hkaiser changed the topic of #ste||ar to: STE||AR: Systems Technology, Emergent Parallelism, and Algorithm Research | stellar.cct.lsu.edu | HPX: A cure for performance impaired parallel applications | github.com/STEllAR-GROUP/hpx | Buildbot: http://rostam.cct.lsu.edu/ | Log: http://irclog.cct.lsu.edu/ | GSoD: https://developers.google.com/season-of-docs/
nikunj97 has joined #ste||ar
K-ballo has quit [Quit: K-ballo]
hkaiser has quit [Quit: bye]
Coldblackice_ has joined #ste||ar
Coldblackice is now known as Guest51721
Coldblackice_ is now known as Coldblackice
Guest51721 has quit [Ping timeout: 268 seconds]
quaz0r has quit [Ping timeout: 245 seconds]
quaz0r has joined #ste||ar
tarzeau has joined #ste||ar
<tarzeau> if i wanted to generate a debian/copyright file of hpx 1.3.0, how could i find the copyright years+authors+email of the stuff with BSL-1.0 license
<tarzeau> i'll just take 2007-2018 Louisiana State University i guess
<tarzeau> and maybe the ones in CMakeLists.txt
K-ballo has joined #ste||ar
<mdiers_> heller: yt?
<heller> mdiers_: Hey
<mdiers_> heller: hi, there is pr #3928 that fixes our problems in combination with slurm. maybe there are others that could test that? i don't want to cause anything serious there.
hkaiser has joined #ste||ar
hkaiser has quit [Quit: bye]
<parsa> wash[m]: ping
<heller> mdiers_: thanks, I'll have a look ASAP
<heller> tarzeau: copyright ste||ar-group would be better
<mdiers_> heller: thanks
<mdiers_> heller: I have some problems/questions with sanitizing. Are you available early next week?
<tarzeau> if someone has debian/ubuntu and wants to review it: http://phd-sid.ethz.ch/debian/hpx/
<tarzeau> i was thinking deb/ub needs something too looking at https://repology.org/project/hpx/versions
hkaiser has joined #ste||ar
<Yorlik> What would be a good and typesafe/-correct way to implement a set of flags in a field of bits?
<Yorlik> (in some uint64_t for example or with a configurable underlying type)
<Yorlik> I'm playing with enum and enum class types but running into oddities when oring or assigning values.
<Yorlik> It's more a theoretical interest as in how to think about flags in a bitfield type structure.
<Yorlik> My current thought is, that using any enums for the bitfield is fundamentally wrong,
<Yorlik> because it would be too hard to represent all flag combinations.
<Yorlik> Ideas/Suggestions?
<Yorlik> I'm currently messing around with this and that's where I saw the issues: https://godbolt.org/z/Ghe4h_
<Yorlik> My current thought is if I really want to have this correct I'd have to make a flagset class derived from a bitset and override the bitwise operators to only allow certain parameter types which would be either another flagset or any of the allowed single flags from an enum class.
<Yorlik> Maybe it's just crazy - no idea :/
<hkaiser> Yorlik: I usually use enums in combination with functions that set/reset/check single flags
<Yorlik> So - the compiler might just optmize away the call?
<Yorlik> I just wondered how to protect against illegal use of uint42_t and other things people might do.
<Yorlik> Even played with the thought to derive from std:bitset and ovveride operator|, etc ...
<Yorlik> Even played with the thought to derive from std:bitset and overide operator|, etc ...
<Yorlik> woops
<Yorlik> That was a Discord reflex ;)
<hkaiser> sure, if you inline your functions they most probably go away
<Yorlik> I mean - I have something that works, but ... well .. trying to do things the C++ way :/
diehlpk_mobile has joined #ste||ar
<diehlpk_mobile> hkaiser proposal is submitted. However, no e-mail confirmation yet
<diehlpk_mobile> I will forward the PDF of the uploaded proposal with the confirmation email later
<Yorlik> hkaiser: Ended up with something along this as a possible "solution" (not saying I like it ...): https://godbolt.org/z/PQ1ayN
<hkaiser> why dynamic_cast?
<hkaiser> Yorlik: ^^
<hkaiser> and you probably mean operator|(*this, f) (note the star)
<Yorlik> Isn't dynamic checking more in depth? I might getting it wronmg here - C++ catsa are on my current learning wall
<Yorlik> err - not cats - casts :)
<hkaiser> static_cast should be just fine
* Yorlik needs to read more :)
<Yorlik> hkaiser: I'll probably just keep using simple integral types and drop remarks in comments for my fellows. :) This was more for learning. Thanks for the heads up on the cast.
diehlpk_mobile has quit [Ping timeout: 252 seconds]
<hkaiser> and the missing star
<hkaiser> this is something entirely different from *this
<Yorlik> woops :)
<Yorlik> This is funny: Why did it not just explode ? Someone insid ethe compiler must have done some implicit conversion for me ...
<Yorlik> like create a temporary copy ...
<hkaiser> Yorlik: well, you dynamic_cast it, so it does not check whether its valid
<Yorlik> works even with static
<Yorlik> _cast
<hkaiser> ok, interesting
<Yorlik> I wonder if there would be a trick along the lines of crtp maybe to simply allow subclass types as parameters for all operators of a std::bitset
<Yorlik> After all the purpose is just to add restrictions to what you are alloowed to OR in or AND out
<Yorlik> a set of constexpr can serve as replacement for an enum of flags
<Yorlik> I'm just imagining using a bunch of different sets of flags and someone oring in a flag which should not exist in that set, but accidentally has the same value
<hkaiser> heller: yt?
nikunj97 has quit [Ping timeout: 250 seconds]
hkaiser has quit [Quit: bye]
<lsl88> hi!
<lsl88> I have written a draft of my proposal for GSoD
<K-ballo> buenas tardes
<lsl88> K-ballo: hahah buenas tardes ;)
<K-ballo> lsl88: I hear you are from Argentina.. me too, Bahía Blanca, born and still living there
<lsl88> oh, I know Bahía Blanca, nice city :)
<lsl88> I live in Buenos Aires city, which is a nice city but too crowded IMHO
hkaiser has joined #ste||ar
<lsl88> parsa: please, let me know when you are available :)
<parsa> lsl88: hellp
<parsa> hello*
<lsl88> hi!
<parsa> i haven't had a chance to read your proposal... i'll take a look this evening
<heller> hkaiser: hey
<lsl88> sure, no problem, there is still time :) simbergm and dihelpk have been discussing some ideas with me. I don't know how much I should expand the proposal
<hkaiser> heller: hey
<hkaiser> sorry, we can't find a common time
<heller> Seems so
<hkaiser> would you have time over the weekend?
<heller> I think I can make some
<hkaiser> well, let's play it by ear
<heller> Should we try tomorrow at the same time again?
<hkaiser> ok, might actually work
<heller> Or earlier/later?
<hkaiser> no 2pm should be fine
<heller> Great
<hkaiser> ok, ttyl
<heller> ttyl
<wash[m]> parsa: pong, hit me up on Monday about the callback thing, I've identified who I need to talk to
<parsa> wash[m]: will do. thanks a lot!
diehlpk_mobile has joined #ste||ar
<Yorlik> hkaiser: things slowly going into the right direction (line 199 refuses to compile as wanted): https://godbolt.org/z/gRXQob
diehlpk_mobile has quit [Ping timeout: 252 seconds]
nikunj has quit [Remote host closed the connection]