problem in the next example code: namespace boost { namespace iterators { class iterator_core_access { public: iterator_core_access(){} ~iterator_core_access(){} }; } using iterators::iterator_core_access; } class testFriend { private: // don't work, compiler error: error: namespace "boost" has no actual member "iterator_core_access" friend class ::boost::iterator_core_access; // here is the problem // it works, compiles ok, this was my cure //friend class ::boost::iterators::iterator_core_access; }; int main () { // it works ::boost::iterator_core_access a; return 0; } Do you guys know what is the reason of this ? With msvc and g++ works fine. Thanks and best regards, Alin

2 1

[build] check __aarch64__ Aarch64
by Oliver Kowalke 24 Feb '15

24 Feb '15

Hi, I've added an pull request for jam.h (build/engine) to set os.platform correctly to 'ARM' on Aarch64 architecture. This is required in order to build boost.context correctly on the64bit ARM platform. thx, Oliver

1 0

Any online C++ compiler with Boost version to select?
by Boris Schäling 23 Feb '15

23 Feb '15

Does anyone know of an online C++ compiler which does not only come with Boost but allows you to select the Boost version? While http://coliru.stacked-crooked.com/ comes with Boost 1.57.0, http://ideone.com/ and http://cpp.sh/ with 1.55.0 and http://codepad.org/ with 1.34.1 (really!), I can't even tell whether these websites support Boost explicitly or whether it's more of an accident that I can include Boost headers. :) http://isocpp.org/blog/2013/01/online-c-compilers also mentions http://liveworkspace.org/ and http://gcc.godbolt.org/. But I can't open these websites (not sure whether they are temporarily broken or gone for good?). Boris

3 3

[config] test macro for std::index_sequence<>?
by Oliver Kowalke 23 Feb '15

23 Feb '15

Hi, I can not find an macro (BOOST_NO_CXX11_...) for testing the absence of std::index_sequence<> - may I have overlooked it? so long, Oliver

2 2

Re: [boost] [gsoc15] Attn: We need 2015 mentors and project ideas for Boost!
by Antonio Colino 22 Feb '15

22 Feb '15

Here is a random thought for a project idea. Agner fog created a vectorization lib for the sse and avx instruction sets as part of a project called vector class. Intel created the equivalent opencl kernels as part of a project that is called ispc ( Intel simd program compiler ). Research both and submit a single vector lib that auto vectorizes avx, sse both gpu and cpu overloading the std::vector class operators.Just a random thought. Intel SPMD Program Compiler)

1 0

[mpl] Fix to Ticket #11001
by Bruno Dutra 22 Feb '15

22 Feb '15

I just finished a pull request <https://github.com/boostorg/mpl/pull/14> with a fix for Ticket #11001 <https://svn.boost.org/trac/boost/ticket/11001>, which, by the way, I reported myself, so I'm quite confident it's fixed. Could one of the maintainers kindly review it? Oh and it is worth it to mention that I also included a new test case to ensure the fix is working. Kind regards, *Bruno C. O. Dutra*

1 1

GSoC: Enhanced vector and deque containers
by Benedek Thaler 22 Feb '15

22 Feb '15

Hi, I'm looking at the "Enhanced vector and deque containers"[0] item of the GSoC projects. I have to following interface idea: boost::devector: Like std::vector, but the first call to push_back inserts the element to the middle of the buffer. Likewise, reserve puts the contents to the middle of the new buffer (if any). This allows a cheap push_front, pop_front, push_back, pop_back. boost::deque: Like std:deque, but segment size can be specified in the constructor. To make buffer operations easier, a `segment_end` would be provided: iterator segment_end(iterator begin) Which returns the end of the largest consecutive range starting at `begin`. I'm also thinking about the following twist: Assuming this layout: [a,b,c,d] [e,f,g,_] Brackets denote segments, letters are elements, underscore is empty space. Here, inserting `x` after `d` normally results several moves and this layout: [a,b,c,d] [x,e,f,g] But instead, we could allocate a new segment: [a,b,c,d] [_,x,_,_] [e,f,g,_] To make the insertion cheaper. This behavior is similar to a segmented list, i.e: std::list<boost::devector<T>> Are these similar to the "extra functionality" mentioned in the docs you were thinking about? Thanks, Benedek [0]: https://svn.boost.org/trac/boost/wiki/SoC2015#a5.Enhancedvectoranddequecont…

1 0

Enquiry about gsoc project Enhanced vector and deque containers
by Jasdeep Singh 21 Feb '15

21 Feb '15

I had a few doubts regarding this project and was wondering if someone could help me out.It would be much appreciated.

1 0

[MultiIndex] Memory consumption of ordered indices
by Olaf van der Spek 21 Feb '15

21 Feb '15

Hi, What's the memory consumption of ordered indices? Is it implemented as a linked list, so two pointers per element? In that case for a write-only data structure an external ptr vector would make more sense wouldn't it? -- Olaf

3 4

[filesystem] temp_directory_path() behavior on Windows
by Beman Dawes 20 Feb '15

20 Feb '15

temp_directory_path() on Windows is currently implemented by calling the Windows API GetTempPath() function. Ticket https://svn.boost.org/trac/boost/ticket/5300 points out that GetTempPath() does not work as expected for environmental variables longer than roughly 130 characters. I've added a test to Boost.Filesystem that verifies that boost::filesystem::temp_directory() is affected. The suggested fix is to use GetEnvironmentVariable to in effect implement GetTempPath() the way we would like it to work. That makes sense to me, so I looked up the specs for GetTempPath(): The GetTempPath function checks for the existence of environment variables in the following order and uses the first path found: 1. The path specified by the TMP environment variable. 2. The path specified by the TEMP environment variable. 3. The path specified by the USERPROFILE environment variable. 4. The Windows directory. (4) Makes me queasy. I suspect it will fail anyhow because of permissions errors, but I would prefer to have temp_directory_path() report an error if 1, 2, and 3 are not present. temp_directory_path() already reports an error if a path is found but it is not a directory. OTOH, excluding the Windows directory is a breaking change for anyone currently depending on that behavior, so I thought it best to ask for comments before charging ahead. Thoughts? --Beman

8 39

[log] custom architecture and address-model
by Vladimir Prus 19 Feb '15

19 Feb '15

Hi, it seems that Boost.Log defines, and checks, it's own log-address-model and log-architecture and log-instruction-set properties. Naturally, if I'm trying to make regular address-model and architecture properties always set, these are getting in the way. Do we really need per-library properties? If a user wants to build Boost.Log with a particular architecture, I would think he can do: ./b2 --with-log <whatever> If so, is everybody fine if I eliminate these custom features eventually? -- Vladimir Prus CodeSourcery / Mentor Embedded http://vladimirprus.com

3 6

[Review] Boost.Endian mini-review
by Joel FALCOU 19 Feb '15

19 Feb '15

The mini-review of Boost.Endian starts today and runs through Sunday, February 1. The library repository is available on GitHub at https://github.com/boostorg/endian The docs for the library can be viewed online at https://boostorg.github.io/endian/ The objective of the mini-review is to verify that the issues raised during the formal review have been addressed. Those issues are documented at https://boostorg.github.io/endian/mini_review_topics.html which also describes the resolution of each issue. Comments and suggestions unrelated to the mini-review issues are also welcome, but the key question the review manager needs your opinion on is this: Is the library ready to be added to Boost releases? -- Joel Falcou, Endian Review Manager

16 57

[endian] Post-mini-review status: endian ready to merge to master
by Beman Dawes 18 Feb '15

18 Feb '15

Here is a post-mini-review status report to support a pull request to merge endian into the boost super-project master branch as a submodule. (It has been a submodule of the super-project develop branch for months.) The mini-review ran from January 23rd through February 1st. There were over 50 messages posted in the main review thread ("[Review] Boost.Endian mini-review"). The key question Joel Falcou, the review manger, asked was: Is the library ready to be added to Boost releases? All seven responses to the question were "Yes" votes (Peter Dimov, Nevin Liber, Philip Bennefall, Jason Newton, Jeremy Maitin-Shepard, Cliff Green, Bjorn Reese), and none of these votes were conditional. Other things that came out of the mini-review discussion: * Major concern: The default alignment needs to revert to unaligned. There should be no problem completing this before the final 1.58.0 release, and probably before the beta, but I need to carve out a quiet couple of hours to review the discussions before actually doing the revert. * Major concern: Floating point support needs an additional test program tailored to the characteristics of floating point numbers. A warning needs to be added to the docs alerting users that FP support is much less mature than integer support. The plan is to create the new test program in the next few days, and ask on the list for contributions of test cases from FP experts. * Requests for new features: A couple of simple to analyze and implement features have already been added. Other possible new features were added to a "Future directions" section of the docs so they won't be forgotten. * Technical fixes and suggestions. Most of these have already been applied. * Documentation fixes and suggestions. Most of these have already been applied, with three or four docs change emails still to be processed. Today is the cut off for 1.58.0 new libraries, so I'm planning to submit pull requests to merge endian submodule support from the boost super-project's develop branch into its master branch. (endian submodule support has been in the boost super-project's develop branch for months.) Thanks to all of those who contributed to the mini-review! --Beman

1 0

[GSoC] interested in implementing numerical algorithms for uBLAS
by Ganesh Prasad 17 Feb '15

17 Feb '15

Hi I am interested in implementing numerical algorithms optimized for distributed/parallel architectures for Boost.uBLAS. I have been applying for this project for previous 2 years, so I have definitely done some research in this field. Also, I have been working on Clusters for last 7 months as an Academic project (Product Development). So, basically I am acquainted with programming for distributed architectures with MPI. best wishes Ganesh Prasad

2 1

[threads] Win32 Events on POSIX
by Fu ji 17 Feb '15

17 Feb '15

Hello, I want to port win32 events on Posix using Boost. To avoid the re-invented the wheel I use PEvents: https://github.com/neosmart/pevents which is on MIT licence. But with my current knowledge I do some mistakes because I have deadlock. Maybe someone with bigger knowledge about differencies in boost and pevents might help with this ? I know that there is also thread-safe boost.signals2 with observer pattern but it does not provide WaitForMultipleObjects functionality. Best regards.

5 16

[nowide] Bug fix for older Windows versions
by Michael Smith 17 Feb '15

17 Feb '15

I found an issue with the implementation of cout/cerr in Boost.Nowide. It will fail in the middle of writing output due to a segfault in VerifyConsoleIoHandle (during WriteConsoleW). See the attached files for a patch and example (I've tested with MinGW-w64 4.8.3).

1 0

Fwd: Boost.Fixed_Point for GSoC 2015?
by Nikhar Agrawal 17 Feb '15

17 Feb '15

Hi everyone, I'm Nikhar Agrawal, currently an undergraduate student at IIIT-Hyderabad pursuing a degree in Computer Science and Engineering. I have been a succesful GSoC participant twice in a row. In 2013, I did GSoC under Boost C++ libraries under the mentorship of Christopher Kormanyos with continuous guidance from John Maddock and Paul Bristow. In 2014, I did GSoC under GNOME. In 2014, I had also submitted a proposal for Boost.Fixed_Point. Here's the link to the proposal I submitted last year: https://docs.google.com/document/d/1Ylc_OJBpoN_lNgMrYh0l-O_qCb-EDJQy0JGoh5u… Unfortunately, the project couldn't go through because of lack of slots. I would like to try again this year. Christopher isn't available as a mentor though. Chris and Niall suggested that I float a mail through the mailing list to see if there are any takers for this project who'd be willing to mentor it. The proposal hasn't changed much from what I submitted last year. We may have to limit it's scope to something that'd be ideal for one summer though. I'll be more than happy to discuss it with anyone who's interested in the project. Thanks. Nikhar P.S.: I'm attaching some mails from the conversation I've had with Niall and Christopher in this regard. ---------- Forwarded message ---------- From: Niall Douglas <s_sourceforge(a)nedprod.com> Date: Wed, Feb 11, 2015 at 8:12 AM Subject: Re: Boost.Fixed_Point for GSoC 2015? To: nikharagrawal2006(a)gmail.com Cc: Christopher Kormanyos <e_float(a)yahoo.com> Nikhar, Firstly my commendation to you on keeping up interest after not getting a place last year. For your information, we requested 12 slots and received just 7, and after I begged Google for another we got eight for 2014. That meant four projects did not get funding, and yours was one. Can I suggest that you perhaps consider approaching potential mentors in Boost and see if you can find one interested? Please do CC my email in your approaches so they know I have vetted you, and mention you were a successful student in 2013. Given that Chris won't be available this year, you may need to consider not doing fixed point as I am unaware of a suitable alternative mentor who is available. When approaching potential mentors, try to suggest ideas that are something the mentor badly wants in their particular specialism. Try searching the mailing lists and past GSoC ideas to see which mentors might have which pet projects which get them enthusiastic. Just this week a student called Damian Vicino managed to persuade Robert Ramey to mentor work on safe_numerics. I did not know Robert had an interest in this, but Damian obviously figured it out and won a mentor. Good luck in your efforts. Niall On 10 Feb 2015 at 21:24, Christopher Kormanyos wrote: > Hi Niall, > I will not be available as a GSoC mentor this year. > Just a heads-up, Nikhar Agrawal, a successful veteranGSoC candidate, has expressed interest in thefixed-point project. > I do not know if anyone would like to mentor it. > > See text below. > Thank you for your continued support. > Best regards, Chris > > > It seems that the discussions for GSoC 2015 have begun> on the mailing list. > > The fixed point project couldn't make through last year.> Is there interest in it for this year from you guys?> Would you like me to post about it on the mailing list?> I'm still very much interested in this. I contacted about this> project a couple of times last year after the summers> but everyone was busy with other stuff. And I think the> GSoC deadlines provide a lot of motivation anyways. > So let me know what you guys think about it and> if I should apply for the project. > Cheers. :)> Nikhar > Hi Nikhar, > Thank you for keeping in touch. > I will not be available as a GSoC mentor this year.In my opinion, the fixed-point project is a good one.We just need to find a mentor for the project. > The fixed-point project needs to have a clearly definedscope. The preliminary specification suggested byLawrence Crowl represents man-years of coding effort. > > If there is a way to limit the scope of the project toa reasonable level for one summer and find a mentor,then go for it ! > Please post any suggestions and or comments or questionson the board. > You were a very successful participant in GSoC 2013.With help from John and Paul, the ground work you did hasbeen included in Boost. We now have Bernoulli, high-precisiontgamma(), and also polygamma() for integer order. > > Boost is a long-term endeavor. I encourage you to stickwith it --- and keep up the good work ! > Good luck. > > Cheers, Chris > > > On Tuesday, February 10, 2015 4:29 PM, Nikhar Agrawal < nikharagrawal2006(a)gmail.com> wrote: > > > Hi everyone, > How are you all doing? > > It seems that the discussions for GSoC 2015 have begun on the mailing list. > The fixed point project couldn't make through last year. Is there interest in it for this year from you guys? Would you like me to post about it on the mailing list? I'm still very much interested in this. I contacted about this project a couple of times last year after the summers but everyone was busy with other stuff. And I think the GSoC deadlines provide a lot of motivation anyways. > So let me know what you guys think about it and if I should apply for the project. > Cheers. :)Nikhar > > On Tue, Apr 22, 2014 at 1:25 AM, Nikhar Agrawal < nikharagrawal2006(a)gmail.com> wrote: > > Hi all, > GSoC results are out and unfortunately, the fixed-point project didn't go through. However, the last couple of months or so, having invested time in discussing various design issues and going through various code prototypes, I have become really interested in developing a fixed-point library. And I stick to my word of being willing to develop library outside of GSoC as well. It would be great to have your continued guidance. Though with GSoC in summer, I wonder you'd have your official projects as first priority. I had submitted another proposal to GNOME and I got into GSoC via that as well. > So, I was wondering what could be a good way forward? How I could contribute to fixed-point library? The first question is if there's an interest in the library still from all your sides as well now that it's not an official GSoC project? If yes, I guess we could continue our collaborations just like they happened during last GSoC? Though probably less frequently? What do you guys think? > Also, if it's not an inappropriate thing to ask, could you kindly elaborate why the project didn't go through? Seeing the comments of everyone involved, I had a feeling that proposal was very well received. > Thanks.Nikhar > > > > > > > > --- Boost C++ Libraries Google Summer of Code 2015 admin https://svn.boost.org/trac/boost/wiki/SoC2015

1 0