Repository navigation
P2P Backplane: create, drive, stream and view threads across all machines and clients - #53
Merged
Merged
Conversation
…ar/rpc), hub stage and owner handler
…ect add/browse to the owning machine; picker machine row; making state until the mirror lands
…, server queue/receive
…folder, reply fix) with laws
… note channel for far turn ends
…list, replies fixed
…ds elsewhere; phone View.fv
…w), they cache and show
…nch, farlive tests, live e2e
…, sync-off machines not chosen or routed, forks shared, bring back routed, picker resets machine listings, projection version in the epoch
…ly to owner machines
… times, one locked install that never rolls a file back, deletions follow the owner, a failed sync is not acknowledged and answers down, phones schedule one ask, nothing opens before the cache folder is known
… and go only to machines with sync on; one prefilter for far requests
…llows a remote hub's files into a local cache
…k, a far request answers only once stored, /give elsewhere is a queued handoff (thread_handoff) reported once taken and never also read as a local word, thread_read asks the owner, thread_wait by the clock, machine_list from project changes only
…clears only an older stream, a late delta never shows, a newer one that came first stays), bounded snapshots, truncated buffers go whole, far turn ends alert without touching another hub's Live Activity, a tapped far alert opens on the hub that holds it, island keyed by the owner's id, previews read only when sent
…view's root), checked against this hub's farfs
…nd retried sends stop when sync goes off, a too-big live request is not taken for an idle peer, a history reset drops its live text, models are never deleted by another client's list, snapshots 16K, multiline /give, phones open a far file only once its cache is known
…s own repository and never a hidden file, cache rounds ordered by a ticket taken as their list arrives (a restored older file is installed, a stale round installs nothing), a failed round waits before the next, copy paths never prefixed twice (child sheets), a comparison's working copy read from the copy; bots_test schema text
Snaiderbg
pushed a commit
to Snaiderbg/Backplane
that referenced
this pull request
Oct 9, 2026
Bots answer people on other machines
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
P2P Backplane: every one of the owner's machines and every client of any hub act as one system.
machine,<link>~<id>,>machine:slug). /give to>machine:slugand%machine:projectqueues a handoff the same way a local /give does, and says which machines took it. The prompt gains a<machines>line.Reviewed adversarially by GPT-6.1-Sol in three rounds; the findings were fixed or are listed below.
Tests: scripts/check.sh (PROOF: All terms check), scripts/test.sh, and new e2e tests with two real hubs each: create_far, agents_far, live_far, viewers_far, remote_view, remote_view_far, plus far_e2e.
Known limits: a thread that lives elsewhere is given from its own machine; history versions of a mirrored thread aren't read from another machine; sender ids through a third machine aren't translated; catching up a very large mirror re-scans it once per chunk (this predates the PR).