February 2015 - Boost - lists.boost.org

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