compacting macros

[proven totem]

https://github.com/Eczbek/xieite/blob/main/include/xieite/containers/tupleify.hpp
I have this ton of macros to generate repetitive code, but is there some way to reduce their quantity?

[echo hearth]

When your question is answered use !solved to mark the question as resolved.

Remember to ask specific questions, provide necessary details, and reduce your question to its simplest form. For tips on how to ask a good question use !howto ask.

[merry storm]

I think repr has a cursed macro for that

o nvm it's not finished https://github.com/Tsche/repr/blob/master/include/librepr/reflection/detail/visit_aggregate.h but I think @tawdry badger had a godbolt with some macros intended for it

[proven totem]

o

[proven totem]

@tawdry badger can you share pls?

[merry storm]

might of been @dapper garnet 's

[proven totem]

haha yes bring in more people to this thread

[dapper garnet]

@empty marlin show them

[proven totem]

lol

[dapper garnet]

No, I use empy to codegen the nasty shit. The only macro ways are either the boost pp way (hand written incdec macros to count up) or a beautiful but exp time solution baconheadcrab came up with that counts in unary

[proven totem]

my goal here is to be able to generate as many overloads of this function as is reasonable using only macros

because no external tools or prior configuration >:(

[dapper garnet]

If you care about compile times the only good way is to codegen beforehand

[proven totem]

well i did once

but the file took up like 3 megabytes?

this is already a big improvement but i think it could be even less

[tawdry badger]

i just wrote a codegen thing for it

https://github.com/8ightfold/godbolt-embeds/blob/main/deserializer/unpack_as_list.hpp

[proven totem]

hmm

so you'd recommend I just dont make macros?

yours is already 107 KB

[proven totem]

is it not possible to automatically detect the number of elements?

[merry storm]

you can automatically count a lot of aggregates members https://godbolt.org/z/odovb4ers

[dapper garnet]

Fwiw https://github.com/Tsche/repr/blob/master/include/librepr/reflection/detail/visit_aggregate.h.in this is the codegen source for repr

https://github.com/Tsche/repr/blob/master/tools/codegen.py and the codegen script

Sorry about the embeds, can't seem to remove them on mobile

[merry storm]

I wrote some absolute spaghetti codegen for a thing like this 😛

[dapper garnet]

Hey, that's @noble plover 's algo

[noble plover]

I knew I should have patented it

Mfw genuinely advanced c++ auto-reflection out of stubborness on a lol

[proven totem]

ooo

[proven totem]

"some_fairly_long_complex_identifier"

[dapper garnet]

Not that I care but I don't think that's an appropriate way to use otherwise licensed code @merry storm - that algorithm was committed under MIT license, so you need to at least mark it as such - just stating which project you lifted it from isn't sufficient afaik

[proven totem]

is it limited in identifier name length?

[proven totem]

gib source

[dapper garnet]

Especially given this looks more like an exact copy than a reimplementation 😛

[merry storm]

is hardly a copy paste but ya I'm no legal expert

[dapper garnet]

https://github.com/Tsche/repr/blob/master/include/librepr/reflection/detail/arity.h that link isn't going to be stable btw, been considering restructuring a bit sometime soon

[proven totem]

aa

[dapper garnet]

If you find someone who knows please let me know, would be good to actually know ig

But yeah same here, pretty much just guessing

[proven totem]

so if i reimplement this, who do i credit?

[noble plover]

me of course

[merry storm]

uhh ya prolly him 🙂

[proven totem]

aha

[dapper garnet]

I'd probably copy the license header from the project and place a permlink to the specific commit

[merry storm]

ya commit perma link would be an improvement at least

[proven totem]

what should i include? i generally just put something like ```cpp
// Thanks to lapinozz (https://github.com/lapinozz) for original code

[noble plover]

too bad you used such a permissive license on your mom

[dapper garnet]

Where you've taken it from 😛

[proven totem]

aha

internet decided to die :)

[proven totem]

from godbolt link

also wait this requires boost?

[empty marlin]

@proven totem you can use Boost.pp to implement the iteration

[proven totem]

no boost >:(

[empty marlin]

Then, open the library files, study how they work

[merry storm]

the arity/member counters shouldn't require boost

[proven totem]

boost is included in the godbolt link for some reaosn

[empty marlin]

It doesn't, but it's a starting point to understand arithmetic with the preprocessor

[noble plover]

when I made it it had 0 dependency

[merry storm]

o I'm not using any boost headers/boost code here

I'm not sure why the library was ticked on in the godbolt 🙂

[empty marlin]

I'm not sure I understand this sentence

[dapper garnet]

It still has zero dependencies in repr

[noble plover]

[dapper garnet]

I think I only did some slight refactorings to it

But yeah, repr is zero required dependencies

[noble plover]

my lawyers will contact your shortly

[merry storm]

I apparently had a library on in the godbolt (though I'm not using it) https://godbolt.org/z/odovb4ers

[dapper garnet]

The only thing you can prevent me from doing is change license

[noble plover]

can't even give you a kiss?

[dapper garnet]

Illegal

[merry storm]

no, just doing some scatter tests

[proven totem]

@noble plover should i link a particular repository of yours here? do i include its license?

[noble plover]

link to repr

[proven totem]

i'm in the process of rewriting it completely

[proven totem]

the arity thing?

[noble plover]

yeah

[proven totem]

got it 👍

[noble plover]

you can link to the original commit

[proven totem]

yeah

i'm following good programming manners :)

[tawdry badger]

pretty much

u only need it in one place

and expanding the macros takes up unnecessary time

[proven totem]

i may have found a better solution

if i can detect the number of elements at compiletime, i just need one overload

[tawdry badger]

u can

in 26 u can also use auto&& [args...] i think

[proven totem]

and so far it seems like it wont limit me to 256 elements?

[proven totem]

oh excellent, my library already has a requirement of c++26

wonder if gcc has implemented it yet though

[merry storm]

a chain of if constexpr's was limited to 121 in some compilers, though seems like an overload-approach accepts more

[tawdry badger]

its cool how far this has come

[proven totem]

is that pack indexing?

[tawdry badger]

https://github.com/8ightfold/godbolt-embeds/blob/main/deserializer/extended_bindings.hpp this was my original arity detection method for clang lmao

[proven totem]

i can just make a fold expression i think

[tawdry badger]

u should test the compile times

idk if the if constexpr is gonna be faster

[proven totem]

https://github.com/Eczbek/xieite/blob/main/include/xieite/functors/templatify.hpp

:)

taking a byte from runtime to compiletime

[tawdry badger]

shouldnt use static_cast<void>

[proven totem]

it surely has absolutely no downsides

[proven totem]

why

[tawdry badger]

literally just because its weird for the sake of it

ppl are used to (void)

[proven totem]

ewwwww

c-style cast

[tawdry badger]

thats the convention

[proven totem]

gross

fine maybe i'll consider it

opening an issue

[tawdry badger]

i still use c style casts in certain places

just not in templates

unless its a base cast

[echo hearth]

@proven totem Has your question been resolved? If so, type !solved :)

[proven totem]

no

[proven totem]

you use size_t outside the std namespace 🤮

[tawdry badger]

ong

[proven totem]

line 39

actually unreadable

[merry storm]

you're right, I should include stdint.h 😈

owait that's stddef.h in C

[tawdry badger]

cstddef

cstdint has the intX_ts

[proven totem]

[merry storm]

but ya that is one thing I'm not in the strong habit of qualifying and probably should be

[merry storm]

now this convo is dragging me away from the project I want to be finishing to fix some things in reflection 😦

[tawdry badger]

lucky it died...

[proven totem]

for now

[proven totem]

i dont see that

only attribute specifiers in c++26

[merry storm]

what's the phrase...

structured bindings can introduce a pack?

something something https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2023/p1061r5.html

[proven totem]

o

[merry storm]

probably not finished yet if it is going to make it in

[proven totem]

i see

[tawdry badger]

well

it should work here https://godbolt.org/z/vv3ssTMP5

but it just crashes lmao

[proven totem]

doesn't even work with a normal struct smh