reply on: The Curious Case of Digital Signatures \ stacker news ~crypto

pull down to refresh

0 sats \ 40 replies \ @ek OP 24 Feb 2024 freebie \ parent \ on: The Curious Case of Digital Signatures crypto

found the public key from https://github.com/spesmilo/electrum/blob/master/pubkeys/ThomasV.asc

Imported the key, also there are other devs are listed on the site.

Good, I think using Github as the source of trust is okay. But remember: the more sources that say that this is indeed the correct key, the better!

one thing really strange is that when I tried this again, it says
gpg: can't open 'electrum-4.5.3.dmg.asc': No such file or directory

gpg: verify signatures failed: No such file or directory

Mhh, and you are sure you didn't (remove) the file? Did you run gpg --verify in the correct folder?

I can't find the SHA256 to continue 😳

If the software you downloaded was signed, then you don't need separate hashes. The signature contains the hash to verify integrity. I can tell from your comment that this is the case for Electrum since the signature is named electrum-4.5.3.dmg.asc and the software is in electrum-4.5.3.dmg.

Sparrow Wallet was just a special case where not the software was signed but the hashes. Then you need to run another command (sha256sum --check <hashfile> --ignore-missing) to verify the software.

I mentioned that I don't know why Craig did it like this, I only had an educated guess:

Conclusion

So what we just did was to basically verify the authenticity and integrity of the file that contained the hashes for all binaries with gpg --verify. When the hashes could be trusted, we could use them to make sure that the software was not tampered with. But why not simply provide a digital signature for the binary itself?

I actually don't know. But my educated guess is that it's related to convenience. Instead of providing a signature for every binary, the hashes are signed. Using sha256sum --check with --ignore-missing then simply ignores all files that don't exist. So I am basically guessing that there is no way to do something similar with digital signatures. Maybe someone knows more?

473 sats \ 38 replies \ @Natalia 24 Feb 2024

deleted by author

480 sats \ 37 replies \ @ek OP 24 Feb 2024 freebie

Yes. The "asc" is the (detached) signature.

The hardest part is verifying the public key but most people just skip that lol

0 sats \ 6 replies \ @Natalia 25 Feb 2024

now I finally understand what you mean here, why not just put each dev's key in GitHub 😨

0 sats \ 5 replies \ @ek OP 25 Feb 2024 freebie

Best way is to spread your key fingerprint around imo.

If you only use one site as the source of trust, it's a single point of failure. Even if it's Github.

I have to do that myself, still figuring things out around PGP keys

412 sats \ 29 replies \ @Natalia 24 Feb 2024 freebie

deleted by author

346 sats \ 28 replies \ @ek OP 24 Feb 2024 freebie

To be fair, I think if the instructions mention to import the key from a site like Keybase like Sparrow does, I think it's fine. Most important thing is to not import the public key from the same site you received everything else and I think if people just follow instructions, they automatically do that.

It just makes me feel uneasy if people are not aware that this is important. The why's and so on.

589 sats \ 27 replies \ @Natalia 24 Feb 2024

deleted by author

0 sats \ 26 replies \ @ek OP 24 Feb 2024

Haha yes. Like a secret key hidden in plain sight.

0 sats \ 25 replies \ @Natalia 24 Feb 2024

deleted by author

359 sats \ 24 replies \ @ek OP 24 Feb 2024

~~Yes~~

~~You just summarized my post with a few words haha~~

Wait, no. The dev signs the software (or whatever). The signature IS the hash "encrypted" with the private key.

view replies

0 sats \ 0 replies \ @Natalia 27 Feb 2024

deleted by author