12:55 <Yorlik> Quick Question: hpx::new_ is nothing like new right? So - if a component goes out of scope and all shared pointers to it are released it should be detroyed, is that correct ?

12:56 <hkaiser> yes

12:56 <hkaiser> components are garbage collected

12:58 <Yorlik> Would calling a component constructor locally work correctly or is it mandatory to use hpx::new_ to have correct registered AGAS ID, etc ?

13:01 <hkaiser> you have to either use new_ or local_new

13:01 <hkaiser> hoever new_<> with a local id will dispatch to local_new anyways

13:03 <Yorlik> Allright - thanks

14:17 <simbergm> hkaiser: do you know if ali is working on rostam at the moment? buildbot is down but not sure if I can restart it

15:15 <Yorlik> Is Catch2 usable to make tests using hpx components?

15:20 <Yorlik> What am I doing wrong here?: return hpx::components::local_new<GameObject>(hpx::find_here()); gives me C:\__A\arc_2017\_INSTALL\hpx\stable\Debug\include\hpx\runtime\components\server\component.hpp(73): error C2664: 'agns::data::GameObject::GameObject(agns::data::GameObject &&)': cannot convert argument 1 from '_Ty' to 'const agns::data::GameObject &'

15:36 <rori> K-ballo: Hey ! I'm trying to find BoostSIMD to check if my cmake changes work and it appears like the github repo is no longer maintained, do you know anything about this ?

16:10 <zao> Proposed Boost.SIMD? :)

16:12 <zao> rori: Judging by the issue and reddit thread I find, it sounds like they decided to go properitary partway into the process and pulled out.

16:12 <zao> Sounds like there was a lot of drama on the list too.

16:13 <rori> yeah actually I found it, it is now called nsimd and it has the same templates we need, only the headers are changing

16:13 <zao> Sounds like their proposed replacement is going to be binary-only bullshit.

16:14 <zao> https://github.com/agenium-scale/boost.simd/issues/545#issuecomment-422435374

16:14 <rori> ah yep

16:17 <parsa> if you're looking for a SIMD library just want to mention agner fogg's released VCL 2 just yesterday

16:17 <parsa> https://github.com/vectorclass/version2

16:18 <zao> parsa: You might be the person to ask out of the blue. Do you happen to know if BLAS/LAPACK implementations have any constraint on whether they may mess with or tolerate changes in the floating point environment like SSE rounding mode or overflow tracking?

16:19 <parsa> zao: sorry, dunno

16:19 <zao> Alright, wild shot :)

16:29 <hkaiser> zao: reminds me to remove support for Boost.SIMD from HPX

16:29 <hkaiser> rori: ^^

16:29 <zao> hkaiser: They went offline.

16:29 <hkaiser> I know

16:29 <zao> (I guess there's logs :) )

16:29 <hkaiser> Joel was totally pissed about how the Boost discussion went a while ago

20:57 <hkaiser> aserio: yt

21:47 <aserio> hkaiser: yes

21:47 <aserio> sorry

21:49 <hkaiser> aserio: see pm, pls

21:54 <diehlpk_work> hkaiser, aserio https://developers.google.com/season-of-docs/docs/participants/

21:55 <diehlpk_work> They announced the Google Season of Docs participants

22:18 <hkaiser> diehlpk_work: nice!

22:23 <diehlpk_work> Rebecca will start writing a blog post about her project

22:45 <parsa> is this file supposed to have a .good extension? https://github.com/STEllAR-GROUP/hpx/blob/457670e0581fe03ee4b4b4b12b0a757eff445307/plugins/parcelport/verbs/rdma/RdmaCompletionChannel.h.good

22:46 <hkaiser> parsa: ask jbjnr_

22:47 <parsa> jbjnr_: ^