On Fri, Nov 09, 2018 at 12:17:50PM +0100, Werner Koch wrote:
On Fri, 9 Nov 2018 11:50,
ijackson(_at_)chiark(_dot_)greenend(_dot_)org(_dot_)uk said:
It certainly would allow serving the data from static files. If you
wanted case-insensitivity and can't configure your webserver to smash
It is not only about the case but about allowed characters in a file
name. In particular '/' and depending on file system the length. Noet
that '/' is a valid character in the local part of the addrspec.
Since you are still in the protocol design phase, you would no doubt
welcome implementation and deployment of an alternative simpler
Nope. It is in use for more than 2 years.
I feel some obligation to push back on this -- if there is no willingness
to deviate from the deployed implementation, why not just document the
existing implementation behavior as part of the implementation's
documentation and move on?
-Ben
The only simpler thing which could have been done would be to skip the
hashing and directly use the z-base-32 encoding. The only drawback
would have been that very long addresses won't work on all file systems.
Changes to the SVR record thing should be possible because I doubt that
this is widely used (Caesonia specifies it use, though).
_______________________________________________
openpgp mailing list
openpgp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/openpgp