New user, thank you for your great work! I ran into some issues, but I might just be using the app in an unintended way.
On 1.5.1 testflight on latest iOS, connected to an Arch Linux server running latest Codex via SSH over Tailscale:
-
I have an ongoing thread on the server started with codex --yolo via ssh and a PC terminal. I can see this thread in Litter and can open it. The moment I open it, I get permission requests popups, despite --yolo. This seemingly keeps happening despite toggling permissions (red lock thing after clicking on the top header of the thread). Shouldn't Litter respect the --yolo flag from CLI? If not, is there a permanent setting where I can set full permissions permanently for a session/server/overall?
-
I kept a ssh connection open on the PC, with the codex thread visible and active. If I connect onto this same thread with Litter, and write a message in Litter, this is not reflected in the ssh thread on PC. It seems that Litter creates a background session which diverges from the PC session and both can keep working in parallel? Is this expected behavior? Is Litter actually taking a session resume log and spawning a new session in the background, or is it actually connecting to an ongoing running codex thread? If it's the former, what's the intended workflow for someone who primarily works from a PC and wants to only occasionally check in on running threads and send off a few commands, and continue working later from a PC again? Constantly starting new sessions via /resume?
-
Once the Litter thread started going in unintended directions, clearly diverging from the PC SSH thread, I saw no obvious way to terminate it, similarly to how closing a terminal window with a active ssh connection terminates whatever codex is doing on PC. Closing the app seemingly did nothing, since it just went into the background. What's the proper process for stopping Codex in Litter?
Thanks!
New user, thank you for your great work! I ran into some issues, but I might just be using the app in an unintended way.
On 1.5.1 testflight on latest iOS, connected to an Arch Linux server running latest Codex via SSH over Tailscale:
I have an ongoing thread on the server started with codex --yolo via ssh and a PC terminal. I can see this thread in Litter and can open it. The moment I open it, I get permission requests popups, despite --yolo. This seemingly keeps happening despite toggling permissions (red lock thing after clicking on the top header of the thread). Shouldn't Litter respect the --yolo flag from CLI? If not, is there a permanent setting where I can set full permissions permanently for a session/server/overall?
I kept a ssh connection open on the PC, with the codex thread visible and active. If I connect onto this same thread with Litter, and write a message in Litter, this is not reflected in the ssh thread on PC. It seems that Litter creates a background session which diverges from the PC session and both can keep working in parallel? Is this expected behavior? Is Litter actually taking a session resume log and spawning a new session in the background, or is it actually connecting to an ongoing running codex thread? If it's the former, what's the intended workflow for someone who primarily works from a PC and wants to only occasionally check in on running threads and send off a few commands, and continue working later from a PC again? Constantly starting new sessions via /resume?
Once the Litter thread started going in unintended directions, clearly diverging from the PC SSH thread, I saw no obvious way to terminate it, similarly to how closing a terminal window with a active ssh connection terminates whatever codex is doing on PC. Closing the app seemingly did nothing, since it just went into the background. What's the proper process for stopping Codex in Litter?
Thanks!