| Commit message (Collapse) | Author | Age |
| |
|
|
|
|
|
| |
them from what a dump keeps (the pane's messages and prompt, the view's leader, rect and tag rows), never the derived copy a step or a park left in Pane.notices or Pane.pv, which a client joined from a dump has not: its kind was told other rows than the session's and the join diverged (`digest N differs in panes pane M.addon`); collectNotices makes the live view's from the same function
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
a key's or a word's into its pane is): a replica has no kind's code, so its guess of a word that opens a pane with no file (Snake, Life, Fake) made the open's outcome a default and was always mispredicted; it waits for the authority's unit now, as do a Look, a revived placeholder and a reload that reach the same open
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
(pollCwds before the no-window return, upstream's test ported onto forkShell's seven arguments), and a replica joined from a dump paints the restored theme's chrome (initDump runs finishThemeInitialization: Core.chrome is local and was never recomputed, so the join showed the default theme's tags until the next Theme, as the ported test showed before the line)
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
Slot.where_text on every core (digested now), which Jump and Recent read: they asked the kind through a taped placeOf that a Shadow answered from the place it last replayed, so a call in a step of another view's (after a window joined) moved the PDF's place unseen and the window's Esc back to it guessed 16:0.000000 1.000 against the authority's 16:0.000000 0.045
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
and replica (std.Io.Dir.path, std.Io.Dir.max_path_bytes), and the window's handover dump written and removed through the shell's io (Io.Dir.writeFile at 0600, deleteFile) instead of libc open, write and unlink
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
restoreName beside undoStep, Query's row makers (clearPlaces, makeAllRows, makeNameRows, feedGrepPlaces) beside produce, look's helpers beside lookAt, lookSpanIn and lookUnresolved, mouse's beside handleMouse and dragUpdate, normal's beside handleNormal and executeNormalAction, and the rest in Core, draw, exec, layout, main, modal, replica, syntax, terminal; Messages.printRight into draw.zig beside joinNotice and edit.spanHas into exec.zig beside operandAt, their only callers
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
that had one mid-file (the locality analysis's 45 and test/hxdiff.zig; gui.zig, macos.zig, look, host_io, 9p_io, terminal, tui, mouse, fs, replica, layout among them, 11,429 test lines), with the functions only tests call, moved after the file's code in their order, a mechanical move (.scratch/suites/tests_to_end.py), zig fmt'd; the unit suites pass as before (tui 1486/1487, gui 1536/1537, core 956/956 on both), macos.zig only parse-checked here
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
over PA_RASTER_MAX is dropped at paint and said once, a kind's frame names its placements by key and their pixels cross before it as `raster` strips of at most 4 MiB over the rows its `src` shows (only those the peer was not sent with the frame before), the peer assembles them by name in a 64 MiB cache let go by least recent use but for what a Shadow shows, a frame naming one it lacks draws its cells and asks once for its pane again, the sender checks every message against wire.max_payload (Debug's panic, release's drop and log) and a reader names an oversized one (said on the client as it joins again, logged at the session) instead of reading it as the link's end, so kinds #21's byte budget goes; wire 23; and pdf.c lays out for a view with pictures over a glyph view until an input chooses (one layout until 7c), so the 110-column window of the stall's repro shows its page
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
field a step reads (§17.2 B 2-3): the Slot says what is the kind's own or a frame's (its code and memory, its jobs' inputs, a placeholder's record, the place its kind says, the thumb, the pointer, whether a frame was drawn) and the walker digests the rest, `placed` and the prefix leave the Shadow's frame (the pointer its one byte), the dump carries the cancelled tokens a replica's job answers are dropped by, and a pane test dumps a session with a PDF and a picture, loads it on another core and compares every Slot field the walk does not leave out, so a field added and not dumped fails a test, not a soak
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
B 1): a step of the session's own (a tick, a job's answer, a 9P request, any input no client sends) tells each view whose facts moved and ticks a focus changed or a deadline come, a client's input ticks only the pane its key, click or word went to (which already waits), so a client never guesses a tick and never waits for one; the tick's prefix is in its taped outcome (a Shadow's tag is the authority's from the tape, `told` and the frame's prefix gone, the dump carrying the prefix), Kind.wake says a focus change, the tick wait and Kind.repaint are gone, the post-unit flush stays for the tick after a key; the stall test asserts the replay (0 mispredicted, 0 waited)
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
view told each open view's facts by id when they change (the slot's view table, a hash a view, a stopped view told cols 0, and a step woken for a view a frame's layout changed) and ticked after, the tick's pa_tick and its prefix (placed and retag gone: the art words go on every pane of a PA_PICTURES kind from the first frame, the same bytes whichever view's step lays the tag out, by the session's first view, which was the 30 ms soak's rollbacks), found's rows the search's or appended to its +Search after, cancel's tokens dropped on every core (on the tape), a picture over PA_RASTER_MAX dropped and said once, a record of at most PA_RECORD_MAX, a Debug call over 16 ms of the kind's own (its allocations and job copies the core's) said once a site and failing --addon-test (which gains view N on|off and prints the tag the core makes), and the record §17.5's: a Restore, an Addon load and a reload read the pane's file as a Look does, a file not there a placeholder saying <path> is not here, record kept, opened by the next Look or watch of it
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
paints it once for the live view (Kind.repaint) before its tag is laid out, where the tag took what the frame before the tick placed and the next input's step another (a replica's guess of that input, the last step's tag, missed: 13 of the 30 ms cell's 14 visible rollbacks), so a pane of a kind's tag has its art words from its first step (the fake and pdf addon scripts' windows a row shorter), and an input's step sends each client what it made of a kind's pane after its unit, not at the next pump (a terminal client looked at the frame from before its own key's tick, the text layer missing from its view)
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
the wire's bound (encodeShadow takes a budget and leaves out a raster past it, where a 17 MB page in a 110-column window ended every client's link and every join's after), a join's dump keeps a kind's tag words and selection (SlotRecord tail and selection, which only Kind.reopen made: a joiner's tag wrapped another way and it diverged on tag_rows at each join), and a speculation whose step would tick a kind waits for the session's unit (Kind.step's rule, read only) where it guessed no tick and rolled back, every visible rollback of the 30 ms soak cell
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
amount (a window passes a touchpad's fraction of a turn over a kind's pane, a terminal whole notches), pa_view says glyphs (no graphics), pa_paint's tail padding holds the pointer the kind asks for (the window shows a hand for PA_POINTER_LINK over its pane, a Shadow's frame carries it), and tick is told PA_FRAMED when a frame was drawn since its last; addon-test gains graphics on|off
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
every pdf_enabled branch with them; look, Jump, Recent, dumps (one from before the addon opens as its pane), the watch, mouse, keys, the rail and the tag go through Kind for any kind's pane; test/pdf.zig's behaviours are T1 cases over 9P (addons_pdf.zig) and pane cases (a Shadow over a replica, a file changed under it), the harnesses load pdf.so by path
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
attach to a session attached already (as it did for one of the other kind), so a second `pardes --attach` or `pardes-gui --attach` on this machine is a replica client through the replica socket, the path a remote one takes; a session tells the daemon its clients (SO_PEERCRED on each replica socket, at every join and leave) and `--ps` lists `client <pid> <command>` under it; the tty's per-terminal Screens (Screen.create/drop/destroy, primary, its stub and Killed slot), the Poster and its `from` input, the Gui's window-fields swap (window_fields, GuiWindow, copyWindow, freshWindow, follow, route, openAnotherWindow, closeView, windowsOpen, closeOtherWindows, resumeWindows), drawOthers and every harness's extra terminals (snapshot.zig's `start ... views=N` and `view <k>`, e2e_harness's initView and drainOthers, PARDES_TEST_VIEWS in both frontends, the GUI feed's `view;<k>`, the monkey's --views) go; the pump draws the one view this process shows (views.shown), through drawOne when a client's input was the last; an Upgrade from a build that kept a second terminal of its own lets it go (its stub told it is detached) and closes its view; deleted: views.snap, views-term.snap and views-pan.snap with their goldens (the in-process second terminal they drove is gone; views.zig's unit tests keep the per-view cursor, layout, register, sizing, closing and panning behaviour, test/replicas.zig and the soak the views of clients); the two views.zig tests that drew through drawOthers draw through drawOne; daemon.py's second terminal is a client listed by --ps, typed into in both terminals, detached from inside and closed; its second window a client listed by --ps and closed; the reference and next-daemon.md say so; no golden changes; gates: core-test 956/956 both, behaviour-test 248/248 both, unit-test 1487/1488 tui and 1530/1531 gui, pane-test 171/172 both, snap all 102 ok, process-test 0 of 37 failed on both, hxdiff 19 and hxparity 6 mismatches (the base's names), hxgolf 0, replicas, addon-test, gui-golden and monkey-gui 0 of 1 failed, daemon.py main, --window-clients and --gui ok, soak local pass (differ, mispred, view_bad, lost 0)
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
(the ~42 fields views.fields named: size, cell metrics, columns, rects, focus, jumps, pointer, gestures, prefixes, frontend), one of Pardes.views[4] (null when closed), the live one views[live] through Pardes.view() (index field `view` renamed `live`); going into another view is the index plus the panes' shares, which stay swapped (park and unpark now move only Pane.pv); views.zig loses fields, Part over Pardes, Parked, kept, emptied and the session half of every copy, and a pane closing leaves every view's layout at once (views.gone, at the two places a slot is freed) instead of the serials each parked view kept; drawOne and within put the drawn view's own share back whole; the views are in the digest now (tape's per-field parts name views[k].field), so a fresh view's col_x/col_w default to zeros, not undefined; the dump keeps every open view in the core and the panes' shares of the others under `shares`; an image from before (live view as the core's own fields, the rest under `views`) loads them into views[] once; the ~2,400 references rewritten mechanically from the compiler's errors (p.active is p.view().active, `const v = p.view()` hoisted in 823 places), Kind.zig (4 lines) and addon_test.zig (3) among them, macos.zig by pattern (not compiled here); tape's watched host writes read the live view's frontend and active; perf.zig's accessor lines change its harness hash, the debug baselines re-keyed to it (their numbers the base's); no golden changes; gates: core-test 956/956 both, unit-test 1487/1488 tui and 1530/1531 gui, behaviour-test 248/248 both, pane-test 171/172, snap all 105 ok, process-test 0 of 37 failed tui, replicas ok
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
architecture §16.9 has it (pa_window 24 bytes and no art bytes, pa_placement 80 with at, from and dirty_y/dirty_h and no stride, pa_word 32 with no size and PA_WORD, tick(st, now_ns, flags) with PA_FOCUSED, PA_OK and the conventions, the dump and load rules, pa_put with attrs, pa_copy and PA_POD, whose fragment gcc -Werror refused as written and is ternaries here) and fake.c and image's table speak it; tick is called at its deadline and once after open, a taken key or click, a word, a job's answer, a load and a change of PA_FOCUSED (the pane has the keyboard in some view, its window focused), nextWake folds the deadlines, and a Debug build hashes the kind's allocations across paint and panics naming it; paint runs once a frame, the slot keeps its prefix and whether it placed a picture, the tag is that prefix, the core's art words when it did and the path, and a frame that changes it owes a step to lay the tag out again; Addon load and reload are an addon effect the host performs (memfd, BLAKE3, dlopen, the head checks and init, the bytes a process holds shared and counted) answered by an addon event carrying what the library brings as data, whose update does the collisions, the reload's validate-then-swap at a new generation and the waiting panes deterministically, while unload is the update's alone and lets the library go by effect; a kind's job is Effect.job run on the host's workers (host_io.KindJob: tui and gui on Io.concurrent, macOS and tests on the loop) answered by Event.job, stale generations dropped and a library closed after its last job; each core has its own Registry (16 kinds, their words, leaders and extensions, digested and dumped; their code a local side array a replica has none of), so a test session no longer unloads the world and Restore and -l keep the process's kinds; every call into a kind whose answer the core acts on is taped as an Outcome (its answer, refusal, message, Look, text pane) and the prefix a step reads is taped too, so a replica's core, which runs no kind's code, steps alike; a pane of a kind on a replica is a Shadow drawn from the authority's frames, painted for the peer's view and sent by Screens as row hashes with the rows that changed and each raster once a key (screen_state.Shadow), its keys going to the authority; the tail, reload and the Addon word and event are Addon.zig's, the slot's record dump.zig's, a tail word's run exec.zig's; --addon-test gains focus, blur, dump, body, ctl and the window line, its jobs done as a host does them; fake.out gains its window line, loses a body row once the tag carries the art words and lands FakeJob at once; tests rewritten: the image dump test (the replica-style load keeps the record), the Petscii/Palette/Ascii tag test in ctl.zig and test/panes.zig's image tag test (a picture decoded and shown first), fs_namespace's decode test (the job performed as a host would), Kind.zig's registry test and Addon.zig's crash-mark test, the addons suite's ABI numbers; new: a job under way across a reload (T1), a Shadow's frame codec, and a replica with no code for a kind showing its pane and passing its keys; gates: core-test 958/958 and behaviour-test 249/249 on both, unit-test 1490/1491 tui and 1533/1534 gui, pane-test 171/172, snap all 105 ok, process-test 0 of 37 failed on both, addon-test ok, hxdiff 19 and hxparity 6 mismatches by the base's names, hxgolf 0, replicas, gui-golden, monkey-gui, daemon.py main and window-clients, soak local pass, perf medium within 1% of the base
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
clock are fields of their nouns: Jump.zig is the jump list's type (list, n, cur, force, its own len annotation) and the session's jumps a Jump where it was jumps, njumps, jcur and jump_force; lsp/Ask.zig is the question's (seq, last, wait) and the session's lsp_ask an Ask where it was lsp_seq, lsp_last and lsp_wait; the clock is the session's clock, a Clock (now_ns, stepped_ns, started) where it was now_ns, stepped_ns and clock_started; a view opened still starts with no query's mark on its jumps and one closed leaves the staying view's (views.zig), as when the mark was a field of its own; callers in 26 files read p.jumps.n, p.lsp_ask.wait, p.clock.now_ns (macos.zig's lsp_wait tests by text, built nowhere here; dump.zig, the persistence agent's, only where it names lsp_seq); a dump written before this change loads these as their defaults once: an Upgrade or Restore from it comes back with no jumps, no language question awaited and its clock unstarted (jumps' kind changed, so the decoder says so), and the wire's semantics change; gates on tui: core-test 960/960, behaviour-test 238/238, pane-test 171/172 (1 skipped), snap all 108 scripts ok; gui unit-test 1518/1519 (1 skipped)
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
rows, an undo entry's bytes) is taken as its length, never its bytes, a text's revision saying what it holds, where it was hashed once and kept as a word by its address (walk.zig's blob words, from when no text was written in place): since the journal a session's edit writes its text in place, and since the change before a replica's does too, so the word went at every edit and the next unit's digest hashed the whole text again, 100 MB of it in a big file, on the session's step and the replica's; the words go with what kept them (walk.forget, wordOf, forgetAll, unhashed, makeWords, the bench's blob_words_on, memory.freeRaw's and unpin's forgets, File.edit's and grow's), and so does the join's hashing thread (replica.Peers' held, texts, hasher and rounds, joins, hashTexts, unhashedDone, the `hashed` note), which made the words of the texts a session never digested before a join's digest took them, its pins with it: a client is joined at the boundary its connection comes at; the semantics change, so the wire's version is 20 (a client of the build before is refused at its hello, as any other build's, and test/daemon.py's own hello says 20); a blob under 4 KiB is hashed whole as before, so a small text's bytes stay in the digest, and the replicas harness (test/replicas.zig) compares every file's text byte for byte at quiescence beside the digest; tests: walk.zig's 'a blob is hashed once and folded in as a word; one written in place is not its word, which a Debug build checks; a freed one's word goes' becomes 'a blob of 4 KiB and up is digested by its length, what it holds told by its keeper's revision; a smaller one by its bytes'
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
as the session's does, and its rollback copy keeps none of their bytes: the epoch notes what each write into a text or an index that was there when the copy was taken took out (rollback.Epoch.wrote, from File.edit and shiftIndex), and putting the copy back writes that back, the newest first, before the saved bytes go back (unwrite), where a replica's edit made a whole new text and a whole new index at every key so that the copy could pin the old ones; the bridge the journal left for this goes with it: rollback.isEpoch, File.keepsBytes (a text a join is still sending is copied on write as before, memory.isPinned), the replica's copy-on-write in File.edit and shiftIndex's `keep` (its in-place shift is File.shiftStarts, which the rollback runs turned round); the journal's steps, its `done` and its entries come back as the copy saved them, an entry a speculation merged into and the redo entries it dropped being saved bytes and put-off frees; File.State's content and line_starts are annotated `.journal`, which the copy neither saves nor pins (line_starts leaves `.blob`); a write that could not be noted (no memory) fails the replica at its rollback; tests: rollback.zig's 'a text written in place under a copy comes back byte for byte: typed into its room, undone, cut, and grown past it' is new, and File.zig's 'a line index shifted in place equals a rebuilt one' becomes '..., and the edit turned round gives the one before back' (its every tenth shift made anew, as a replica's core made them, gives way to the shift turned round in a copy, as a rollback makes it)
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
cap, the newest first, and lets the rest go, where it let the least recently used go until the rest was under it: a text past the cap alone (a big file's, edited since the client joined) took every older one with it, the small ones that fit too, so a client after it with that cap was sent them all again (daemon.py's far clients E and F at a 1 MiB cap: F found nothing on disk, the 6.5 KB notes gone with the 1.1 MB file's texts; now the notes stay); tests: replica.zig's 'the blob cache stays under its cap, the least recently used going first, the file compacted' becomes '... going first and one past the cap alone without them, the file compacted', a 20 KB text past its 15 KB cap put newest, gone, and the three older ones kept
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
Cache, in $XDG_CACHE_HOME/pardes, else ~/.cache/pardes), a file of its own and never the state file (expendable, and up to a gigabyte): `blobs/<hash>` the texts, `seqs/<hash>` and `seen` (a sorted map, a number to its hash, the least recently used first) their order, `size` and `next`; a join's blobs are put in one transaction by the client's storing thread, with those it found there made the most recently used (a hit takes a new number), and at the client's end what it made since is put and the cache trimmed: past the cap (PARDES_BLOB_CACHE_MB, 1 GiB) the oldest go in one transaction and the file is compacted into `cache.xitdb.new` and renamed over it, where a directory of files was scanned, sorted and deleted from at each join; each use opens the file for itself (a database a thread, as xitdb asks), a text read back is hashed again (one kept wrong is a miss, and goes), and a file that is no database is started afresh; replica.zig's Cache (its directory, the hex names, put by rename, trim by scan) goes for Db.Cache; tests: replica.zig's 'a cached blob is hashed again as it is read: one cut short, corrupted, planted or longer is a miss, and goes' and 'the blob cache stays under its cap, the least recently used going first' are rewritten for the file (texts kept wrong under a hash through put, the order made by hits, and the file seen compacted), 'a blob cache that is no database is started afresh' is new, the far join test keeps its biggest blob wrong through the cache rather than a file, and daemon.py's far clients read the cache file's size, cut it short by hand where they corrupted files, and check the cap after a client's end where they planted junk files
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
architecture §14.3), one fsynced transaction: `dumps/<label>` holds the session as a database of its own, which the walk writes by field name into xitdb's maps (walk.Put: a struct a map keyed by its fields' names, a union a map of its one field, records a list, plain numbers one bytes value tagged with their kind and width, an enum its name, a number a uint or an int, an optional an absent key, a type with dumpWrite its bytes, a .blob text of 4 KiB and up its BLAKE3 hash) and another build reads by name (walk.Get: defaults, notes, a pane it cannot read left out, as the byte loader did), with the hashes of the texts it keeps once under `blobs/`, so an unchanged text is written once across dumps; labels are `pardes-<date>-<time>` (`-2` and on in one second), `last` names the newest, the newest 32 are kept and the texts no kept one names are dropped in the same transaction, and at a session's end the file is compacted into `pardes.xitdb.new` and renamed over once it is past twice what it holds plus 64 MiB (another session notices the new file by its inode); a dump travels as an image, a database in memory holding it alone, deterministic (std.Io.failing's zero ids, so the replay's fixpoint holds): what Dump hands its host, what `Dump export <path>` and $PARDES_DUMP write whole (the snapshot goldens' `Restore <path>` unchanged), what a Restore of a label is read into, what a handover carries (`handover-<seq>`), and what `-l` and `Restore <path>` read; Restore takes a label or a path (bare: this session's last dump, else the newest), a label the file keeps none of refused in words (`Restore: no dump labelled X`, asked of the authority as the closed panes are); `/dumps` lists them, `<label> <date> <size> <panes>`, the newest first; Upgrade commits `upgrade-<pid>` before the exec, its record naming the state file and the label (a record without one, an older build's, names its `.dump` file, read as before), and the label stays; the byte dump stays for the join, the closed panes and the old `.dump` files (dump.byteDump), and dump.outPath, uniqueIn and upgradePath go for dump.label; measured (Debug, a session with a 3.2 MB file and three scratch panes, 20 dumps): the image on the loop 64-66 ms (59 ms of it BLAKE3 of the file, 0.65 ms in ReleaseFast) where the byte dump was 0.7 ms, the commit 9-16 ms with its fsyncs where the file was written in 2, a Restore 1.8 ms to read and 11.4 to load, and the file 3.55 MB after the first dump and 379 KB a dump after it with the big text unchanged (3.58 MB a dump with it edited each time, as twenty `.dump` files would be); tests: dump.zig's 'two dumps in one second get names of their own' and 'a dump makes every directory missing on the way to its DumpDir' become Db.zig's 'two dumps with one label get labels of their own' and 'a Dump makes every directory missing on the way to its DumpDir, and the file there', 'a session read by name, field by field as another build reads it, is the same session' becomes 'a session's byte dump read by name...' beside the new 'a session's image, read by name as every load reads it, is the same session, and dumps to the same bytes again', two replica tests compare the join's stream with dump.byteDump, the Handover tests read the carried image, upgrade's record test carries a label, the root listing tests and the monkey list `dumps`, the behaviour suite gains five root-ctl tests (a Dump and Restore round trip through 9P with `/dumps` and the tag's word, a label kept none of refused in words, the 33rd dump dropping the first, an unchanged text written once across two dumps by the file's size, a killed Dump leaving the one before it whole), the test host performs a Dump's write and a Restore as a host does, the monkey lets go of its log follower at a Restore, and daemon.py's upgrade case finds the kept dump in `/dumps`
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
DumpDir (src/Db.zig: the xitdb database the architecture's §14 puts everything pardes keeps on disk in), which every session shares: read at a session's start (now after the startup config, whose DumpDir says where the file is) and again on use when the file's committed size has moved, written at the end of a step that changed it in one transaction under the file's lock, taken without waiting (busy, the next step tries again), another session's commit read and merged under this one's changes first; not fsynced, as the `recent` file was not (a xitdb commit costs 20 us without fsync and 5.2 ms with its two on this btrfs); a file there that is no database of pardes's is kept aside as `pardes.xitdb.broken` and a new one started; Store.zig keeps the closed panes alone (its file half goes: the path and its inode/size/mtime stamp, load, refresh, read, flush and write, and W4's one-time read of the old `store`, which comes back with the migration), Options.recent_path becomes Options.state (off, disk, memory), main.recentPath goes, a Restore's new core takes the old one's database, and the test host's sessions keep theirs in a `.memory` database, the same code; measured (Debug, 200 files listed): flush after an open 133-135 us (W4's file: 124), refresh after another session's write 239-247 us (198), refresh with nothing new 3 us (1), a session started on it 1144-1194 us (1214), and a Recent change adds 12,325 bytes to the file (its 12,089 bytes of lines and 236 of xitdb's); tests: Store.zig's four Recent tests become Db.zig's three (two windows sharing Recent through a busy lock, the newest 200, a file that is no database kept aside; the old store's migration test goes with that read), the closed-pane test keeps its second window on a state file in its DumpDir, the behaviour suite gains 'two sessions on one state file share Recent', and the process tests and the canary look for the state file where they looked for `recent`
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
edits (Text.Journal: entries of an offset, the bytes taken out and the count put in, grouped in steps between marks, each step with the place it began at and the place it was undone from), bounded at 256 steps or 64 MiB, the newest kept; pushUndo marks a step (none when the last changed nothing, as a snapshot equal to the text was none) and every edit goes into the last step, joined to its last entry when they touch (typing on, backspace over it, a write going on where the last ended); undo and redo turn a step's entries round, one in place, several in a copy whose gap moves from one to the next (File.Gap), one edit and one revision as the snapshot's was; the snapshots go (File.Snapshot, the undo and redo arrays and their copies of the whole text at every step); the replica's copy pins an entry's bytes (Entry.removed, a blob) as it pinned a snapshot's; tests: 'undo keeps its steps within a byte bound, the newest always' reads the journal, 'file history owns bounded snapshot storage without per-edit bookkeeping allocations' becomes 'file history keeps its last undo_max steps, and an undo or redo with no memory changes nothing', the far join's blob test makes its two blobs beside the text with a delete and a yank (its history held whole texts before), and the counts of steps read history.journal.done
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
-Dplatform=esp32p4 with its pinned riscv32 target, -Desp32p4-cols and -rows, the Peek, Poke, Hexdump and Gpio words and the Board behind them, the board's 9P engine options, the gpio_toggle host slot), and with it every guard and limit only the board needed: terminal_panes and terminal.enabled (ghostty is in every build), memory.limits' board branching (the desktop values stay), MAX_PANES 64 and MAX_COLS 16 fixed, the board's boot buffer and its embedded boot, the pager's and the terminal's emulator-less paths; tests: the board's GPIO, ring and Board ones go, the boot-OOM and transition-grid ones stay under their own names; the perf harness loses the board's fixture, so the Debug baselines of both frontends are recorded again (--only medium --reps 10)
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
the end of a step that changed it and read back under a session's own changes when another session wrote it since, and the panes a session closes are kept in its memory, the newest 64 within 64 MiB; the append-only CRC log with its resync and compaction, and the closed panes' self-describing dumps in it (only the session that closed one ever listed it), go; Options.store_path is recent_path ($XDG_STATE_HOME/pardes/recent), and the old store is read once while it is the newer of the two; tests: the store's four rewritten through the same windows (the 200 newest, a foreign file, a closed pane's way back from memory), the old store read once, 64 closed panes kept, recent.zig's merge; the process cases read Recent's file; docs/next-store.md is the decision
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
floods (PARDES_SCREEN_STATE=<dir>, which nothing sets) are `zig build replicas -- --screen-state <dir>`, replica.zig's big-file digest bench (PARDES_DIGEST_BENCH) is the last of rollback-bench, and Store's timings (STORE_TIMES) go, no bench step taking a store
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
and taken and clearTaken go for the tape's buffer itself, as replica.zig reads the tape's other fields
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
keepSince's, which confirm's two branches each spelled out
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
batch's), where digestHere was the same function and leaveInput inlined it a third time
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
places each built a writer, described and printed under their own PARDES_REPLICA_TRACE test
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
refuseReplica (its words clamped as the daemon's were), where replica.zig kept copies of writeAll, nosignal and a refusal of its own
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
waitForDel's unused io; daemon's find returns null with its words, drop takes the slot alone, send, recv, Got and setName are private; replica's Joining is the daemon's and PARDES_REPLICA_LAG_KB, which nothing sets, goes; host_io's dead line variable; Handover's own readdir is std.c's)
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
variable goes
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
is inlined in Peer.readLoop, its framer freed before the peer is said gone
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
deliver(.fill) with the sizes given after, as deliver already did
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
one file since the phase 3 frontends went, to src/wire.zig; its imports and the paths that name it follow
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
replica.join (the session dialled, its build checked, its dump loaded and checked, or joined again for its view once its link ends, the core it leaves held meanwhile), where tui.zig and gui.zig each spelled out the dial, the check, the load, the redial and their timing; Link.reset takes the join, and Link.load, Link.joinedAs and Link.redial go
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
delta a client applies; each input, boundary and join a session steps) are kept only when PARDES_REPLICA_STATS reads them or an in-process harness does: a session or a client attached for days grew them a word each, without end
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
|
| |
--tui with --tty kept as its old spelling, the process tests' backend .tui and --backend=tui, docs/typ/builtins-tui.json, the Settings report's platform), and the builtin Tty is Term (its registry entry, SPC n t, every default tag, the docs, tests and snap scripts), Tty still run as Term through exec.acme_words, Tty+fish too on a tag and a ctl; tty as a pane's raw mode and as the OS's terminal unchanged
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
|
| |
|
|
|
|
| |
keeps them): they are painted from what the process has painted, the owner's highlights a frame made and the locations cache, both derived, so a window joined from a dump made a `/` query's rows plain where its session had coloured them (the box soak's `pane 21.file.location_rows` at win207-4's `/r`, unit 4/447); replica.zig's test, main fails it at `pane 1.file.location_rows`
Co-Authored-By: Claude Opus 5.5 <[email protected]>
|
| |
|
|
|
|
| |
digested and dumped, no longer derived: a key's move reads it before its step syncs (halfPageMove, ensureCursorVisible, through File.geometry), and a core loaded from a dump had the default, Wrap on and no row heights, so a window that joined after `SPC t w` scrolled further at the terminal view's next Ctrl-D (the box soak's `pane 23.file.scroll`, seed 7: win207-3 at its join); replica.zig's test, main's fails at `pane 0.file.scroll`
Co-Authored-By: Claude Opus 5.5 <[email protected]>
|
| |
|
|
|
|
| |
the panes opened in it: an Upgrade's clients and one whose bye says it joins again (a divergence's, wire 19's `bye.again`) leave their view kept for them (`views_away`, digested and dumped, so an Upgrade's new build has it; taped `awayView`), the client names it in its next hello (`Hello.view`, handed on by the daemon with the connection) and its join takes it back (taped `rejoinView`) rather than a copy of the session's; the live view is never a kept one, `views.count` counts the others, and a join with every view taken lets one go (every Upgrade made each client's own panes leave its view: the hunt's 80 ms cell lost `+Search` and `+New`)
Co-Authored-By: Claude Opus 5.5 <[email protected]>
|
| |
|
|
|
|
| |
emulator on the C heap, as terminal.create makes one (terminal.emulatorGpa): loadPane named no `.terminal` allocator, so its emulator was the rollback epoch's, and its screen replica's blobs, made with State.gpa, were freed with the C heap's behind the epoch at a client's end and again by it (H21: `double free or corruption` in Epoch.deinit at a hangup, `free(): invalid pointer` at a rejoin), and a rollback freed what a delta under a held copy made of them
Co-Authored-By: Claude Opus 5.5 <[email protected]>
|
| |
|
|
|
|
| |
digest to check its unit's against, and a blob born under the copy and digested there keeps its word past the epoch's free of it (src/walk.zig), so a later blob at that address read as it (the replicas harness, Debug, seed 11 at 207 ms: "not what its kept word was"; in a release build a divergence); a speculation over a terminal is confirmed by its records, as any other, its unit taken at once when more is applied on top
Co-Authored-By: Claude Opus 5.5 <[email protected]>
|
| |
|
|
|
|
| |
client's guess of the session's clocks (wire 18's `stamp`), the guess its speculation records, and the authority's step reads that, not its own clocks (within a second of them, else its own), so a step's clock read never mispredicts: an edit's second across a turn, a message's millisecond (Ctrl-W's, a jump's) were each a rollback (fixer 1: 4 seen in 2 minutes), and every reapply of one now reads the same time
Co-Authored-By: Claude Opus 5.5 <[email protected]>
|
| |
|
|
|
|
| |
while one has the keyboard (H14): what reaches its program is a pty write and stops as before; its view (normal-mode keys, the wheel, Ctrl-u) moves here at once, the rollback copy keeping each terminal's view (it holds no emulator) and a delta setting it, put back between the units it replays in the order the deltas came among them; what was applied over a terminal applied again over each delta or rows that come while it waits, and its unit taken at once rather than kept to replay over later deltas; closing a terminal stops (its pty's end is the authority's)
Co-Authored-By: Claude Opus 5.5 <[email protected]>
|
| |
|
|
|
|
| |
(PARDES_REPLICA_STATS's `stops`: a host function's, a site's, or a poison's reason, which tape.poison now takes; PARDES_REPLICA_TRACE says it with the input); a window client's PARDES_TEST_LATENCY says each input (`I seq ns at-once`) and, after each unit, the first of its own not confirmed (`U ns seq`), so a harness finds the frame that shows an input that waited
Co-Authored-By: Claude Opus 5.5 <[email protected]>
|