09:17 <ms[m]> bhumit: the additional `:` pre and suffixes are a bit of a turnoff for me at least, so I'd probably go with the other option

09:18 <ms[m]> essentially we'd like to be able to link to a whole row of these tables: https://hpx-docs.stellar-group.org/latest/html/manual/optimizing_hpx_applications.html#id10, but in practice you'd end up linking probably to the beginning of the first column of each row

09:18 <ms[m]> note that if you think those would be better presented in some other form than a table, that's also possible, we're not particularly attached to the current way of presenting them

09:56 <bhumit[m]> ms makes sense. The table form looks good to me as lf now. Sent a PR with an example change (#5609). If it looks fine to you, I'll get on adding other links.

09:57 <bhumit[m]> * change ([#5609, * #5609](https://github.com/STEllAR-GROUP/hpx/pull/5609), * )). If

09:58 <bhumit[m]> s/ms makes sense. The table form looks good to me as lf now. Sent a PR with an example change (#5609). If it looks fine to you, I'll get on adding other links./ms makes sense. The table form looks good to me as of now. Sent a PR with an example change ([#5609](https://github.com/STEllAR-GROUP/hpx/pull/5609)). If it looks fine to you, I'll get on adding other links./

10:02 <bhumit[m]> This is how ^ looks with the change, by the way.

10:04 <rori[m]> bhumit: I think it the first line of #5609 looks good but for what I understood the ref should be in other places that are referring to this counter in particular no? But I might have misunderstood the point

10:04 <rori[m]> hkaiser?

10:07 <rori[m]> Actually forget what I said, it looks like it's exactly what ms suggested:

10:07 <rori[m]> https://github.com/STEllAR-GROUP/hpx/issues/4495#issuecomment-613976964

10:08 <bhumit[m]> Haha cool, I was confused too : P

10:08 <rori[m]> sorry and thanks for your work :)

10:09 <bhumit[m]> rori[m]: absolutely no need to apologize and welcome : )

10:09 <bhumit[m]> * apologize and you're welcome :

12:53 <Yorlik> hkaiser: yt?

13:56 <pedro_barbosa[m]> Should I install valgrind in the HPX directory in order to use it? I keep getting this error... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/1305e45a332c6b0c5048924d988d7d5b5d485a44)

14:28 <gonidelis[m]> So I 've changed whatever instance of pointer to a vector but it get's 10^2 slower. Check my gprof file

14:28 <gonidelis[m]> https://gist.github.com/gonidelis/be803cf25c6bc4468e568cfdfa9f2416

14:28 <gonidelis[m]> hkaiser: ^^

14:28 <gonidelis[m]> vector allocation take up 10% of the time

14:28 <gonidelis[m]> 20% ^^

14:34 <gonidelis[m]> It's probably easier to debug, but doesn't it make sense to lose performance?

15:08 <hkaiser> gonidelis[m]: let's have a look at the code whether we can circumvent this

16:34 <gonidelis[m]> hkaiser_: i will ping you

16:37 <gnikunj[m]> hkaiser_ yt?

16:48 <hkaiser_> gnikunj[m], gonidelis[m]: here

17:18 <gonidelis[m]> How can I comment out an assertion and it still hits?

17:22 <hkaiser_> lol

17:22 <hkaiser_> you either commented a different assert or you didn't rebuild

17:25 <gonidelis[m]> I didn't make clean as it seems

17:25 <gonidelis[m]> I hate caches

17:50 <gonidelis[m]> how very much possible is that I don't get what I want if I `reinterpret_cast<const char**>` a `vector<vecotr<char>>`?

17:50 <gonidelis[m]> K-ballo:

18:40 <gonidelis[m]> I mean, how does `reinterpret_cast<const std::pair<long, long> *>` work on a `const char **`

18:40 <gonidelis[m]> ?

18:40 <gonidelis[m]> K-ballo: ^^

23:44 <gonidelis[m]> hkaiser: do we embrace our `task_block`s ? ;p

23:47 <hkaiser> gonidelis[m]: sure, why not