✧✦Catherine✦✧@whitequark@social.treehouse.systems: did a crackhead design this

Home | Notifications | New Note | Local | Federated | Search | Logout

Note Detail

Reply to @whitequark@social.treehouse.systems
✧✦Catherine✦✧@whitequark@social.treehouse.systems (2026-09-15 00:42:42)
why the hell do the modulus and exponent of the RSA public and RSA private keys have different order in the serialization stream
---Reply---
✧✦Catherine✦✧@whitequark@social.treehouse.systems (2026-09-15 00:43:19)
did a crackhead design this
---Attachments---
image: https://cache.treehouse.systems/media_attachments/files/117/270/157/366/179/093/original/2084db81643249de.png


---Replies---

grawity@grawity@social.treehouse.systems (2026-09-15 00:50:08)
@whitequark for what it's worth, the private-part format is AFAIK taken directly from the "ssh-agent protocol" format that's now documented in RFC 9987

(just because that frankly seems a bit more readable than whatever the heck sshref's syntax is. I've never heard of "uint32 allocator" terminology before o7)

in other words, I *suspect* they reused the existing (de)serializer code, which evolved independently for the public half in ssh proper and the private half in ssh-agent, leading to a somewhat funny combined result