diehlpk_work has quit [Remote host closed the connection]
diehlpk_work has joined #ste||ar
gonidelis[m] has quit [Ping timeout: 260 seconds]
gonidelis[m] has joined #ste||ar
ms[m] has quit [Ping timeout: 260 seconds]
jbjnr1 has quit [Ping timeout: 260 seconds]
ms[m] has joined #ste||ar
jbjnr1 has joined #ste||ar
nikunj97 has joined #ste||ar
<tiagofg[m]>
Hello,
<tiagofg[m]>
How do I do for creating a class component which is derived from two other base class component?
karame_ has quit [Remote host closed the connection]
<tiagofg[m]>
The only examples I found are just a class derived from another class, one to one. What I'm looking for is a class component derived from two or more classes components
<tiagofg[m]>
maybe you guys can help on that
<tiagofg[m]>
For now, what I know is that I need to define the wrapping_type, type_holder, base_type_holder and some more like this stuff. Besides that, I need to use HPX_REGISTER_DERIVED_COMPONENT_FACTORY
<ms[m]>
heller would maybe know? ^
hkaiser has joined #ste||ar
weilewei has quit [Remote host closed the connection]
<ms[m]>
hkaiser: good morning
<hkaiser>
hey ms[m]
<ms[m]>
I didn't do anything about the linker errors on the actions PR... did you still want help with that?
<hkaiser>
ms[m]: yes, please
<hkaiser>
I have not looked at this since we talked either
<ms[m]>
ok, will try to do something about it
<hkaiser>
thanks!
<hkaiser>
it's just on the old clang, I don't have access to that
Nikunj__ has joined #ste||ar
nikunj97 has quit [Ping timeout: 265 seconds]
nikunj97 has joined #ste||ar
Nikunj__ has quit [Ping timeout: 260 seconds]
nikunj97 has quit [Read error: Connection reset by peer]
Nikunj__ has joined #ste||ar
nikunj97 has joined #ste||ar
Nikunj__ has quit [Ping timeout: 260 seconds]
<hkaiser>
ms[m]: how do I disable the tests for NVCC? Does cmake have a variable to know it's being compiled with it?
<hkaiser>
Or should I change the code using __NVCC__?
<gonidelis[m]>
hkaiser: apologies for not responding yesterday, I was away
<hkaiser>
np
<gonidelis[m]>
so what was it?
<hkaiser>
pls have a look at #4821 (and #4822)
<hkaiser>
I think something similar should be done for for_each now
<gonidelis[m]>
oh ok! I saw that message... I am checking on them right now actually
<hkaiser>
:D
<ms[m]>
hkaiser: hmm... maybe `if(NOT HPX_WITH_CUDA_CLANG)`? I don't think we specifically have a variable for that, but if the cuda modules are enabled it's either clang or nvcc
<hkaiser>
ok, thanks
diehlpk__ has joined #ste||ar
<ms[m]>
hkaiser: I wanted to check if the future serialization error happens on windows as well (since it's even more secretive about symbols in libraries I would expect it to happen there as well), so I enabled one of the tests that tries to serialize futures on the windows builders
<ms[m]>
the address of that static function was not unique across shared libraries if built with hidden visibility
<ms[m]>
not sure if there's another way to provide a per-type stable id to use there
<hkaiser>
darn, you're right
<hkaiser>
how did you solve it?
<hkaiser>
excellent catch, btw!
<ms[m]>
for now one can explicitly set "default" visbility just where it's needed with pragmas (unintuitively enough it can't be set just around the static function definition or its use, but it has to be further down)
<ms[m]>
I'
<hkaiser>
ms[m]: we have the EXPORT_ALWAYS (or somsuch)
<hkaiser>
we use that for exceptions
<ms[m]>
that didn't seem to have an effect either... I might be putting the export in the wrong place though
<hkaiser>
HPX_ALWAYS_EXPORT
<hkaiser>
another option would be to explicitly instantiate that template for the 3 archive data types we use in a cpp file and expose the address through a member function that is defined in that cpp file
<ms[m]>
right, that might actually be cleaner
<ms[m]>
will try that out as well
<ms[m]>
just pushed a fix to the actions pr as well, seems to build ok now with networking off (it was another missing header)
<hkaiser>
ms[m]: as expected - thanks!
<hkaiser>
I really appreciate your help!
<hkaiser>
gonidelis[m]: btw, you might want to read https://wg21.link/p1895 for more context
<gonidelis[m]>
hkaiser: got it ;)
<gonidelis[m]>
hkaiser: So just to get this clear, what you are trying to accomplish with #4821, is the thing we were talking about on Friday?
<ms[m]>
not sure if the rebase was 100% clean though, but if it looks ok later go ahead and merge it
<hkaiser>
yes
<hkaiser>
that and more
<hkaiser>
the use of customization points will allow for easy extensibility of our algorithms in the future
<gonidelis[m]>
Yeah, I have some questions on the "and more" part
<hkaiser>
ms[m]: thanks
nanmiao11 has joined #ste||ar
<hkaiser>
gonidelis[m]: let's talk whenever you want to
<hkaiser>
ms[m]: btw, I'm seeing a ton of deprecation warnings for result_of now
mcopik has joined #ste||ar
<hkaiser>
should we enable a way of supressing those?
mcopik has quit [Client Quit]
<hkaiser>
something like 'define HPX_IGNORE_RESULT_OF_DEPRECATION_WARNING to acknowledge you've seen this warning'
<ms[m]>
I'd rather not... but if you think it's worth it sure
<ms[m]>
what is the warning flag that reports it? (linux or windows?)
<ms[m]>
you can always do -Wno-X at least on windows
<ms[m]>
in the end they're just warnings, but warnings for a good reason
<ms[m]>
or just build with HPX_WITH_DEPRECATION_WARNINGS=OFF
<hkaiser>
windows, it's reported at the namespace hpx { using util::result_of; }, btw
<hkaiser>
we might not need that using to begin with
<ms[m]>
ah, yes, that's a false positive
<ms[m]>
we can disable the warning there, but you're right about removing that altogether
<ms[m]>
doesn't make much sense to add that typedef if it's deprecated
<ms[m]>
it's a new typedef anyway
<hkaiser>
yes
<diehlpk_work>
hkaiser, Meeting?
karame_ has joined #ste||ar
<hkaiser>
nanmiao11: please see #1210, this should fix your problems with __ne etc.
<hkaiser>
diehlpk_work: sorry, can't do it today?
<diehlpk_work>
hkaiser, Katie already told us
<nanmiao11>
hkaiser Thanks!
<nanmiao11>
hkaiser meeting at 11:30?
<hkaiser>
nanmiao11: can we meet 15 minutes later, please?
<hkaiser>
11:45am?
<nanmiao11>
sure
<hkaiser>
nanmiao11: I'm ready whenever you are
<nanmiao11>
Will join
akheir has joined #ste||ar
nikunj has quit [Ping timeout: 256 seconds]
nikunj has joined #ste||ar
diehlpk__ has quit [Ping timeout: 246 seconds]
mcopik has joined #ste||ar
mcopik has quit [Remote host closed the connection]