summaryrefslogtreecommitdiff
path: root/src
Commit message (Collapse)AuthorAge
* vaxis_input: the core's cell styles as vaxis draws them (style, color, the ↵Gabriel Schneider2026-11-05
| | | | | | | underline mapped by its name) and a pasted key's text (pasteBytes) are vaxis_input's, the tui's copies gone; the gui's test-mode stdin takes a paste's keys through the same pasteBytes, which also knows the keypad's Enter (its own copy did not) Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tui: a Screen makes, drops and destroys its own heap copy (Screen.create, ↵Gabriel Schneider2026-11-05
| | | | | | | drop, destroy) where attachTerminal and resumeTerminals each built one inline and dropScreen repeated freeScreen; a Restore closes each pty through closePty instead of a copy of it Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tui: helpers with one caller are their caller's lines (copyToClipboard, ↵Gabriel Schneider2026-11-05
| | | | | | | requestClipboard, paintCursor, updateCoreTerminalSize), and kittyPlacement returns at once in a build without PDFs instead of wrapping its whole body in the comptime check Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tagline: a tag and a header take the keyboard through one takeKeyboard (the ↵Gabriel Schneider2026-11-05
| | | | | | | cursor pulled back inside the text, nothing selected or pending, normal mode) where enterTag and enterHeader each ended in their own copy of it Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* Presentation: a cell is in a box by one test, animation.Box.contains, now ↵Gabriel Schneider2026-11-05
| | | | | | | the centre test its comment always said (its one other caller passes whole-cell pane boxes, where centre and corner agree), so boxContainsCell goes; the aliases used once are spelled where they are used Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* pardes: TAG_GAP goes, a knob with one value (0): TAG_TEXT_INSET is the ↵Gabriel Schneider2026-11-05
| | | | | | | gutter, the tag's zero-width gap fill and its two offsets in draw and mouse go, and the test blocks behind `TAG_GAP > 0` that could never run Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* draw: parameters nobody reads go (renderBody's prebuilt text, null at every ↵Gabriel Schneider2026-11-05
| | | | | | | call; paintPaneTag's capacity; renderHeaderLayer's and paintHeaders' arena), a pane tag's look-hover preview is painted once where it was painted twice over the same cells, the column grip's ink is inkOn's instead of a copy of its sum, boxBorder, inactiveDim, paintPaneTag and paintHeaders are draw's own, and animation's two switches with same-result arms keep one arm each Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* panel_compositor: a slide and a zoom draw through transformClipped with the ↵Gabriel Schneider2026-11-05
| | | | | | | screen for their clip, so the near-twin transform goes (it sampled track.to, which is contentBox for every track but a closing one, and only a vertical one closes); the tests fill their cells with one testGlyphs call where each had its own five-line loop Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* colors: the chrome's fade is one concrete ChromeAnimation over ChromeTheme, ↵Gabriel Schneider2026-11-05
| | | | | | | not animation.Fade and animation.Immediate (generics each instantiated once, picked by a comptime switch the runtime animate_theme_changes flag already decides: the board snaps because finishThemeInitialization sets it from theme_animation); the fade mixes straight to the thousandth it always stepped in, without the integer round trip the generic needed, and its tests run over ChromeTheme values instead of a test-only TestColor Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tui: the terminal-input tests share one TestInput fixture (vaxis, its ↵Gabriel Schneider2026-11-05
| | | | | | | writer, the queue and a feed, set up in place) and one TestReader (parts a read each, a gap, asked or not) where each test had its own twelve lines of setup and its own copy of a reader Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* tui: the loop's queue carries a named Msg union, not the type of a global ↵Gabriel Schneider2026-11-05
| | | | | | | var declared only to be named (`@TypeOf(Command.value)`), and the Windows branches go (startInput's `loop.start()` and seven test skips: the tty cannot build for Windows, it calls sigprocmask, sigwait and shm_open) Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* drawing: dead code goes: colors.carryTheme and Motion.speed (no caller ↵Gabriel Schneider2026-11-05
| | | | | | | anywhere), animation.progress (only its own tests called it), Layer.pointAt (never called), Glide.kick (one caller, the line now), a stale first column header in the frame dump, panes.zig's unused vaxis import Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* pdf_view.applyNormal: the pane's PDF state is the pointer the slot holds ↵Gabriel Schneider2026-11-05
| | | | | | | (the kinds line's `pdf: ?*State`), not a pointer to the slot: the editing line's one key layer rewrote the call on the older inline State, and the two met here Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* modal, edit, normal: the selection ops splice their primary range's offsets ↵Gabriel Schneider2026-11-05
| | | | | | | (Text.primaryRange): d, c, the selection's text, r, R and the chords' replace, ~ ` Alt-`, C-a and C-x, ms, md and insert-mode Delete; selRange and the cursor-pair half of modal's row/col edit layer go (rangeBytes, rangeText, deleteRange, replaceRange, replaceChars, deleteChar, clearLine, lineFrom), with their unit tests, and deleteLines no longer builds the removed text no caller read; a key-level test pins r and Delete on whole graphemes; the insert-mode and paste ops keep insertAt, deleteSpan, pasteLineBelow and advanceBy, whose callers hold cells; no key does anything new Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* Text, edit, normal, look, Query and their readers: the second selection form ↵Gabriel Schneider2026-11-05
| | | | | | | msel (whole rows r0..r1, for search-result highlights) goes; a visited search result and a query preview's place line are an explicit vsel like any other, back from the line's end (to the result's location start, the line's start for ordinary path:line rows; to the start for a preview line), and every msel branch in the edit ops, the drawing, the terminal's position tracking and the chords goes (visible: the result's highlight covers its line's characters and newline cell instead of the whole row width, with the ordinary selection colours; y, d and c on it act on that range as after x, its newline included, so a yanked result pastes as a line); selections.md and the ; Alt-; X Alt-x rows of helix-keys.md no longer describe msel; Text's walked shape changes (wire semantics) Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* modal, normal, pdf_view, Text, syntax, edit: one key layer: an Action is the ↵Gabriel Schneider2026-11-05
| | | | | | | role the key named and what it was given (a count, a literal, ] or [), not a second union of sixteen sub-enums the roles were mapped onto one if at a time; a context's roles are tried in their enum order, the pending prefix and m sub-key are roles too, the text's executor and the PDF's switch flat on the role, and pdf_view.applyNormal(core, pane, action) does what a page cannot (Last, the leader, the tag, a search) instead of handing a seven-value enum back; no key does anything new; Text.normal's walked shape changes (wire semantics); helix-keys.md says how the keys are read Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* File: a +Pager's text is told a diff by diff.looksLikeDiff, the one ↵Gabriel Schneider2026-11-05
| | | | | | | heuristic command panes use (a diff --git line, or a ---/+++ pair over a valid @@ hunk, in the first 200 lines), not by a second one of its own (visible: a pager text with a diff line but no unified hunk, or ---/+++/@@ lines apart, is no longer shown as a diff) Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* File, edit: dropped marks go: an edit frees the location rows and pager ↵Gabriel Schneider2026-11-05
| | | | | | | colours it drops, and an undo back to the same text no longer brings them back (visible: a +Pager or a results buffer undone to its text is shown plain); the pager test in ninep/ctl says so and Stream's no longer asks for the field; File.State's walked shape changes (wire semantics) Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* pardes, normal, edit: multi_on goes: outside a replay the replay state means ↵Gabriel Schneider2026-11-05
| | | | | | | one range, its first pass (index 0, count 1, first true), so setYank, registerValue, pushUndo and the open-lines and Tab checks read it plainly; Pardes's walked shape changes, so a client and a daemon from either side refuse each other (wire semantics) Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* regexp, sam_edit, ninep/addr: a pattern's compile error is told by ↵Gabriel Schneider2026-11-05
| | | | | | | Regex.message(err), which replaces the five e_ constants and the two switches that mapped errors onto them Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* modal, normal, Text, config: goto_view_center goes: its chord list was empty ↵Gabriel Schneider2026-11-05
| | | | | | | (gc runs Callstack, its hx case waived), so the role, the Goto value, its arm and gotoWindow's center never ran; config.match_bracket, read by nothing, goes too; helix-keys.md's gt/gc/gb row says what gc is Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* File, terminal, edit: undo and redo are one undoStep(back) each, as ↵Gabriel Schneider2026-11-05
| | | | | | | Text.step already is, and their callers (normal, builtins, edit_cmd, file_watch, ninep/ctl's pager test, test/panes) say which way Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* File: recolorSyntax paints a grapheme's cells in one loop (syntax colour, a ↵Gabriel Schneider2026-11-05
| | | | | | | diff's emphasis ground, a result's emphasis, a pager's colours, in that order) instead of four loops over the same cells Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* Text, normal: verticalMove and visualMove are one verticalMove(width): width ↵Gabriel Schneider2026-11-05
| | | | | | | 0 (j/k, Ctrl-d/u) moves by lines in one step, a wrap width (gj/gk) row by row, and the rows are read through the file's line index Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* edit: normalDelete's bare cursor is the 1-wide selection on its cell and ↵Gabriel Schneider2026-11-05
| | | | | | | goes through the selection's path, whose cursor is clamped into the text after the delete (a d of a text's last newline left it one row past the end) Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* edit: LineAt, lineCountOf, rowOf, lineAt and indexedLine, copies of File's ↵Gabriel Schneider2026-11-05
| | | | | | | indexed text functions, go: typing, backspace, J and editTextEol ask File.textLine, textLineStart, textLineCount and textPosition Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* edit, Text: the four one-edit multi-range ops (J, Alt-(, &, Ctrl-c) widen ↵Gabriel Schneider2026-11-05
| | | | | | | their rows with allSelRows and end with setEditCells, which puts the text in and selects the ranges' cells, instead of four copies of both Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* syntax: highlightLocationRows takes []const locations.Row instead of anytype ↵Gabriel Schneider2026-11-05
| | | | | | | tested with @hasField, its tests build Rows, the diff window is a named DiffWindow paintDiffPiece takes by type, and the unused look import goes Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* selection_pipe, Pipe, edit_cmd: a pipe's inputs are plain byte slices, not a ↵Gabriel Schneider2026-11-05
| | | | | | | one-field Input struct, dupeAll takes them concretely (no anytype, no length argument), and an Edit's request leaves the single command and cwd empty, which runJob never reads when every input has its own Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* normal, modal, File, edit: small ones: tryFind inlined (its clamp never ↵Gabriel Schneider2026-11-05
| | | | | | | bound), the smart-case haystack is allocLowerString, the two WordTargets nothing produces go, File.lineCount is modal.cursorLineCount's alias, and pasteRegister and ClipRequest share one Paste enum Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* modal: kindOfCodepoint answers the WordClass wordClassAt returns, and the ↵Gabriel Schneider2026-11-05
| | | | | | | Kind enum and the switch between the two go Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* edit: setClipboard, with one caller, goes: clipYank names the + register ↵Gabriel Schneider2026-11-05
| | | | | | | first and its PDF branch yanks into it Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* selection_pipe: runOne, a wrapper only its tests called, goes; the tests ↵Gabriel Schneider2026-11-05
| | | | | | | call runIn, which is private now and carries the doc comment Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* edit: normalYank's tail is selText's: the selection, else msel's rows, else ↵Gabriel Schneider2026-11-05
| | | | | | | the 1-wide range at the cursor Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* Pipe: a pending pipe's command, directory and inputs live in one arena (as ↵Gabriel Schneider2026-11-05
| | | | | | | edit_cmd.Pending's do), so its three failure paths and its deinit stop freeing four allocations by hand Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* Pipe: a pipe's answer is spliced in by sam_edit.apply, its cuts and outputs ↵Gabriel Schneider2026-11-05
| | | | | | | as ops, and each range is put over its output by a running delta, not by a second hand-written copy loop Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* Text, normal, modal: the last f/F/t/T is kept as its Find kind (find_op: ↵Gabriel Schneider2026-11-05
| | | | | | | ?Find), not as the config character that bound it, and findMove and findTarget take the kind instead of two bools; Text's walked shape changes, so a client and a daemon from either side of this change refuse each other (wire semantics) Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* modal, edit: advanceBy takes insertedEnd's body (the last newline and a ↵Gabriel Schneider2026-11-05
| | | | | | | count, not a byte loop) and insertedEnd goes, and a typed run's end is computed once, for the splice and the cursor both Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* File: the four anonymous usize comparators handed to std.sort.upperBound are ↵Gabriel Schneider2026-11-05
| | | | | | | one ascending fn Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* modal, normal: the twelve prefix roles take config.zig's names (goto_prefix, ↵Gabriel Schneider2026-11-05
| | | | | | | find_char_fwd, register_prefix...), so normalInput reads every binding by the role's own name and roleBindingName, which mapped the twelve, goes Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* syntax: the two worker parses are one parseApart (its parser, Reader and ↵Gabriel Schneider2026-11-05
| | | | | | | halt callback written once), and highlightApart no longer sets a parser runQuery never reads Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* modal, selection_pipe, syntax, sam_edit: helpers used once are inlined where ↵Gabriel Schneider2026-11-05
| | | | | | | they are used (consumeCount, clampLineCol, codepointAtCursor, isWordBoundary, wmIsPrev, tokenStopped, dropDiffPieces, Exec.text), and moveWord's doc comment goes back above moveWord Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* ninep tree: a ctl line's bytes are the gpa's until its write is answered, ↵Gabriel Schneider2026-11-05
| | | | | | | where they were the step's scratch: a `size` line resizes and syncs (a step, which resets the scratch under it) and then quotes its value in a refusal, so the refusal read freed memory; the base kept the arena's buffer by luck, and the merged line's bigger steps freed it (a segfault in the behaviour test 'size sets the screen of a session no frontend is attached to') Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* panes: a pane's PDF state is its own allocation, `pdf: ?*Pdf.State` (made by ↵Gabriel Schneider2026-11-05
| | | | | | | openPane with pdf_gpa, freed by deinitPane), as a terminal's is, where every Pane held the State inline, a text file's and a terminal's too: a Pane is 49232 bytes where it was 88936 (the State 39704), and a replica's rollback copy of a pane without a PDF is that much smaller; the dump and the wire change with it, the walk writing the slot as a pointer to a State where it wrote the State in place, so a dump written before this, or a peer of an older build, gives its PDF panes no PDF (the walk notes `Pane.pdf: its kind changed; its default taken`, and dump.ensureKind makes each a terminal); the 97 uses change form only (a `|*x|` capture or a `&` dropped, normal.zig's one among them), applyNormal untouched Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* pdf: the test PDFs are files under test/pdf-fixtures, byte for byte what the ↵Gabriel Schneider2026-11-05
| | | | | | | generators wrote, where pdf.zig wrote each object by object (139 lines fewer there): testFixture reads one, the two outlines sized by the bridge's limits stay generated, and the sections bench embeds its outline Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* pdf: a render request crosses the bridge as one pardes_pdf_request, where ↵Gabriel Schneider2026-11-05
| | | | | | | its six sizes were six arguments of each of the four calls (measure, render, highlight paint and rows) and of the geometry, each call checking them itself (pardes_pdf_request_ok now); boundedRequest makes the struct the Zig side passes Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* Messages: a clock's text is one function, Message.clockText(seconds, ↵Gabriel Schneider2026-11-05
| | | | | | | gmtoff), which the message stamp (with localtime_r's offset) and the tape's guessed clock both call where each wrote its own, and PARDES_NOTIME is read in one place, Message.noTime, which lsp_client's hideTime asks Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* pdf: searchRows prints each hit's row straight into the caller's buffer, ↵Gabriel Schneider2026-11-05
| | | | | | | where it printed each (up to 512) into the arena and copied it over; a row that does not fit ends the list as it did, and look.zig no longer hands it an arena Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* pdf: every by-value copy of a pane's PDF state (about 39 KB, its 256 raster ↵Gabriel Schneider2026-11-05
| | | | | | | slots) taken to read a field or two is a pointer capture, in panes, edit, mouse, tagline, look, exec, the core, 9P ctl, fs, draw, macos and the sections bench; the hot one is pdf_view.moving, which Pardes.continuous asks on every pass of the shell loop and which copied every PDF pane's state each time Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* terminal: resizeGrid takes the pane and the size, where it took an allocator ↵Gabriel Schneider2026-11-05
| | | | | | | too and threw it away (the emulator resizes with its state's own): the core's two calls (pardes.zig), the 9P pty's two and the kitty tests' two pass none Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge