On Mon, 28 Oct 2019 12:28, Daniel Kahn Gillmor said: > one of them. But while GnuPG's errors are implementation errors, SKS is
There was actually only one error: Two limits on the size of a keyblock which did not match so that only at a later point in processing the too long key size was detected. The second limit was actually failsafe code. The problem with all that is that _users demanded_ the ability to create large keyblocks, because they went to so many key signing parties and liked to include arbitrary data. Thus a small and reasonable limit would have annoyed a just few but nevertheless vociferous folks. Just look at the silliness of the large RSA key ability in GnuPG. It costed us several weeks to come to some kind of compromise - because of one key. Look at the removal of v3 keys - technically the correct solution but still a lot of complaining and whining. We have always also designed for minor use cases but the costs for this is very high. The slowness during import is due to the oh-so-cool cleaning of misplaced signatures in keyblocks: "Look, it just does not take a lot of time; let's enable this by default". > call "web of trust". I'm not convinced that it has ever worked reliably > for most people or most use cases, unfortunately. i say that as a It will never work as we expected it 25 years ago because it is too complicated and I bet the majority of participants don't diligently check the identity. If it makes sense at all to have such passport approved identities. > Debian developer, one of the groups that is the most heavily invested in > the idea. I know Arch cares about it too. There are actually even commercial groups who rely on it. > I'd love for it to be more reliable than it has been in the past, and > getting that to happen requires fixing a lot of different pieces. > Please, help us get them fixed! Agreed. But it is and will be a minor use case. > There's a pretty broad consensus that tools should validate input, and > that GnuPG needs to take more serious steps than it has in the past to > defend against potentially malicious inputs. You can't. Neither can a webbroweser avoid downloading huge images or other stuff from stupid web pages. Salam-Shalom, Werner -- Die Gedanken sind frei. Ausnahmen regelt ein Bundesgesetz.
signature.asc
Description: PGP signature
_______________________________________________ enigmail-users mailing list enigmail-users@enigmail.net To unsubscribe or make changes to your subscription click here: https://admin.hostpoint.ch/mailman/listinfo/enigmail-users_enigmail.net