summaryrefslogtreecommitdiff
path: root/src
Commit message (Collapse)AuthorAge
* terminal: an authority with screen-state replicas sizes a terminal after the ↵Gabriel Schneider2026-10-30
| | | | | | step's unit, not inside it (resize_to, made by Peers.unit, null from `create` on): a client typing at another size than the last typist read history rows its replay held as holes, and the rows the session sent for them came in the new grid's numbering, which the replica cannot take, so its replay read blank text (the soak's local80 cell: jumps, njumps, jcur, pane 1.body at a b; fixer 2's H18, pane N.ovl); test/replicas.zig sizes them after its fill too; an in-process test reads the history in the step that resizes, with the session's order of flush, fill and unit, and another makes a terminal from 0xAA memory and finds no size waiting Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Places: its regex made again from its pattern when a set holds none: `re` is ↵Gabriel Schneider2026-10-30
| | | | | | derived (in no dump and no rollback copy), so a client that joined took a loaded search for a list's and found nothing an edit made, and diverged at the next edit, again at every rejoin (the soak's local207 cell: pane 3.places.hits, and pane 17.cwd, which inherits pane 2's and hashes it whole); a test loads a session's dump and makes the same edit in both Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Stream: a search run on the loop (a 9P write's Grep or Find, which answers ↵Gabriel Schneider2026-10-30
| | | | | | whether it found anything) is a tape site, `search`, its batches what the replay takes, as it takes a worker's (Event.search): its file reads were no site, so a replica of a session a script greps in read no file and joined again (`a syscall that is no tape site, made replaying`, a terminal client live, 2 of 2 before, none after); test/cross_replay.py's session greps and finds again Co-Authored-By: Claude Opus 5.5 <[email protected]>
* terminal: a step that reads a terminal's rows as the core keeps them marks ↵Gabriel Schneider2026-10-30
| | | | | | | | | | it read too: `fill` sent a client that came since (its history all holes) none of the rows the step read shellRows set `read` only when it built the rows, not when it handed out the ones kept since the last write, so after a client joined (or was started afresh) a step reading through the cache left the client's holes unsent, and it replayed the step on blanks. Found on the way to H18, whose fix is fixer 1's (wwkwwlzt). Test (failing before): a step that reads a terminal's rows as the core keeps them sends a new client the ones it holds as holes. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* walk: an enum is digested by its name (its word, a fixed-width hash), not ↵Gabriel Schneider2026-10-30
| | | | | | its number: registry.Builtin, made at compile time, has Tty9p in a Linux build and not a FreeBSD one, so a +LocationsConfig (its output's origin a Builtin) digested otherwise in each, and a Linux client of a FreeBSD session holding one could not join (the hunt's H19: `the session's dump loaded as another session`, main's window client of main's session on the box; after, a terminal and a window client here join one with a +Search, a directory, a renamed file and a +LocationsConfig, and agree); walk's test, and test/cross_replay.py's session holds a +LocationsConfig and a renamed file (a FreeBSD tape replayed here diverged at pane 2.file.output before) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* web: the build links no libc, so what is the C heap's is a hosted build's ↵Gabriel Schneider2026-10-30
| | | | | | alone: terminal.create takes emulator_gpa (std.heap.c_allocator) only in a hosted build, as only one has a screen-state replica, and walk's blob lists (a replica client's, a join's) only with libc; `zig build web` failed on main at terminal.zig:601, and at walk.zig:115 behind it; it builds, and test/web-snapshots/touch.mjs passes (the driver's own CDP-deadline test fails as on main, untouched) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Adopt and a terminal reader's failure are inputs (Pardes.adoptShell, ↵Gabriel Schneider2026-10-30
| | | | | | readerFailed; tape.Input adopt_shell, reader_failed): tty.zig's and gui.zig's adoptShell ran the Tty word and cleared the pane's greeting straight on the core, and the window's reader failure set its pane's mode, between steps and untaped, so every client of a session that took in a dead session's shells joined again (the hunt's H15: `the unit of 0/48 diverged: the core called poll_frame where the tape has a host (10)`, gone after); replica.zig's test, which the old writes fail Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Messages: a detached session's messages go on the clock in the pump, after ↵Gabriel Schneider2026-10-30
| | | | | | its wait, not in the host's wait: the host's call (tty waitInput, gui waitDetached) was no input, its clock read untaped, so a replay (and every replica of a session nobody attached) kept a message its session had let go; with MessageLinger 0 (test/daemon.py's sessions) a tape whose session said a failure never replayed, on main too (pane 0's msg, digest 6, 7 of 7 runs; 0 of 6 after); test/cross_replay.py's session says MessageLinger 0 again Co-Authored-By: Claude Opus 5.5 <[email protected]>
* remote clients: a pointer's hit carries the place it was made on ↵Gabriel Schneider2026-10-30
| | | | | | (Layer.Place, on the wire, version 17) and a step resolves it from that, never from this process's frame (`surface`, `.local`), which a session never draws for a client: a window's pointer over a tag diverged it at once (the user's `tape byte 119`, 15 rejoins in 25 s) and a click there aborted it; a replay that meets a site its tape lacks, or makes a syscall that is none, is a divergence that goes on harmless (poisonOf), not a panic, and a speculation that makes one waits (Tty's shell lookup, untaped, is a value site now, as Shell's was); a copy over a queue is refused, not asserted; a session's last client, while no one else watches it, leaves it its layout (leaveView, settled as closeView is), which its rejoin copies; a redial waits only after an Upgrade's `refused` (a divergence's redialled a gone session 20 s); a dying session shuts its clients' links before its trace, and an --attach client ends its ssh and socket whatever way it exits; a far join's hashing stops at a blob once its client is let go (its leave waits for it in the step); a confirmed speculation's state is checked against its unit's digest, and the trace gives a key's and a pointer's words; test/daemon.py: 200 far joins level, a window's pointer swept over tags and body, ssh stand-in clients refused and detached, a client of a killed session ending in 2 s; in-process tests for each Co-Authored-By: Claude Opus 5.5 <[email protected]>
* screen_state: a terminal's absolute rows numbered so that its history's ↵Gabriel Schneider2026-10-30
| | | | | | oldest is row 0 or later: a reflow to fewer columns puts more rows above the last active top than ever scrolled past it, and a peer started afresh at `top - history` overflowed (the session aborted as a window client's font resized a shell); a test narrows 200 long lines Co-Authored-By: Claude Opus 5.5 <[email protected]>
* c_heap: a heap's allocator read as its bytes are, through a volatile ↵Gabriel Schneider2026-10-30
| | | | | | pointer: std's stateless allocators (c_, page_, smp_, wasm_allocator) leave `ptr` undefined, and one never written outside a test (replica.zig's zstd heap over c_allocator, web.zig's over wasm_allocator) is a constant to a release build, which folded `known.ptr == current.ptr` to false: every malloc named a new allocator and the 65th panicked (the box's session aborted at about its 16th far join; the core's table held 64 copies of {ptr 0, c_allocator's vtable}); its test runs built ReleaseSafe in unit-test too, where it panicked before Co-Authored-By: Claude Opus 5.5 <[email protected]>
* views: a view drawn or read outside a step (a replica client's own while ↵Gabriel Schneider2026-10-30
| | | | | | | | | | another is live) leaves the live view's slot and its panes' shares holding nothing, as unpark and drawOthers leave them withParked parked the live view and copied it back, but left its slot (column tags' texts, a chord's argument, the presentation) and each pane's share (its selections) pointing at what the live fields own. A rollback copy walks every slot (src/rollback.zig): once the live view freed one of them and the heap gave the memory to another, the copy saved that and put its old bytes back over it, which `emptied` exists to prevent (the slot left as H5's hypothesis had it, in a path it did not name: Replica.draw and Replica.within). Test (failing before): a view drawn outside a step leaves the live view's slot holding nothing a rollback could put old bytes back through. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Places: the lowered copy a smart-case search reads is zeroed where it is ↵Gabriel Schneider2026-10-30
| | | | | | | | | | made: its pieces not lowered yet were whatever the heap held, in every digest and a join's dump, so two processes' places differed (H5's and H17's `places.hits|hay`) freshHay took `memory.unfilled` and lowered a piece at a time as the search reached it; the digest and the dump take the whole slice. Zeroing it costs 20 us for 1 MB, 0.4 ms for 10 MB and 10 ms for 100 MB in one frame (ReleaseFast, page faults included; lowering the same text is four times that, spread over frames), the order of the copy an edit of that text already makes. Test (failing before): two sessions searching one text, each heap handing out other bytes (rollback.Epoch.fill), digest the same places. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* views: a view opened or closed settles in its own input: a unit taped ↵Gabriel Schneider2026-10-30
| | | | | | | | | | outside a step (a leave's close_view, a join's open_view) ended with effects queued, which a replica's copy asserted on and a joiner's dump dropped A client's leave is the session's closeView, whose sync sizes again a terminal the closed view sized (`sizes`) and queues its pty's resize; the session did it at its next pump, another unit. A replica that replayed the close was left with the effect queued between steps, and its next idle copy (Replica.ready, rollback.Copy.take's assert) aborted it: a client killed when another left (H10). A client joining at that boundary was dumped with the effect still queued, which a core loaded from a dump has none of (Replica.initDump), and diverged at the session's next pump (H6: "the core called poll_frame where the tape has a host (12)"). closeView and openView settle now, inside their own taped input, so every side does the same there; openView's covers what else was queued since the last pump (a 9P request's step). Test (failing before at each of the two asserts): a client leaves, and one joins: what the leave queued is done in its unit, on every side, before the join's dump. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* replica: what only the authority holds, a step reads off the tape ↵Gabriel Schneider2026-10-30
| | | | | | | | | | | | | | | | | (tape.authority): the store's closed panes and Recent, a script's hold on a pane's events, the 9P socket, the session's config directory and startup file; a replica's own are never read, and never put back by a rollback The Panes picker's closed rows, Recent's rows, a closed pane brought back (a Look at @pN, a Panes row), a jump to a closed file and the +Jumps rows read `p.store` and `p.recent`, `.authority` state a replica keeps on its own: its store has no file, its Recent only what it replayed since it joined. The rows made of them went into digested state, so every client of a session diverged together after a Del (H4), every client that joined after one diverged at its first unit forever (H11: trackJump dropped a closed file's jumps at every step when Recent did not know the file), a client's Del of a file the session's Recent knew made the stat the session skipped (H12: noteClosed looked in Recent first), and a client on another machine read its own store (H13). - tape.authority: poison, then a value read: the authority's answer is taped, a replay takes it, a speculation waits for it. One site, `.authority`. - Readers through it: Query's Recent and closed-pane rows, Store.reopen's record, Pardes.recentByHash (jumps, +Jumps), mouse.reportRange's readers check (a script reading a pane's events takes a look or exec: `pane.fs.readers`, the 9P server's), Tty9p's socket check, Pardes.configDir and startupPath (ThemeFile, DumpThemes, Config, DumpConfig, an init file saved), DumpConfig's dump directory. - trackJump no longer reads Recent at every step: a closed file's jump stays; the jump to it drops that file's entries when Recent has forgotten it, as when the file is gone. - noteClosed asks the host (recentKeeps, a taped stat) before it looks in Recent: the same syscall on every side. - The tty toggle, read on every key, is session state: Pardes.tty_toggle, from the options when the core is made, dumped. - The rollback copy no longer saves `store` and `recent` (rollback.zig `unsaved`): their bytes put back pointed at lists the host's heap (recentGpa) had freed, which the epoch never holds. A speculated Del then a Panes listing read a freed name (H8, the segfault). Tests, each failing before: a rollback leaves the store and Recent as a speculation left them; a replica joined after a pane closed has the session's jump to it, its Panes and Recent rows, asked by either (and the variant hunt's Recent written by another window); a replica's Del of a file its session's Recent knows asks the host what the session asked. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* pardes: a core's first frontend is an option (Options.windowed, taped with ↵Gabriel Schneider2026-10-30
| | | | | | the rest), not its build's default, and every hosted build takes a window's font words (observe_font, acknowledge_font, reject_font, take_font_request): a window's build replaying a terminal's session started as a window (its init file's window settings applied), and a terminal's replica of a window's session diverged at the window's first font Co-Authored-By: Claude Opus 5.5 <[email protected]>
* regexp: one step budget in every build mode, 32M: what a search finds is the ↵Gabriel Schneider2026-10-30
| | | | | | core's state, and with 4M in Debug a Debug replica of a ReleaseSafe session answered TooSlow where its session found no match (`\s*(\w+)\s*=` over 200 letters, the variant hunt's test, here) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* tape: an error crosses by its name, a stat as fs.OsStat, and the tape's ↵Gabriel Schneider2026-10-30
| | | | | | shape is in wire.semantics: a replica of another build read its session's units as its own build's, so a window client of a terminal's session, or of the box's, read every error by a number its build gives another (451 of 505 names numbered otherwise in the Linux terminal's and window's builds, 244 between Linux and FreeBSD: pardes-gui replaying a terminal's tape panicked at `unexpected error code, found error.NonLocalPath`, and so did main's window client of a session on the box at its Look of a fifo) and a FreeBSD stat out of step (nlink 8 bytes there, 4 here); a name this build's set lacks is a divergence (error.NoSuchError), never another error; an anyerror (a host's report) reads back as one of tape.Failure's, a Debug build panicking at one outside it as it is taped; tape.shape (each taped type's fields, tags, enum names and numbers) in semantics refuses, in words, builds whose tapes differ (the stat taped as before: the Linux and FreeBSD builds hash to two), and a site whose result tape.results does not list is a compile error; walk's test of an error by its name Co-Authored-By: Claude Opus 5.5 <[email protected]>
* replica: the two tests that count a replica's fast confirmations start early ↵Gabriel Schneider2026-10-30
| | | | | | in a second: an edit's file time is the clock's second, which the replica guesses, so a second's turn between a key and the authority's take of it was a misprediction (another client's input first: expected 4 fast, found 3, in 11 runs of 40 of a loaded Debug window build) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* remote client: a write to the session that fails is its hangup (a client cut ↵Gabriel Schneider2026-10-30
| | | | | | off as it lags said nothing and exited 0 when its own write found the link shut before its reader did); a window client's frame is marked rendered as it presents, so the live harness reads its render as its core's time; test/daemon.py's window clients: one far (PARDES_REPLICA_FAR) put off the session's state rejoins from the core it left, and again across an Upgrade (its `refused`); a client's echo stats in PARDES_REPLICA_STATS; --ps beside them lists the session alone, a window's as (gui, attached), its clients none Co-Authored-By: Claude Opus 5.5 <[email protected]>
* pty: what a run's command printed taped (tape .prompt), as a step's read of ↵Gabriel Schneider2026-10-30
| | | | | | its terminal after output is: a screen-state replica's emulator is its deltas', which a lagging client's are states behind, and a run's end replayed there read pins into pages it never had (ghostty's leftWrap overflowed: a window client stopped while a shell flooded the session aborted 3 runs in 8; none in 6 since) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* gui: the window a replica client (pardes-gui --attach=<socket>:<name> and ↵Gabriel Schneider2026-10-30
| | | | | | <host>:<name>, as the tty's, far through ssh as amort's: its blobs by hash and its stream in zstd) and a window session serving replicas (a terminal's client or a window's): the client feeds its replica what its window does, read over its own view (views.within) and applied after, draws its view between steps, takes the stream on its own reader and rejoins as the tty's does (replica.Link, the client's half shared with it, the far path's Have with it); the session takes clients at its step boundary, one person's input a wait, a client's view no window of its own, its font and opacity answers taped; a client says its frontend first (Event.frontend, in wire 16 beside the far word) and the tty and window builds step alike (one desktop semantics: each has the other's words, which a view takes by its own frontend, builtins.registry.here); a check asked 300 ms after the last input and check, not per wake; test/daemon.py --window-clients Co-Authored-By: Claude Opus 5.5 <[email protected]>
* daemon: pardes --ps, --kill=<name>, --kill-all: the daemon's registry as a ↵Gabriel Schneider2026-10-30
| | | | | | tree (the daemon, each session with its pid, kind, state and name and the shells it holds, then the shells no session holds, which Adopt takes) and sessions ended with their shells: a TERM, a KILL 2 s later, gone once the daemon no longer lists them (a zombie still answers kill 0), a killed session's held masters closed rather than kept for an Adopt and --kill-all's every kept one; in --help, the guide and the reference; --help's ssh default no longer -C; test/daemon.py --manage Co-Authored-By: Claude Opus 5.5 <[email protected]>
* 9p_io: the parked-write test rests again until the client's release has ↵Gabriel Schneider2026-10-30
| | | | | | landed before it reads the body: a write through an open is held (pane.zig batchFor) until the open is let go, and the client lets it go without waiting, so under load the test read the body truncated and the bytes not yet in (it failed on the first run of a loop beside builds, then 30 of 30) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* replica: a far client's content cache and zstd: the session lists its join's ↵Gabriel Schneider2026-10-30
| | | | | | blobs by BLAKE3 hash first (wire 16 `blobs`), the client finds each in the core it leaves or on disk ($XDG_CACHE_HOME/pardes/blobs, hashed again on load, a bad file a miss, the least recently used evicted past 1 GiB) and asks for the rest (`want`), and everything after the welcome goes in zstd envelopes (level 6 over 8 MiB for the join, an undo snapshot being a near copy of its text, then level 3), so `ssh -C` leaves the default command; a far client (through ssh, or PARDES_REPLICA_FAR) keeps what it was sent and, at its end, what it made; a hangup ends a remote client the normal way; through the box a client holding the blobs is drawn in 2.4 s where main took 7.0 to 11.9, and a first join of a session with undo in 7.1 s where main took 11.9; libzstd 1.5.7 vendored, its allocations on c_heap, a second copy refused by shared-libs-check; tests for the cache's verify-on-load, its eviction and a far join's blobs sent once; test/daemon.py's far clients Co-Authored-By: Claude Opus 5.5 <[email protected]>
* remote client: a link's end read before the drain, not after it: the link ↵Gabriel Schneider2026-10-30
| | | | | | reader posts `link_gone` just before it marks its end, so a drain that tested the mark last could leave that post queued, and the next link (a rejoin after divergence or an Upgrade) took it and ended at once, the client saying 'let this client go' right after 'joined again' (test/daemon.py's forced divergence, about one run in ten under load) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* walk: an optional slice or pointer loads through a local, not into an ↵Gabriel Schneider2026-10-30
| | | | | | optional set to an undefined payload: a slice's null is its address, so that optional was undefined and ReleaseSafe's optimizer took it as null; a command pane's `?[]u8` command read one byte long, every field after it out of step, and Restore of a session with a command pane in it said "end of dump" (Debug's 0xAA filler and ReleaseFast read it right); a test of an optional slice and the field after it aborts in ReleaseSafe without it Co-Authored-By: Claude Opus 5.5 <[email protected]>
* main: tcgetattr and tcsetattr take libc's place on Linux glibc, the kernel's ↵Gabriel Schneider2026-10-30
| | | | | | own ioctls on the 36 bytes both layouts begin with: glibc's struct termios is 60 bytes and Zig 0.17's std.c.termios the kernel's 44, so libc's tcgetattr wrote 16 bytes past every Zig caller's (vaxis's makeRaw, the stub's, the GUI's) and a ReleaseSafe build trapped at its first terminal (Debug and ReleaseFast ran on by luck of the stack); a test with a canary past the struct fails without them Co-Authored-By: Claude Opus 5.5 <[email protected]>
* terminal: a shell's cells painted from the row's own pin (cellAt), not ↵Gabriel Schneider2026-10-30
| | | | | | getCell's walk from the viewport's top to the row for every cell: a frame of a 120x40 terminal over 64 KB to 8 MB of scrollback 107-127 -> 94-102 us against base, ReleaseFast; the 'regression' phase 5B read against base was a Debug-core perf binary (core_optimize debug, terminal safe) beside a ReleaseFast one, file gestures level once both are ReleaseFast Co-Authored-By: Claude Opus 5.5 <[email protected]>
* `pardes --attach=<host>:<name>` runs `ssh -C` when $PARDES_SSH is not set: ↵Gabriel Schneider2026-10-30
| | | | | | through the FreeBSD box a 3.9 MB join took 7.8 to 7.9 s plain and 5.4 to 5.8 s compressed (ssh brings the forward up in 2.0 s of either) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* tty: a terminal attached while a replica client's view is live takes the ↵Gabriel Schneider2026-10-30
| | | | | | | | session's own screen: the live view has no terminal here, and the attach returned without a word, its stub waiting and its terminal blank (the main test's third terminal, on the box) test/daemon.py: the clients test attaches one while a client's view is live. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* a replica client's join costs the session's step its mutable state alone: ↵Gabriel Schneider2026-10-30
| | | | | | | | | | the dump leaves the blobs out (walk.zig `Dump.deferred`: a `.blob`'s texts of 4 KiB and up, noted where they go) and its terminals' emulators (a size: the first delta brings each), the client's writer thread writes the welcome and the dump with the blobs put back in place, and each blob is pinned until written (memory.zig `pin`: its free put off); the texts the session never digested are hashed first on a thread of their own (walk.zig `unhashed`, `makeWords`), the join waiting a boundary for them The client is unchanged: what it reads is the dump `dumpBytes` makes, byte for byte (a test splices one and compares). ReleaseFast, the session's worst boundary while three clients join one after another (`t2/join2.py`), before and after: typical (4.1 MB of files, 1 MB of scrollback, an 8.1 MB dump) 12.4 to 1.6 ms; big (38.6 MB, 8 MB, 76 MB) 51.3 to 1.5 ms; a 100 MB file open beside typical's (254 MB) 172.5 to 1.6 ms. Of the dump the step makes 113 to 122 KB. PARDES_REPLICA_STATS says each join's time and the state's size. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* remote client: predicted echo (src/echo.zig) over its own view, judged a ↵Gabriel Schneider2026-10-30
| | | | | | | | | frame at a time and woken while predictions are out; PARDES_NO_ECHO turns it off; a session's step that reads whether a terminal is at its prompt after its output (the first prompt's greeting, a command waiting for the prompt) is taped (`.prompt`), as a screen replica's emulator is the last delta's until the next Found by a client opening a shell of its own (` nt`): the session's layout after the shell's first output saw the prompt and typed its greeting (`ls`, after a pty resize), and the client's replay of it, its emulator a delta behind, did not: it diverged at once. At 207 ms through the relay (Debug): a key into a shell on the client's screen p50 9.1 ms with prediction, 213.6 ms without. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* replica clients join again: one that diverged (a digest the session's unit ↵Gabriel Schneider2026-10-30
| | | | | | | | | | | | did not end on) and one the session let go to upgrade dial it again, the same join as the first, and say why in a line over their frame until a key; a client of another build is told "session upgraded to a newer pardes: upgrade this client too" tty.zig `remote` is a loop of joins over one terminal: the link ended, its reader joined, the replica let go, the session dialled again (an upgrade's exec and load, and its daemon's, retried for 20 s), its dump loaded, its view's size sent again. `Remote.notify` draws the line (`noticeLine`, the client's own, never the session's core). The welcome carries the session's semantics: the daemon judged the hello by the hash it last heard, which an upgrade under way may not have said yet. replica.zig `Peers.refuseAll`: before an Upgrade's dump each client is sent `refused` after what it was sent (up to 200 ms) and let go, its view with it. "A newer pardes is installed" is a Host call (`newer_build`) the core makes in its focus step, so replica clients take the line from the tape; the GUI still says it itself. PARDES_TEST_DIVERGE=<n> puts a client's core off the session's once quiet after its n-th unit. test/daemon.py: the clients test has one rejoin and agree; the upgrade test has a client of the old build told to upgrade and one of the build running join again and agree. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* replica clients' terminals as screen state: the session's Peers keep a ↵Gabriel Schneider2026-10-30
| | | | | | | | | | Screens (rterm's), a peer each: the rows a pump's steps read before its unit, the deltas after it, and before each client input's step, the rows it read before its unit; a joiner's terminals afresh after its dump; acks, rows asked and resets from the client (0x04) taken at the boundary; units carry no pty bytes and a word of digest The client applies a delta or rows (0x06) in stream order and asks for the rows it shows and lacks after each frame; its dump's emulators are loaded on the C heap (walk.zig `.alloc` of a pane's `terminal`, dump.zig `terminal_gpa`), as `terminal.create` makes a screen replica's. Its load checked under its own tape (no emulator in the digest). tty.zig: with replica clients, a wait takes one step of a person's input and none after a pty's output (`split`): the next is held for the next wait, so a step reads the terminals as the deltas after each unit bring them. Screens: `dropPeer` (clients come and go), its numbers reused. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* replica clients over a socket: the session streams each its dump and every ↵Gabriel Schneider2026-10-30
| | | | | | | | | | | | unit after it (replica.zig `Peers`, a writer and a reader thread a client, units at each step boundary, a client's input stepped there in its own view), and `pardes --attach=<socket>:<name>` and `--attach=<host>:<name>` (through `ssh -tt -L`) are a client in the tty host, its keys applied at once by its own replica and drawn by its own process wire.zig version 15: welcome (the client's id and view, the dump's length, the session's UTC offset and digest), dump pieces, units (with the authority's clocks), terminal messages both ways, a check. A unit carries one digest, at its end (`unit_digests`). The client: tty.zig `Shell.remote`, `feed` (a terminal's input to the replica and the session); the session's clock guessed from the last unit's (tape.zig `Guess`); a pump woken only by a client's post reads no new time (`mail_only`), and a unit of another's ending where the last ended leaves the speculation as it is (`neutral`); the rollback copy taken while idle (`Replica.ready`); views a frontend here does not draw left undrawn. A lagging client never holds the session: its queue is cut off past 64 MiB (PARDES_REPLICA_LAG_KB). Detach in a client ends that client; a Restore lets every client go. daemon.zig: a replica the daemon hands over is the session's loop's (a note), not a stub thread's. A window session refuses one in words. test/daemon.py `--clients`: two clients (one through the delay relay) type into a detached session and agree with it, one killed outright, one detached, a third typing into a shell. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* replicas' core for a socket: a client's input settles as a `step` (no clock, ↵Gabriel Schneider2026-10-30
| | | | | | | | | | | no wait, no frame) on both sides, every kind of input speculated, the session's clock guessed, views no frontend here draws left undrawn, and a drawn view's selections given back to its slot pardes.zig: `step`, the middle of a pump, is a client input's own top-level input on the authority and in the replica's speculation, so the two tapes read alike; a pump skips the live view when it is `undrawn` (a replica client's, drawn by its own process); `settle` is the pump's whole middle, the Upgrade's too. tape.zig: a `.step` input; `unattended` is a watched field; a speculation's clock reads answer a guess of the session's clocks (`Guess`); `resync` says the host and its writes again for a joiner; `adoptCore`; a short ret is shown in a trace. views.zig: `drawOne` draws a replica client's own view between steps; `drawParked` gives a drawn view's selections back to its slot, for drawOthers too. rollback.zig: std's lists by walk.isList (a safe build's have a third field). Co-Authored-By: Claude Opus 5.5 <[email protected]>
* echo: a replica draws a key typed into a terminal at once, mosh's way (a ↵Gabriel Schneider2026-10-30
| | | | | | printable character at the cursor, Backspace, Left and Right, in tty mode on the primary screen), from an overlay of its own that no digest, dump or rollback copy sees, painted over the terminal's body at the end of renderBody (the tty's and the GUI's); judged at each delta once the authority has the key and 50 ms have passed, every prediction taken back on a mismatch; drawn at once at a prompt pardes's marks vouch for, elsewhere only once an echo was confirmed, never after a drawn misprediction until one is, so a password prompt or a program that ignores input shows nothing typed; the keys typed behind one not predicted (Tab, Enter) predicted over the screen that shows its doing; zig build replicas -- --echo types a corpus into a played shell at 30/80/207 ms: a key to its character p50 239 -> 8.9 ms and p90 248 -> 15.8 ms at 207, no wrong character drawn; docs/next-daemon.md: the decisions, the hook for the remote client, the numbers, the gates and what is not done Co-Authored-By: Claude Opus 5.5 <[email protected]>
* 9p: a mount keeps its connection and sends each operation's walk, stat, open ↵Gabriel Schneider2026-10-30
| | | | | | and first read or write in one flight, a new connection's version and attach before them: an operation through it is one round trip where it was six or seven (a directory, or a file of no stated length, two: 9P's only end is an empty read); a file whose stat gave its length is read ahead, its Treads 14 at a time; reads are timed by their progress, so a large remote file opens at 80 and 207 ms (1.3 MB timed out); a write refused unopened behind a truncating open that parked is sent again after it; Unmount lets the connection go; the CLI, --wait and Handover dial afresh, pipelined; reference.typ says how a session elsewhere mounts through ssh -L Co-Authored-By: Claude Opus 5.5 <[email protected]>
* the walker's digest: a blob's hash made once and folded in as a word (a 100 ↵Gabriel Schneider2026-10-30
| | | | | | | | | | MB file was 5 ms a digest, two a unit, on the authority); blob fields annotated `.blob`, which the rollback copy's pins read too walk.zig: a field named in a type's `replica.blob` (a file's text, line index and location rows, an undo snapshot's text, a register's values and joined text, the rows an edit dropped, a query's saved rows) holds immutable blobs. A digest hashes one of 4 KiB or more once, keeps the word by its address and length, and folds the length and word in; `forget` drops the word, from memory.freeRaw and freeUnfilled, through which every such blob is freed raw. Only a Debug build (the tests, the harness, replay) hashes the bytes at every digest all the same and panics where the kept word is not their hash: a blob written in place, or freed past its word. A release build, ReleaseSafe the user's own as well as ReleaseFast, takes the word and never panics on a stale one: at worst a divergence, which a replica records and then stops at (nothing resets or rejoins it yet; the notes say so). File.zig: an edit's free of the old text went to `rawFree` itself and so kept its word; the next text of the same length at the same address folded in the old one (the ReleaseFast harness diverged in 162 of 180 replica runs, in `file.content`), and it goes through freeUnfilled now. Registers.zig: `put` makes a new values array rather than writing the old one. rollback.zig: the copy pins what `.blob` names (`pinned` and `pinnedDeep` gone). Measured, ReleaseFast on the perf cores: with a 100 MB file open (PARDES_DIGEST_BENCH=100, replica.zig's bench), a unit of a key into it went from p50 10.45 ms, p99 41.6 ms to p50 24 us, p99 6.1 ms (the first, which makes the word), a digest from 5.18 ms to 10 us. The harness (20 seeds x 120 inputs x 3 clients): rollback p99 9.5 -> 2.6 ms at 30 ms (1.5 and 1.4 ms at 80 and 207), copy p99 0.27 ms; none diverged, none unequal, no delta wrong. Tests: a blob hashed once, its kept word no longer its hash once written in place, its word gone when freed. docs/next-daemon.md: the follow-up. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* views and the rollback copy: a live view's slot keeps no pointer into what ↵Gabriel Schneider2026-10-29
| | | | | | | | | | the live core owns; a register's texts and a file's location rows are pinned, not copied views.zig: when a view goes live (to, close, drawOthers putting the live one back, only), its parked share is emptied (`emptied`): it was left holding copies of the pointers it handed the live fields (a column tag's text, a chord's argument, the presentation's), stale until it parked again. The rollback copy walks every slot: a stale pointer whose memory the live view had freed and another taken saved that other's bytes and put them back. A pane's share holds no pointer but its selections, which `unpark` already hands over. The test fails without it: a tag's text freed under view 0's stale slot, its memory taken by an allocation the copy has no business with, written under a held copy, and the copy put back. Registers.zig, locations.zig, rollback.zig: a register's values and joined text, and a file's location rows (the dropped ones, a query's saved), are replaced whole and never written, so the copy pins them as it does a file's text, freed raw (a safe build's free would write over one a copy still holds); rows pin whole with what each owns (`pinnedDeep`). In the harness at 30 ms the big copies were a file's location rows, 5-6 MB after the flood, not the register. ReleaseFast, 20 seeds, three clients, p99: copy 5.1 -> 0.22 ms at 30 ms (0.19-0.20 at 80 and 207); rollback 14.9 -> 9.5 ms at 30 ms (1.3-1.4 at 80 and 207), what is left being the replay's digests (half the harness's CPU is Wyhash over the whole core per replayed unit). Tests: a 2 MB register and 500 location rows pinned, not saved, and whole after a speculation drops them and the copy is put back. docs/next-daemon.md: the follow-ups. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* upgrade: a newer build at the path a session started from is said at a ↵Gabriel Schneider2026-10-29
| | | | | | focus-in, once a build, the path read just after the first frame (off a startup's way); an upgraded process lets the binary it replaced go (held, its deleted file kept its blocks) and sweeps its old image's prompt files (kept for an exec that failed); test/daemon.py --upgrade asks through /ctl, checks the fds and prompt files level, and runs on FreeBSD (procstat, the daemon's registry, no kept old binary there) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* upgrade: the Upgrade word (SPC s U, a tag, 9P) moves a running session onto ↵Gabriel Schneider2026-10-29
| | | | | | the pardes installed where it started, in place, by exec at a step boundary: its dump in the DumpDir, its pty masters, terminals, stubs and daemon link kept open and named on the new command line, every view and running shell its own again; the daemon re-execs too, its registry on its command line; a build that cannot load the dump falls back to the old binary, or starts afresh with the shells adopted and the dump kept (test/daemon.py --upgrade, --upgrade --gui) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* screen-state replicas: a replica's emulators on the C heap, outside its ↵Gabriel Schneider2026-10-29
| | | | | | | | | | rollback epoch, so a delta written under a held copy is no copy's to keep or put back (a ReleaseFast run found the heap corrupted) terminal.zig: State.gpa, what an emulator and its State are allocated with: the core's, but on a screen-state replica the C heap (emulator_gpa). replica.zig applies deltas and rows with it; the epoch's pause is gone (rollback.zig back to part A's but for the rows cache). The harness in ReleaseFast corrupted the heap after 20 seeds: a safe build's check named an allocation the copy saved and a delta freed while the epoch was paused, reached as `Text.own` through a view's stale slot (views.zig keeps the live view's parked share as stale copies, and the copy walks it), the replica's old terminal's pin map at that address. With the emulator outside the epoch, nothing of it is saved, kept or put back, and every free under a held copy is put off again, as in part A. Debug, 20 seeds at 80 ms and the defaults: no divergence, every replica and terminal the authority's. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* screen-state replicas: what a terminal's ingest tells the positions pardes ↵Gabriel Schneider2026-10-29
| | | | | | | | | | | | keeps over it taped, the history a step reads sent before its unit, rows sent a blob once a message, and the rows cache out of the rollback copy terminal.zig, tape.zig: ingestWithPositions tracks its rows only where an ingest happens; what the ingest told them (each tracked row's row after it and whether it was erased, a clear, whether the pages lived, a pin not made) is a taped value (Site.positions), which a screen-state replica, ingesting nothing, takes; the logic after it runs alike on both. On the stack, so the local path allocates nothing new. buildRows marks a terminal read (State.read) outside the render (terminal.drawing). screen_state.zig, replica.zig: the authority keeps each peer's holes as its replica does (Peer.holes, from the deltas it sent), and Screens.fill sends a peer the rows it lacks of every terminal a step read, after the step and before its unit, so a replay reads what the step read (a text-mode cursor moved into history the flood scrolled past diverged). encodeRows sends each distinct blob once a message and the rows as addresses (a fill after yes is a third). The rows cache is dropped when a copy is taken and emptied when it is put back, not saved. test/replicas.zig: Ctrl-b in the mix (a shell out of tty mode: text, cursors and selections over it), and a shell's session (prompts, titles, directories, clears) between the TUI and the flood. 4 seeds x 120 inputs x 3 clients at 30, 80, 207 ms: no divergence, every replica and terminal the authority's, 4,653 deltas checked. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* screen-state replicas: a rollback never brings a terminal's past back: the ↵Gabriel Schneider2026-10-29
| | | | | | | | rows the core keeps read of a terminal marked stale after a delta and after a copy is put back, and a test that a delta applied under a speculation outlives its rollback replica.zig: the core's cache of a terminal's rows (shell_rows) is invalidated by feedOutput, which a replica no longer calls; a delta and fetched rows now mark it, and so does a rollback, whose copy holds whatever rows were read before it was taken (a replica draws between steps). The test types into the text pane, lets the shell print under the held copy, lands its delta, reads the rows, and rolls back: the terminal is the authority's newest, the history as long, the rows the screen's. Without the stale mark it reads an older grid's rows; without the epoch's pause the rollback frees the emulator's pages and it crashes. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* screen-state replicas: a flood's rows renumbered in place, not grown and ↵Gabriel Schneider2026-10-29
| | | | | | | | | | trimmed by the thousand (a yes delta applies in 97 us, was 2.4 ms), holes blanked when they come into view, and input into a terminal left to the authority screen_state.zig: when a delta's scroll is more than the replica's screen and history hold, nothing it holds is in the new state; its rows stay where they are, renumbered, the active ones and the tail written from the delta, the rest holes, and only as many more rows grown as the authority's history is longer. A hole has whatever that row held before, so `blank` clears the holes in view before they are asked for. The bench's floods run under the C allocator, as a replica does. replica.zig: a replica's input whose live pane is a terminal is not speculated (its step reads the emulator, the last delta's here and the pty's newer one there): it waits for the authority, as a key in tty mode already did at its pty write. ReleaseFast, 200x60 yes at 16 MB/s: apply p50 97 us, p99 284 us; encode a frame 41 us at one peer and at three; 64 KB/s a peer. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* phase 5 part B, terminals: a replica's terminal fed by screen-state deltas, ↵Gabriel Schneider2026-10-29
| | | | | | | | | | | | | | several in flight against the version last sent, its history the authority's to the row, and the authority's own steps batched into a unit a frame screen_state.zig: each state the authority sends is a version of its terminal (Track), made once for every peer; a peer's delta is against the version it was last sent (the stream is ordered and loses nothing), up to 32 in flight, a peer at the bound waiting for an acknowledgement and then getting the newest; peers at one base are sent the same bytes. A replica refuses a delta against another base and is started afresh. It keeps the authority's history extent to the row (no byte limit of its own, trimmed to the authority's count), takes its viewport and its cursor's semantic content, and grows a flood's rows in bulk. The soak now keeps several in flight, lands them late in order, loses one now and then, and checks the history extent and the viewport too. replica.zig: Screens (the authority's tracks and peers) and Replica.screen/screenRequest, messages of u8 pane, u32 serial, u8 kind (a delta, rows; an ack, a row request, a reset). Deltas are applied with the replica's epoch paused, so an emulator outlives a rollback. Authority.output/pending: its own steps (a pty's output) go into one unit, sent before any other input it steps and at a frame, with one digest; under screen state a digest is one word. tape.zig, terminal.zig: under screen state an output is recorded without its bytes, pty writes are off the tape on both sides, no digest covers an emulator, and a replica's core never ingests or resizes one (a resize's own core work, the pointer selections cleared, still runs). A terminal counts its writes (State.changes); its MarkedHandler cycle rides in each delta. test/replicas.zig: the session's shell runs a TUI, then 8 MB of yes read 4 KB at a time, while three clients type; deltas go out before each input the authority steps and once a frame, acknowledgements and row requests come back over the links, every applied delta is checked against the authority's screen at that version and every terminal at quiescence. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* views: a view drawn while another is live keeps its selections, given back ↵Gabriel Schneider2026-10-29
| | | | | | to its slot after the frame; a headless replica, which draws nothing, kept them and diverged (three-view monkey seeds 5 and 35) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* views: a loaded session's live view's slot owns nothing, as `installed` ↵Gabriel Schneider2026-10-29
| | | | | | makes it, and so does the one view a revived session keeps Co-Authored-By: Claude Opus 5.5 <[email protected]>