daemon status says "update needed" on 0.16.0.0 cli. it shouldn't because that's the latest, no?

woodser: yea will be fixed with next release

I'm having a hard time getting and keeping peer connections to my daemon on stagenet.

It'll get a connection, then it'll drop. I have port forwarding enabled for 38080.

luigi1111w: also supercop repo PR 3 please if you are ok with it

and that Windows CI PR

nvm... just saw that 6582 was merged!

Looks like CI is progressing smoothly

vtnerd_ around?

luigi1111w: paste.debian.net/hidden/8b16462b

ah wrong link

was the gui one

luigi1111w: cli paste.debian.net/hidden/80e823ac

you got mac built for gui?

yep

also not sure that's as good as my existing description

yes very good is also good

thanks

🍄

yes

an hour late, but around

luigi1111w ^

just painting some bikesheds

oh yeah the boost thing, Ill try to come up with something that is the least gross

no supercop

why not ge25519_scalarmult and move the doublings to their own function or the host function, or ge25519_scalarmult8 ?

the iamamyth comment? yeah I need to respond

ah that one too

but the mooo comment above it

also a follow up comment on the thread about comments being wrong would be appreciated

ah I somehow missed that comment about renaming the function, and yeah it would be for the better

will get to these because this is kind of useful

and if your other question was about why not calling ge25519_scalarmult ... because there is none for these libs

the only constant time mul function is fixed to the basepoint because the signing libraries never need constant-time arbitrary point mul

we could switch to the variable time functions which would be faster, but with the downside that this theoretically does change the possible viewkey scanning behavior

look at ge25519.h in either variant folder

no, not calling that function, naming it that

yeah ok I wasnt certain if you were suggesting changing the function too at the same itme

just wrt clarity surrounding the mul8

ge25519_scalarmult8 is a more accurate name and can be reused

fine by me

ohh I see what you meant now sorry

yeah if they were separate it would be even more useful

because then it exposes ECDH on ed25519 directly

cool

yeah Ill expose a public symbol for these

I don't know if aynone noticed, but this can be built and installed as a standalone static/dynamic system lib

I agree with leaving it constant time, though would be interested to see the perf difference if you ever care to check

yes, it might be worth exposing this for advanced used if there is a decent bumb

bump

Looking for opinions on this: monero-project/monero #6600

Would it be worthwhile to add a "sign with both keys" option, as opposed to separate sigs with each key on the same message?

The idea is that asking for a signature with a spend key does not guarantee that the signer has spend authority... you could have the spend private key but not the view private key