Re:Guile security vulnerability with listening on localhost ...
by paroneayea on Friday October 14, 2016 @10:30AM
(#12708293)
It does look like there's a way to "forward" a unix domain socket over ssh enough to do that kind of debugging:
Re:Guile security vulnerability with listening on localhost ...
by qwertyuiop924 on Friday October 14, 2016 @11:48AM
(#12708871)
That's certainly a potential solution.
Re:Guile security vulnerability with listening on localhost ...
by tetrep on Friday October 14, 2016 @11:04AM
(#12708517)
Why not have the REPL (or whatever else you're exposing to localhost) require a token for authentication of each request?
Re:Guile security vulnerability with listening on localhost ...
by qwertyuiop924 on Friday October 14, 2016 @11:31AM
(#12708735)
Because it's usually a raw socket, not an HTTP interface, so that's kind of tricky. But it could work.
Re:Guile security vulnerability with listening on localhost ...
by fulafel on Friday October 14, 2016 @10:49AM
(#12708421)
Networked REPL doesn't preclude security. See eg Twisted's SSH based twisted.conch.manhole
Re:Guile security vulnerability with listening on localhost ...
by qwertyuiop924 on Friday October 14, 2016 @02:06PM
(#12709957)
Yeah, but until you linked that, I'd never seen a networked REPL that actually provided it.
I'm glad that Guile kept the option to run a REPL via a port, because that's really useful for debugging nonlocal servers (investigating production issues, etc.).