13:27 <hkaiser> simbergm: I'm now seeing 'CMake Error at src/CMakeLists.txt:360 (target_link_options):

13:27 <hkaiser> target_link_options called with invalid arguments' here https://github.com/STEllAR-GROUP/hpx/blob/master/src/CMakeLists.txt#L360

13:29 <simbergm> hkaiser: hmm, I think it might need a `PUBLIC` keyword

13:29 <simbergm> would you mind trying that?

13:30 <hkaiser> yah, that works - however PRIVATE should be fine

13:31 <simbergm> or that if that's enough

13:31 <hkaiser> I can create a PR

13:32 <simbergm> I don't know if that flag needs to be present in dependent projects

13:32 <simbergm> thanks!

13:33 <hkaiser> no, it's just needed to link the core library

13:39 <hkaiser> simbergm: #4440

13:40 <hkaiser> that has one more change also

13:41 <hkaiser> I'd appreciate it if this could be merged asap

13:44 <simbergm> hkaiser: already merged

13:44 <simbergm> thanks a lot!

13:54 <hkaiser> simbergm: thanks!

14:12 <hkaiser> simbergm: wrt #4434

14:12 <hkaiser> I think you missed at least two fixes

14:13 <simbergm> hkaiser: :(

14:13 <simbergm> tell me where :)

14:14 <hkaiser> the add_custom_target() change and the missing template arguments for basic_string<CharT, ???>

14:19 <simbergm> hkaiser: indeed, thanks

14:19 <simbergm> I've pushed again

14:20 <simbergm> hkaiser: heads up, I just merged the PR to change the way we handle the c++ standard as well as it's closely related to the target PR

14:21 <simbergm> let me know about any problems as usual

14:36 <hkaiser> simbergm: ok - give me a sec

14:36 <simbergm> hkaiser: ah, no rush

14:41 <diehlpk_work> parsa, Do you have any AGAS plot?

14:41 <diehlpk_work> parsa, I have new two 10124 nodes runs for you

14:42 <diehlpk_work> Still waiting for 1500 to 1900 runs

14:42 <diehlpk_work> hkaiser, I could finish two 1024 node runs and it seems with the older version of octotiger everythings works again

14:43 <hkaiser> ok, cool - are you sure that the older version is ok to use?

14:43 <hkaiser> what did Dominic say?

14:43 <diehlpk_work> We used it before already.

14:44 <hkaiser> ok

14:44 <diehlpk_work> I just accidentaly updated it while I compiled the newer APEX version

14:44 <diehlpk_work> My script checked out master and not the previous version

14:45 <diehlpk_work> The latest version had the radiation merged and something caused an issue

14:46 <hkaiser> great, I'm glad you resolved the problem

14:49 <diehlpk_work> hkaiser, Can you add some text for hpx and hpx's performance counters by tomorrow morning?

14:49 <diehlpk_work> it would really help if out group did something

14:49 <hkaiser> diehlpk_work: will try, could you give me the link to the document one more time, pls

14:52 <diehlpk_work> hkaiser, see pm

16:03 <hkaiser> diehlpk_work: done

16:09 <hkaiser> simbergm: our circleci runs for phylanx now fails horribly: https://app.circleci.com/pipelines/github/STEllAR-GROUP/phylanx/11/workflows/4eaad2a3-567c-4aac-850c-cc5163c1e6b0/jobs/34884 :-(

16:12 <simbergm> hkaiser: you do something funny with the cmake in your docker image? it says 3.11, whereas ours has 3.13

16:12 <diehlpk_work> hkaiser, parsa Eitherone of you broke the LaTex compilation

16:12 <simbergm> (and I only bumped the version checks in the c++ standard PR, which probably isn't on dockerhub yet)

16:13 <simbergm> hkaiser: my point is that we require 3.13, except that it's not enforced in our cmake in that commit

16:20 <hkaiser> diehlpk_work: compiled for me :/

16:20 <hkaiser> ahh, so it's a cmake problem? we can certainly go up to 3.13

16:20 <hkaiser> parsa: yt?

16:20 <diehlpk_work> Ok, it us your part using utf-8 characters, I will fix it

16:21 <hkaiser> diehlpk_work: thanks a lot - probably a copy&paste snafu

16:35 <simbergm> hkaiser: yep, once 4378 gets pushed it'll complain much earlier

16:35 <hkaiser> k

16:35 <hkaiser> simbergm: do we have 3.13 in the base image now?

16:36 <hkaiser> I'll talk to parsa to make it's being used

16:36 <hkaiser> make sure

16:36 <simbergm> hkaiser: yep, we do

16:36 <hkaiser> ok

16:37 <simbergm> that's why I was surprised you have 3.11 for phylanx :)

16:37 <simbergm> but I guess it was to have something newer than the ancient one we used to have in the base image

16:37 <hkaiser> we installed that explicitly before hpx was updated as we needed a newer version

17:20 <hkaiser> diehlpk_work: yt?

17:21 <diehlpk_work> yes

17:21 <hkaiser> see pm, pls

17:26 <parsa> hkaiser: we had that ancient cmake issue with highfive, so we installed cmake 3.11 (and raised the minimum we require) temporarily until we came up with a more permanent solution

17:26 <hkaiser> parsa: yah, 3.13 is now in the HPX base image and our 3.11 causes problems

17:26 <hkaiser> would you be able to remove our special handling, please?

17:27 <parsa> i'll remove the 3.11 installation, but maybe we should also raise the minimum to 3.13

17:27 <hkaiser> indeed, we should

17:27 <hkaiser> I can do that once the image has been updated

17:47 <parsa> hkaiser: #1136 should do the job

17:47 <hkaiser> thanks

19:40 <diehlpk_work> hkaiser, I started the cross listing of the course in EE

19:40 <diehlpk_work> Can you get in touch with Golden for CS?

19:40 <diehlpk_work> Course selection will start soon

22:17 <gonidelis> hkaiser sorry for the late response. The last couple of days I was studying the C++ STL concepts like containres, iterators and ranges. Must admit, ranges were quite hard to comprehend but I have catched the basic ideas (any extra sources on ranges would be appreciated). Tomorrow I will start working on Range Based Parallel Algorithms by picking

22:17 <gonidelis> one and trying to implement it. Let me address some questions though: 1) I have studied the all_any_none_of commit as an example and I think I will try to imitate the way it was implemented in order to adapt one the missing algos from here(https://github.com/STEllAR-GROUP/hpx/issues/1668). The question is: What is the difference between the three

22:17 <gonidelis> head (hpp) files:

22:17 <gonidelis> #include <hpx/parallel/algorithms/all_any_none.hpp>#include <hpx/parallel/container_algorithms/all_any_none.hpp>

22:17 <gonidelis> hpx/include/parallel_all_any_none_of.hpp

22:18 <hkaiser> our range algorithms are just wrappers around the iterator based ones at this point

22:18 <hkaiser> so it's fairly easy to create those

22:18 <hkaiser> the third is just a forwarding header, nothing special

22:19 <hkaiser> the first two have the actual implementation and the range wrappers in them

22:20 <gonidelis> by saying wrappers you mean....?

22:21 <gonidelis> oh nevermind! I get it!

22:30 <gonidelis> Under what directory does the thirs header reside? I can see that `hpx/parallel/algorithms/` exists no more on master...

22:30 <gonidelis> third*

22:32 <gonidelis> sory i mean the first* header, as you can see from the directory that I mention, my bad.

22:40 <hkaiser> hpx/hpx/include

22:40 <hkaiser> it's an old directory that might goa away at some point

22:46 <gonidelis> While all_any_none.hpp resides under hpx/libs/algorithms/include/hpx/parallel/algorithms...

22:57 <hkaiser> yes, the algorithms are all in the algorithms module (under libs/algorithms)