summaryrefslogtreecommitdiff
path: root/test
Commit message (Collapse)AuthorAge
* hxlib: hxdiff's cases on the library, against helix's goldens and against ↵Gabriel Schneider41 hours
| | | | the file pane (docs/next-editor.md, R6): test/hxlib.zig types each case into a `pardes-editor` value and prints its result as hxdiff prints a pane's, doing what an app does with the s, S, K and Alt-K it is handed back (the keys to <ret> its pattern, <esc> no answer) and printing a case that reaches any other action (a view, n and N, a macro, gw) as the caller's; hxdiff takes another driver's results (`--actual=`), and `zig build hxlib` (`-- --parity`) holds them to the goldens or the file pane; the key notation is test/hxkeys.zig, which both harnesses read; what the cases found, fixed: paneCursorLines gave the library's body no lines (mm, mi, ma, md, mr, insert-mode arrows), and the library's selectRegex keeps its pattern in `/` as pardes's prompt does (look.submitSearch); a page key in insert mode moves a view and is the caller's; test/lib's `zig build bench -- <KiB> <keys>`: a key's cost, the pages it changes, a parse; hxlib 537 of 658 compared with 0 mismatches (5 waived as for pardes, 121 the caller's), --parity 617 of 738 with 0 mismatches and 0 waived; hxdiff 658 / parity 738 / hxgolf 124 with 0 mismatches, core-test 811 of 950, shell-test 990 of 993, replicas pass, lib-test pass
* lib: the text pane as a library, `pardes-editor` (docs/next-editor.md, R5): ↵Gabriel Schneider41 hours
| | | | src/lib_editor.zig is an editor a program keeps as plain data (`Editor(capacity)`: a gap buffer for the text, Text's cursor fields by name, two byte rings of whole-text snapshots for undo and redo, registers as records; no pointer, no allocator), thawed into a Text and Registers in the allocator a call is handed, run through normal.zig and edit.zig as a pane's keys run, and frozen back, a key that does not fit refused whole (`key`, `run`, `selectRegex`, `setText`, `setRegister`; `Outcome` done, pending, unbound, action, refused), and `Syntax`, a parse derived from a grammar handed in; the seam is `pardes_config.lib`, true in the library's build alone: Text.pane is null at compile time there, a body's characters and undo are its own (flatSurface, editText, setEditTextSpan, pushUndo, undoStep), bodyFile and filePane are none, Esc and `*` do the text's part only and `+` is a plain register, and config.zig's check of the builtins' leader paths is the editor's; normal.textKey is the mode dispatch keys.zig and the library share, handleNormal hands back the parse and runAction runs an action across the ranges, and applySelRegex's search over the ranges is selectByRegex, which the library calls with a pattern; build.zig: `lib` in the options, mvzr's patch and uucode as functions (patchedMvzr, uucodeModule) the editor's module takes too, editorModule, and `library` exports `pardes-editor` to a package that asks (`.editor = true`: it fetches uucode, mvzr and vaxis, of which it compiles the grapheme width), so a pardes-syntax consumer fetches what it did; test/lib: a value typed into, copied elsewhere and its memory gone, types on equal to one never moved, undo past the ring and redo back, a text that does not fit refused with the value unchanged, an action handed back and a regex done by the caller, a language handed in parsed and coloured; core-test 811 of 950, shell-test 990 of 993, hxdiff 658 / parity 738 / hxgolf 124 with 0 mismatches, replicas pass, lib-test 8 of 8
* edit: the routing of a text's characters takes the editing context ↵Gabriel Schneider41 hours
| | | | (docs/next-editor.md, R2, first slice): flatSurface, editText, editTextEol, setEditText, setEditTextEnd, setEditTextSpan, paneCursorLines, paneByteAtDisplay, pinPaneCursor, paneWrapWidth, treeIndent, pushUndo, undoStep, stepOwn, selText and vselText take `e: Editor` where they took `*Pardes`, the pane they route to reached through `e.p`; 149 calls in 16 files hand `p.editor()`, nothing else changes; core-test 810 of 949, shell-test 989 of 992, hxdiff 658 / parity 738 / hxgolf 124 with 0 mismatches, replicas pass
* perf: the terminal stress cell compiles and its tail check wants what the ↵Gabriel Schneider41 hours
| | | | cell draws: measureTerminalSession and the gesture cells read the emulator through Term.emulatorOf (Term.State became a handle in the terminal's own module, and test/perf.zig's two `state.vt` reaches no longer compiled, so pardes-perf, and the unit-test perf gate with it, did not build on this line); checkTerminalTail wanted a tail cell's colour as palette index 2, written when theme 0 was helix, and since orchard became the default theme (682e237df) a cell with the pane's filter off (as the cell sets it) is the theme's own colour 2 as an RGB: it wants pardes.ansi.inkOf(theme).palColor(2), and asserts the filter is off; test code only; `zig build perf -- --terminal-mib 1` passes (1 MiB in 121 batches)
* integration: the terminal a module of its own, with the plumber whole and ↵Gabriel Schneider41 hours
|\ | | | | | | | | | | | | | | | | the walker's cheaper compile A merge of term-phase1a and pardes-next, no conflicts. core-test (the core with no terminal) 411 passed, 535 skipped of 946; shell-test (the same tests, the terminal compiled in) 989/989, on tui. Co-Authored-By: Claude Fable 5.2 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNBECMWcJD5pdS4SR6T
| *-. integration: the plumber whole and the walker's cheaper compile, with the ↵Gabriel Schneider41 hours
| |\ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | terminal behind src/Term.zig A merge of pardes-next, plumber and walk-fix. Only the changelog conflicted: both sides' entries kept, and the first plumber entry's "A look does not ask it yet" dropped, a look asking it now. core-test 990/990 on tui; behaviour-test 269/273 (the four addon ABI tests the base fails); the look, diff, pager and tree process cases pass, goldens unchanged; replicas pass. Co-Authored-By: Claude Fable 5.2 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNBECMWcJD5pdS4SR6T
| | * | plumber: the pager sends (docs/next-plumber.md, phase 4): a +Pager filled ↵Gabriel Schneider41 hours
| | | | | | | | | | | | | | | | with a diff (File.looksLikeDiff, pager_diff) is a plumbing message, as acme's `plumb -i` sends what it shows: fillPager's last call is Plumb.paged, which sends src pager, dst edit, wdir the +Pager's directory, type text/x-diff, `action=showdata filename=<the +Pager's name>` (quoted as mesg.c quotes), the text as data; the default rules' last set makes it showdiff (attr delete action, attr add action=showdiff, plumb to edit), and the edit port's showdata and showdiff (Plumb.show, acme's plumbshow) show the data in the pane filename names, the sender's own first, else a new clean pane of that name, showdiff then restacking it in a +Diff in its place (hunks.paged, the code the pager's hook called, unchanged); text paged that is no diff is not sent, where the record had every page sent: the look rules would read it whole, `type is text` holding (a `matches` over a MiB of output, up to the regular expressions' 300 ms budget, at each page), and a page of one line naming a file would ask the disk; the rules in force are read once and kept until their text changes (Plumb.parsed, by its hash, derived state: each process's own), measured in a fast build per send: 29k instructions to read the default rules, a URL's send 37k before, 8.3k after, a word's look 69k before, 40k after, a look that opens a file 262k before (its send 63k), 233k after (34k); test: a diff paged is stacked by the default rule, text that is no diff stays a +Pager, without the rule the +Pager stays as text, and another sender's showdata and showdiff; the design record gains what phases 2-4 did otherwise (detached starts with no output kept, only diffs paged sent, own ports asking nobody, the rules read once); plumber.zig's defaults end with the paged rule; the reference, the default rules' file and the changelog say so
| | * | plumber: external ports (docs/next-plumber.md, phase 3): every port a rule ↵Gabriel Schneider41 hours
| | | | | | | | | | | | | | | | names has a file, plumb/<port> (0444), listed after rules and send in the order rules named them and never dropped (Plumb.ports, dumped: seeded at boot from the rules in force, grown by each apply, refused past 32), its node a third family in the tree beside panes and columns (Node.ofPort, Target.port; port_bit is 1 << 40, clear of the host tree's and the archive's high bits, which 1 << 57 hit); an open reads (Open.what.port, an events.Queue of its own, 64 KiB): the messages sent from then on, packed, one a read, a short read the front and the next the rest, a read with none held until one comes (p.fs.news, answered by the listener as event reads are); every open gets its copy (fanOut, plumber(4)'s queuesend), one over the queue is not sent to it and the log says so; a port pardes reads itself (edit, web, a loaded kind) hands its readers copies without asking anyone, so a look stays speculated whole; another port's readers are the authority's to know, asked through tape.authority (a replica's own look waits for the first answer and guesses the last), with none its rule set's start runs, or its client, the message held in p.fs.plumb_held for the port's first opener (Plumb.opened, plumber(4)'s queueheld); test: plumber.zig, two opens of plumb/edit each read a look's message while the file opens, an open made after reads nothing, a client's message reaches the first opener and the next goes to the reader with nothing started, a port is not written; monkey9p.zig gains the plumb/ files, rules and messages that start no program, the ports' reads as held reads, and plumb/rules' rest read at the close (whose refusal is the log's alone, as a line file's: restAtClose); replicas.zig: a look to a port a program reads, the second of a replica's own speculated whole and confirmed, digests equal; tree.zig's readdir test lists web and edit; the reference's tree and plumber section, fs-help and the changelog say so
| | * | plumber: a look sends what it looks at (docs/next-plumber.md, phase 2): ↵Gabriel Schneider41 hours
| | |/ | | | | | | | | | after the pane's own readings and once a name already open in a pane is focused (never sent), lookAt sends the trimmed name, type text, src pardes (term from a raw terminal, a kind's name from its pane), wdir the pane's directory, with what resolveLookTarget found of it (Plumb.Seen: isfile and isdir take it, so no second stat and Look's own readings of a URL or `//` stay no file); a message the rules give back to `edit` as sent is Look's own (Sent.look) and runs lookAt's switch on that target, one nothing takes or a started word declines is Look's own too, whose `.url` arm is now the word's search; the edit port's open of another message runs with look_text, which now means "unplumbed", so the edit port's lookAt never sends again (the recursion trap); the three diff hooks become the default rules of the record's §5.2 (a bare .diff/.patch to `Diff $file`, a commit hash in the editor to `Diff $0`, after the two edit sets), and `Diff` takes a file or a revision, a rule's Diff declining in silence (Pardes.plumb_declined) where it cannot (no hunk, not readable, no .git above): hunks.lookedAt and lookedAtCommit go, their bodies in hunks.word, the `needs` check of the bare form in the word, PaneKind.diff gone; the old one-line `Plumb` word, Core.plumb, look.Port, plumbRuleError, plumbTarget and resolve's rules block go: bare, `Plumb` opens <config>/plumbing as Config opens init (both through builtins.openConfigFile; `include basic` in a new one), an argument is refused in words, Save applies it (host_io.writeFile beside reapplyInitFile, Plumb.saved), and boot reads it (Plumb.boot in applyStartupConfig); a started program now runs detached as Plan 9's plumber's (the coordinator's decision for the user's question 4): Effect.start, every host's host_io.startDetached (double fork, setsid, /dev/null for 0-2, every other fd closed, argv exec'd in wdir with no shell, the child's errno back through a close-on-exec pipe, said as `plumb start: <prog>: <why>`), so exec_dir goes; resolveLookTarget's and Plumb.Send.search's directory walk is one look.Dirs; test: look.zig's Plumb test rewritten onto the rules (a rule's word, an open pane unsent, the hover's resolve, a URL with no web rule searched for) and a hash inside and outside a git work tree; addons.zig's Plumb test onto plumb/rules (a port no kind reads, its kind gone: Look's own); Plumb.zig: the word, the save and boot; replicas.zig gains a URL (speculated whole, confirmed, no rollback) and a hash, each the authority's and a replica's own; every look golden unchanged; the reference quotes src/plumbing whole and says the order, the changelog says so
* | / term: the terminal a module of its own (docs/next-terminal.md, Phase 1): ↵Gabriel Schneider41 hours
|/ / | | | | | | src/term/term.zig roots it, importing std, ghostty-vt, wuffs, zstbi and plain, never the core, and gives the core one table, `ops` (plain.term.Ops: about 60 entries, the records they take and give and the opaque handles State, Encoder, Track and Replica, in src/plain/term.zig), its own functions put in it through a cast refused at compile time unless each parameter and result matches the table's but for the handles and every error is one the table names; src/Term.zig, the core's side, takes `ops` from the term module when the build's term_options say the core has one (every core module but core-test's), null otherwise, and Term.State is a handle with the walk's hooks, so a dump keeps a terminal as it did (its Parts by name through walk.save, then ghostty's snapshot) and the digest is the emulator's; echo.zig comes back to the core (src/echo.zig: predicted echo is the replica's presentation, its five reads of the emulator through Term.grid, modes, prompt and Term.echoCell); screen_state's Peer is a plain record both sides name (its noteSent a function of screen_state's), its Encoder, Track and Replica handles made and freed through the table (replica.Screens makes its encoder at the first encode); a prompt's cut is told in columns and measured in its row's text by the core (Term.promptPrefix, File.rawAtDisplay); TerminalImages' decode bound moves from memory.limits to the decoder it bounds; build.zig: core_lean, the core module without the terminal, is core-test's, and wireCore no longer gives any core module ghostty-vt, wuffs or zstbi (the term module has them), so core-test's compile names none of ghostty, wuffs, zstbi, simdutf, highway or -lc++ (from --verbose); term-test runs the module's own tests (the pager's, screen_state's, the luminance table's); the tests of a terminal through a core move as they were to test/terminal.zig, in pane-test, which imports both (`of` reads the emulator behind a handle; the dump round trip frees through State.dumpFree), and tests elsewhere that read the emulator do it through Term.emulatorOf, a test in core-test skipping first; core-test skips a test whose terminal Term.open refuses; Authority.init releases its tape when the core it boots fails (it leaked); tests: core-test (no terminal) 409 of 941, 532 skipped, shell-test (the same tests with the terminal) 984 of 984, pane-test 180 of 181 (1 skipped, as before), term-test 15 of 15, kitty-test 45 of 45, replicas checked/wrong 0, hxdiff 658 cases 0 mismatches, behaviour-test 268 of 272 (the base's four addon ABI tests), lib-test
* | integration: the terminal behind src/Term.zig, with the diff viewer and the ↵Gabriel Schneider41 hours
|\| | | | | | | | | | | | | | | | | plumber's first phase A merge of term-phase0 and pardes-next. File.zig's recolorSyntax keeps the theme's grounds and names Term.Filter and Term.Span; the changelog keeps both sides' entries, the palette sentence in the place of the entry the terminal work cut down. core-test 985/985 on tui; the diff, pager and tree process cases pass, goldens unchanged. Co-Authored-By: Claude Fable 5.2 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNBECMWcJD5pdS4SR6T
| * integration: the diff viewer on general mechanisms, and the plumber's first ↵Gabriel Schneider41 hours
| |\ | | | | | | | | | | | | | | | | | | | | | | | | phase, together A merge of diff-lib and plumber-1; the changelog keeps both sides' entries; core-test 984/984 on tui, the process suite's tree cases pass (fs-help.txt holds both the rows and the plumb lines in its 45). Co-Authored-By: Claude Fable 5.2 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNBECMWcJD5pdS4SR6T
| | * plumber: Plan 9's plumber in the core, its language and its files ↵Gabriel Schneider41 hours
| | | | | | | | | | | | | | | | | | | | | (docs/next-plumber.md, phase 1): src/Plumb.zig owns the rules, kept as their text (Core.plumber, digested and dumped; `include basic`'s, the embedded src/plumbing, until any are written, so a session and a dump from before it start with the defaults) and parsed again for each message, and a matcher over them read from rules.c and match.c: the message's pack and unpack (plumbunpackpartial's whole, short of its data, malformed), packed attributes read and quoted rc's way, rule sets of `object verb arg` lines, variables, include (basic, the configuration directory's, an absolute path, through the taped read site), rc's quoting and `$` expansion at parse and at match, every parse refusal in rules.c's words with its file and line, `$1`-`$9` refused in words (the regular expressions have no capture groups), `matches` over regexp.zig whole (`^(...)$`) or around a click counted in runes with Pike's two patterns, `is`, `set` rewriting for good, `add`, `delete` failing on an attribute not there as match.c's code does, `isfile` and `isdir` finding a name as Look does (past its place, the message's directory, a look's jump list's, Incl's) at the taped resolve sites, once a send; dispose to a port pardes reads (edit: opened as a look opens it, at addr=; web: the host's link opener; a loaded kind by its name), to a rule set's start or client run as exec runs a line (a builtin in place, else a command pane in the message's wdir: Core.exec_dir, exec.runCommand's reuse skipping the sender), or Plan 9's words for what went nowhere; rules print back in the order they were written, a variable where it was set, so a printed text reads back the same; plumb/rules (a write adds rule sets as each is whole, the rest at the close, `>` replaces them, one writer at a time) and plumb/send (a message put together across writes) in the tree, with fs-help's line, the reference's section and tree, and the fstree test's paths; nothing calls the plumber from Look yet; test: Plumb.zig's eleven (every parse error, print and read back, plan9port's basic less its submatch rule sets, its ¡-￿ classes UTF-8's bytes, whole and click matches with horse.gift, quoting, a set's rewrite kept, delete of an absent attribute, dst filtering, isfile and isdir, dispose, a dump keeping the rules), test/behaviour/plumber.zig's two through the tree (Session records the link a host would open); the root's listings in tree.zig's and the process tests gain plumb, and the README folds a blank line to stay 45 lines; the changelog says so Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNBECMWcJD5pdS4SR6T
| * | lib: the diff is a language of pardes-syntax: `diff` (Walk, targetAt), ↵Gabriel Schneider41 hours
| | | | | | | | | | | | | | | | | | | | | `highlightDiff`, the style byte's `syn_bits` and grounds and colors.Grounds, a theme's colours for them; test/lib colours a C patch through it (each side's code as C, the rows' grounds, Orchard's, and the place a `+` row names) Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNBECMWcJD5pdS4SR6T
| * | test: a +Diff pinned before it moves: process-test's `diffstack` pages a ↵Gabriel Schneider41 hours
| |/ | | | | | | | | | | | | patch into `pardes -` and snaps the +Diff it is restacked into (its hunk found two lines down where the file moved it, under the declaration it stands in, `a.zig:3` and `const S = struct {`, its label one level deep, `a.zig:5 *`; the changed word on a stronger ground in both lines, a line added whole on a faint one, the markers in their tints' hue), as text and as styles; replicas opens with a .patch looked at by the authority and as the replica's own input, the first stopped at its first read and confirmed with no rollback, the replica's digest and +Diff the authority's either way; the soak's fixture is a repository whose last commit changes code.zig, and `gitshow` pages `git show HEAD` into `pardes -` from a shell Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNBECMWcJD5pdS4SR6T
| *-. integration: the drag guides' removal, the build-time highlight filter and ↵Gabriel Schneider41 hours
| |\ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | the lean-core changes together A merge of drag-guides, syn-buildtime and lean-core; the changelog's three new entries are kept, each whole. Co-Authored-By: Claude Fable 5.2 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNBECMWcJD5pdS4SR6T
| | | * mouse: no drag guides, and columns resize by their grips aloneGabriel Schneider2 days
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | A drag draws no guide lines: the dashes where a pane or column would land, the carried column's rail and the held seam's dashes are gone, with the guide region kind, Surface.overlayDash, the rail's Region.clip and MAX_GUIDES; the gui's floating draw group is the debug box's alone. The line between two columns is no resize handle: Drag.border_v and its corner variant are gone, with clampBorderCol and the unused chromeTarget. A press there is a press in the pane under it (a sweep). A column resizes by its grip, a pane by its grip or, in a window, the rule above it (border_h, unchanged). Tests: the corner and border-drag unit tests and the cornerdrag and cornerdrag-left tty cases are deleted; windowops loses its border drag and tagalign2col widens its column by the grip; colgrip, shrinktag and windowops goldens drop the guides; a unit test pins that a press on the line between columns sweeps and resizes nothing. The cheatsheet, themes, render-pipeline notes and changelog say so. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | | Term: the replica half and the seam's last reaches: screen_state.zig's ↵Gabriel Schneider41 hours
| | | | | | | | | | | | | | | | terminal half is reached through the interface (Term.encode, Term.encodeRows and Term.untrack on the authority, Term.wasRead and Term.clearRead for the rows a step read, Term.apply with the prompt marks riding a delta, Term.applyRows, Term.shown and Term.blank on a replica, Term.Encoder, Track, Peer and Replica its handles), so replica.zig reaches no emulator state; its kind half (a kind's frame as a Shadow, its strips and Blobs, encodeShadow) is the core's and moves, as it is, with its tests, to Kind.zig beside the Slot it shadows; predicted echo is Term.Echo; the draw entries (Term.recolor, Term.images, Term.petscii, Term.togglePetscii, Term.hasImages, Term.Filter, Term.FilterPalette, Term.palColor) are reached through the interface with their signatures as they were, not yet plain data (Term.zig says so); src/Term.zig is the one file outside src/term/ that imports one inside it (fs.zig embeds terminal.zig's text as a served source), panes.terminal, pardes.screen_state and pardes.echo go, the comments that named ghostty for the emulator say the emulator, and the tests that read emulator state read the records but three (Core.zig's palette, pty.zig's output pin and page limit); the changelog says so; nothing else changes
* | | | Term: text through the interface, the editing of the copy back to the core: ↵Gabriel Schneider41 hours
| | | | | | | | | | | | | | | | Term.scrollback and Term.scrollbackLen are its whole history as plain text (its body, Save), Term.textRows its rows as text with a prompt's own cells cut off (a plain Term.Rows), Term.viewport the rows its view shows, Term.promptPrefix how many bytes of a row are its prompt and Term.markRead says a step read them; the editing of the copy of a terminal's rows is the core's again, in edit.zig beside the text it edits: shellRows and the rows' cache (edit.RowsCache, edit.drawing), terminalLines (the rows with the copy spliced in), and in edit.zig's own flatSurface, editText, setEditText, pushUndo and undoStep the terminal's arms written out where they were dispatched to terminal.zig; the body's walk over the rows and the copy (BodyWalk) is body_layer.zig's, beside the bodyText it makes; screenTextAlloc, screenTextLen, cursorLines, flatSurface, editText, setEditText, pushUndo, undoStep and bodyText leave terminal.zig; nothing else changes
* | | | Term: prompt and commands through the interface: Term.prompt is the plain ↵Gabriel Schneider41 hours
| | | | | | | | | | | | | | | | record of where the shell is in its command cycle (its phase, commands started, prompts drawn, lines declined, a continuation, the last command to finish and its status, input ready, input empty, the cursor at a prompt), Term.commandOutput what the newest command printed and Term.endOutputHere ends a command's output at the cursor; the pending command queue is the core's (Pane.pending_command, Pane.queuePendingCommand and Pane.releasePendingCommand, released at the shell's start or its prompt where Core.zig hears of them), and so are Pane.greeting, Pane.noteCommand, Kill's "the command pardes typed runs" (builtins.zig) and the newlines that keep a command's output below the edit copy (exec.zig); /pty/status, /pty/run and predicted echo read the record; promptInputReady, promptInputEmpty, commandRunning, armShellSpawn, shellSpawned, deinitPendingCommand and padOutputBelowEdits go; nothing else changes
* | | | Term: input through the interface: Term.key writes a key's bytes as its ↵Gabriel Schneider41 hours
| | | | | | | | | | | | | | | | program's keyboard protocol and modes ask, Term.mouse and Term.wheel a mouse event's and a wheel turn's, Term.paste a paste as xterm makes one, Term.modes is the plain record of the modes the core decides by (mouse, all motion, the wheel, focus, bracketed paste, the alternate screen) and Term.promptMoves how far a click on the prompt line moves the shell's cursor; the mouse's records (MouseMods, MouseButton, MouseAction) are the interface's; keys.forwardKey (the bytes queued, an error reported) and keys.enterTty (raw mode, the cursor handed to the shell) are the core's, in keys.zig, pardes.keys for the tests; bracketedPaste, reportsMouse, reportsFocus, ownsWheel, tracksAllMotion and onAlternateScreen go into the record; the tests that called them follow; nothing else changes
* | | | term: the terminal's own files move into src/term/ (terminal.zig, ↵Gabriel Schneider41 hours
|/ / / | | | | | | | | | screen_state.zig, echo.zig, TerminalImages.zig), as they are: names and signatures kept, only the import paths change (theirs to ../, pardes.zig's, panes.zig's, tape.zig's, replica.zig's and Kind.zig's to term/), the comments that name them by path follow, and the embedded sources serve terminal.zig at src/term/terminal.zig; the first step of docs/next-terminal.md's Phase 0, a move, nothing else
* / / text: a terminal's edit overlay and its undo are the Text's: EditBuffer, ↵Gabriel Schneider41 hours
|/ / | | | | | | Snapshot and history_max move from terminal.zig into Text.zig, as they are, so Text names no terminal type (terminal.zig and panes.zig's ovl name them there, as test/panes.zig does); a move, nothing else
* / addons: an addon can hand the frontend a GPU scene (ABI 3.1), and a snake ↵Gabriel Schneider41 hours
|/ | | | | | | | | | | | | | drawn by shaders does; a commit's patch keeps its own +Diff pane; a real ZLS checked through the window build ABI 3.1 appends an optional `gpu` callback and its context to pa_paint: one self-contained sketchbook SGPU v1 scene for the pane, at most 4 MiB, copied during the call, the last successful call winning. The SDL frontend runs it on its own GPU through sketchbook's executor and composites its target texture into the pane; replicas carry the scene beside the authoritative pane frame; cells and rasters stay the fallback for terminals, older hosts and frontends without an executor. addons/snake_shader is such an addon (`zig build snake-shader`, the SnakeShader word: a board that follows its pane, buffered turns, a CRT look), with test/gpu_snake.py, test/gpu_snake_interaction.py and test/gpu_upgrade.py. Commit patches open as `+Diff-<short-hash>` and reuse only their own commit's pane, so another paged patch no longer replaces one already open. The language-server client looks in the standard per-user bin directory, which desktop launchers often leave out of PATH, and takes an empty answer that beats the startup progress notification onto the wire; test/real_zls.py checks imports, hover, references, rename previews, diagnostics and completion against a real ZLS from a private home. The dump's schema reader skips two fields older builds persisted, and a replica's shadow of a kind ticks when the authority's does. The reference, the guide, the building and setup notes and the changelog say so. The working copy as it stood on 2026-11-12; message written by Claude from the diff, at the author's request.
* zstbi: upstream's main, the fork pinned before being gone: Pivok7/zstbi ↵Gabriel Schneider7 days
| | | | (zig-gamedev/zstbi PR #16, the Zig 0.17 port the manifest waited on) is deleted and a fresh checkout could not fetch it, while upstream has since ported itself (6b9bea1f); its init takes no Io and can fail, and its resize can fail, so TerminalImages.start takes the allocator alone (a start refused leaves PNGs refused, as before one) and a resize that fails is an undecodable PNG; host_io.startWorkers and tape.replayMain lose the Io they only handed on; build.zig's patch of zstbi.c applies unchanged (the file is the same); test: kitty: a PNG longer than 1280 pixels on a side is kept at 1280 (the one path through zstbi, which no case walked: a 1300-pixel one-bit PNG built in the test, its row stored); kitty-test 45 of 45; the changelog says so
* lib: pardes's syntax is a library, as ghostty-vt is ghostty's: a package ↵Gabriel Schneider9 days
| | | | that depends on pardes is given build.zig's `library` and nothing of the editor, one module, `pardes-syntax` (src/lib_syntax.zig), naming what a program that draws source text itself takes, the editor's own code under its own names (highlightFileRange and analyzeSource for a Syn a byte and the sticky headers' declarations, parse and walkNodes over the tree-sitter bindings, theme, page, ink and synInk for a named theme's colours, the line arithmetic, pathFor for a language's name); every dependency but tree-sitter's runtime is lazy in the manifest, so such a package fetches the runtime and the grammars of its tree-sitter set alone, and pardes's own build asks for all it is missing in one round before its first b.dependency; build.zig's tree-sitter and theme-generator blocks become treeSitter and generatedThemes for both to call, a grammar not yet fetched asked for and left out of that round, and the scanners' -include of src/tree_sitter_heap.h is spelled from the root package's directory, where the compiler runs (pardes's own flags are as they were); File.synStyle's switch moves to colors.synInk, so the editor and the library colour a class from one place; test/lib is such a package, depending on pardes by path: it colours, outlines and walks a C and a Zig source in orchard's colours under the testing allocator, and `zig build lib-test`, which unit-test runs, is its own `zig build test` over this checkout's zig-pkg; ts_markdown is pinned by the commit its hash is of (a0a00f81): its split_parser branch moved to v0.5.4 and a fresh fetch no longer matched; building.typ and the changelog say so; syntax snapshots unchanged
* history: each dump is the command that brings it back, `Restore <label>` on ↵Gabriel Schneider9 days
| | | | | | | a line of its own, its date, size, pane count and names indented on the line under it, so the line is a whole command (a sweep of it, a click on its Restore, a look at it) and the details are read, not parsed; the builtins JSON, the reference and the changelog say so Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* test: Session.initShared's startup config lives in the session's arena, not ↵Gabriel Schneider9 days
| | | | | | | the constructor's frame: the core keeps the slice and a Restore applies it again, which read freed memory (an `init file line 1: unknown word` of 0xAA bytes in the log after every Restore in the suite) Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* history: a +History row reads `Restore <label>` before its date, size and ↵Gabriel Schneider9 days
| | | | | | | panes, and a click on that Restore (an exec, or a look) is a Restore of that row's dump, exec.operandAt expanding the word to the label beside it as the row reads, where a bare Restore would have been this session's last dump; a look of the whole row reads past the word; the behaviour test clicks a row's Restore and looks at a row Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* layout: exec.placeNew moves into layout.zig as layout.placeNew, with ↵Gabriel Schneider9 days
| | | | | | | NewPane, the type only it switches on, and acmeColumn and columnOf, which only it calls, before canSplitCol, since 29 of its 41 references are layout's own (acmePlace, takeRoom, splitTallest, placedRoomy); callers name layout, and Dir, mini and Store, which imported exec for it alone, import layout instead Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* hxdiff: a case may name its file pane ("path", whose extension picks the ↵Gabriel Schneider9 days
| | | | | | | grammar), and test/hxcases/indent.jsonl pins the new line's indent in a Zig, a Python, a Rust and a plain-text file against the reference helix with its grammars, typed into a real hx in a pty by hx_indent.py since the harness runs without any; zig build hxdiff runs it before the suite Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* edit: Enter, o and O give the new line helix's indent where the file's ↵Gabriel Schneider9 days
| | | | | | | grammar has an indent query, its indent_for_newline over the file's tree (indent.zig: the reference's containment engine, every capture and predicate the vendored queries use, the hybrid heuristic), else acme's copy as before; and Enter between a pair puts the closer on a line of its own at that indent and the cursor on one a level in, as helix's insert_newline does, in plain text too Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* macos: the AppKit trackpad adapter for the SDL window (upstream a41360e3's ↵Gabriel Schneider9 days
| | | | | | | Trackpad.zig, macos_trackpad.{h,m} and their test as they were), its gui.zig hooks on today's names (pointAt, view().jumps and active, one setFontPx for a key's step and a pinch) and the build's shim per gui module with macos-trackpad-test registered once after the loop, all gated to macOS so a Linux touchpad's pinch and the goldens stay as they were Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* soak: instant counts an input at once when it shows no later than the local ↵Gabriel Schneider9 days
| | | | | | | session's own p99 for its class as well as half a round trip: a tty's wheel glides, its first row two frames on in a local session too (36 ms), which the 33 ms floor at 30 ms called waiting Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* soak: the session keeps a PDF pane open (--pdf, design.pdf by default, in a ↵Gabriel Schneider9 days
| | | | | | | column of its own at the start and Looked at again at a quiet point if a Del closed it), so every cell carries a kind's frames and raster strips to a client without the library, and instant's pointer classes stay in the first column, where the file and the shell are Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* a session's end leaves a slim dump and History lists them, the tags gain ↵Gabriel Schneider9 days
| | | | | | | Follow, History and the window's Fonts: every way out (Exit, the last pane's Del, the host's quit) runs Pardes.endSession, which puts a dump of the session in the state file under the usual label before quitting, slim by the dump's own means (dumpJoin's slim_terminals, and Store's slim-file swap now a Slim struct: no undo, a clean file's text read from its file again by revivePane, a terminal a size for a new shell), none for a session of shells alone; the dump's image carries the panes' names, which /dumps rows end with; the History builtin (SPC f h) lists the state file's dumps in a +History pane and a look at a row is a Restore of its label; the default tags: Follow after Collapse on text panes, History before Debug on the workspace row, Fonts after Themes in a window's view alone (tagline.topbarDefault picks the row per view; both rows are known defaults to the docs and tutor tests); the tty goldens re-recorded (the new words; a narrow pane's tag wraps to a second row, so its body rows shift), builtins JSON regenerated Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* idioms: the edit surface fails by name: the undo history (Journal.mark, ↵Gabriel Schneider10 days
| | | | | | | Text.remember and step, File's, terminal's and edit's pushUndo and undoStep) returns error{OutOfMemory} instead of dropping an undo point, setEditText, enterInsert and insertTab return edit.Error with try where they reported each failure, Registers.put and setYank return error{OutOfMemory} instead of losing a yank, and their other-file callers decide (try under edit.Error, said on the pane by Undo, Redo, Pipe, Edit, Query, keys, look, completion, Incl and Plumb, ENOMEM or null where a 9P write answers, Ask's false); ctl.get and getFile return GetError (fs.ReadError, now named at fs.zig's surface, and InvalidName) for the *anyerror out-parameter and errno, the 9P get, Get and Edit's e mapping it as before; the OOM undo test pins the error; perf.zig's try relabels its harness hash, the three baselines re-keyed to it (their numbers the base's) Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tests: at the end of their files: every top-level test block of the 46 files ↵Gabriel Schneider10 days
| | | | | | | 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
* tests: the gui goldens go (test/gui-goldens.txt, the gui-golden step, ↵Gabriel Schneider10 days
| | | | | | | window.zig's scenes case): their 15 frame hashes were this machine's; what they caught the kept set catches, checked break by break: the core's primary selection drawn as a secondary one (6 frames) fails 12 tty cases, the grip's focus fill and its button a pixel narrow (all 15) fail "every pixel of a grip column is band, button or ring", the theme's rules a pixel wider (all 15) fail "an empty column is ruled on both sides" and "decor is read from the regions"; two caught by the frames alone get unit tests of their own in gui.zig, the scrollbar thumb a pixel narrower (all 15) and the block cursor not reversed (9 frames); breaks the frames never saw (the body's clip and its context border, InactiveDim's fade, an image's scissor, fit and placed rect, the palette, the glyph art) stay the unit tests' and the image harness's or nobody's, as before Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tests: monkey and replay are modes of the T1 runner and snapshot.zig goes: ↵Gabriel Schneider10 days
| | | | | | | behaviour-test runs test/monkey9p.zig bare (its smoke), `-- --monkey <seeds>` drives an in-process session with test/monkey.zig's input (keys, leader bursts, the mouse, tiny resizes, over /kbdin, /mousein and Session.resize) hunting panics and hangs, `-- --replay` records each seed's sessions as a tape in memory and replays it (tape.replayBytes, replayMain's loop: three replicas and the dump), and `process-test -- --replay` records and replays the tty cases' real sessions as `zig build replay` did (with the addons it lacked, so images passes); the snap, snap-driver-test, monkey, monkey-test (it no longer compiled), replay and monkey-session steps, pardes-snap and build/snap.zig (now build/process.zig) go; checked: a counter outside the core bumped at the 40th update diverges every T1 seed and 89 of 103 old tapes, one bumped on terminal output diverges no T1 seed but 69 tty cases and 70 old tapes (why the tty replay stays), a panic on a window under 5x5 is found by the T1 monkey (seeds 1 8 17 20) and the old one (seed 13) Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tests: the terminal snapshot scripts (19: tty ttyclick ttyedit ttyfollow ↵Gabriel Schneider10 days
| | | | | | | ttyhelp ttylook ttyonly ttypaste ttyprompt ttytaken psearch reflow respawn scroll cmdexit look-fd-deep ghostty-mid exec exec-bigdrag) are process cases, test/process/tty_terminals.zig, their goldens unchanged; checked against the old runner by Ctrl-b no longer leaving tty mode (the same 12 fail in both) Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tests: the image snapshot script (images) is a process case, ↵Gabriel Schneider10 days
| | | | | | | test/process/tty_images.zig, its golden unchanged; checked against the old runner by the glyph art's ink and ground swapped (it fails in both) Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tests: the theme snapshot scripts (6: theme themesel syntax markup ↵Gabriel Schneider10 days
| | | | | | | pagercolor cursorline) are process cases, test/process/tty_themes.zig, their goldens unchanged; checked against the old runner by the tty dropping bold (the same 5 fail in both) Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tests: the columns and layout snapshot scripts (14: colgrip collapse ↵Gabriel Schneider10 days
| | | | | | | cornerdrag cornerdrag-left delask emptycol layout-close layout-open shrinktag tagalign tagalign2col windownav windowops wheeldrift) are process cases, test/process/tty_layout.zig, their goldens unchanged; checked against the old runner by the tty reading a drag as a motion (the same 7 fail in both); three core breaks (splitBelow at a third, minRows one more, a dropped column's width to its right) fail neither Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tests: the editing snapshot scripts (41: modes, selections, look, search, ↵Gabriel Schneider10 days
| | | | | | | the language client, diffs, tags) are process cases, test/process/tty_editing.zig, their goldens unchanged; checked against the old runner by two breaks, look's centring a third of the pane down (the same 4 fail in both) and the tty reading a drag as a motion (the same 7), grep the one flake in both, cured by the run alone Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tests: the boot, argv and builtins snapshot scripts (22: argv badutf boot ↵Gabriel Schneider10 days
| | | | | | | builtins dump load restore headersweep headersweep-restored logquiet tutor topbarexec leader mini ninep nested nested-optout shellset msg winch savepane filewatch) are process cases, test/process/tty_boot.zig, each its script's steps as Tty calls and its golden unchanged as its expectation; checked against the old runner by two breaks, the message stamp's spacing (3 fail in both) and the tty reading a middle press as a right one (the same 9 in both) Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tests: the tty's own cases get their runner in the process suite, ↵Gabriel Schneider10 days
| | | | | | | test/tty.zig (a pardes in a pty of its own, driven by a terminal's bytes, its captures compared with test/snapshots/<name>.golden: the snapshot scripts' commands as methods, their capture and settle code moved from snapshot.zig) and process_test.zig's pool (a child a case, a core's worth at once, a failure run again alone with the wider settle probe), with six empty files for the script families the next changes fill, so each family can be reverted alone Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tests: hxparity is hxdiff's --parity mode, one driver: the hxparity step ↵Gabriel Schneider10 days
| | | | | | | goes, `zig build hxdiff -- --parity` runs the pty run against the file run as the step did (738 cases, 0 mismatches, 12 waived, its report still named hxparity), and the docs that named the step say so Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* helix: %S\s+ keeping the line's newline in its last piece is a documented ↵Gabriel Schneider10 days
| | | | | | | pardes choice, waived in words: helix's split_on_matches (helix-core/src/selection.rs:852-880) runs the Rust regex over the whole range, whose \s matches the newline at 14, so the last piece ends at 13, where pardes searches as sam does, line by line (reference.typ, Regular expressions; src/regexp.zig), and \s never reaches a newline unless the pattern names \n (S\s+|\n gives helix's pieces, checked against hx-harness); waivers.jsonl names split-regex-ws and split-regex-ws-tty (hxdiff 2 to 0), reference.typ adds \s to what never matches a newline, and helix-keys.md has its row Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* helix: o on a line indented by less than a unit is a documented pardes ↵Gabriel Schneider10 days
| | | | | | | choice, waived in words: helix's no-language indent_for_newline (helix-core/src/indent.rs:1361-1362, indent_level_for_line at 200-211) opens the line with whole units only, rounding two spaces down to none (and six to four, Enter the same, checked with hx-harness), where Enter, o and O keep the indent as it is, acme's autoindent (reference.typ's Indent row, change tlwwmyym); waivers.jsonl names o-partial-indent-drops (hxdiff 3 to 2), and helix-keys.md's o, Enter and waiver rows say it Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge