summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAge
* tests: the boot, argv and builtins snapshot scripts (22: argv badutf boot ↵Gabriel Schneider13 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 Schneider13 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 Schneider13 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
* gui: the workspace row's underlay is the tagline's ground: the window ↵Gabriel Schneider13 days
| | | | | | | painted every tagline-role cell's band in the tag's colour under the cell's own, so a theme whose tagline is another colour (dark_plus's blue) showed the tag's gray at the row's left inset and between its cells like an anchor that is not there; the surface carries tagline_bg and the rows it covers, and the band takes it there 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 Schneider14 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 Schneider14 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
* helix: the page-scroll mismatches were the view's height, not a rule: ↵Gabriel Schneider14 days
| | | | | | | pardes's scroll (Text.scrollViewMove, halfPageMove, gotoWindow) is helix's commands.rs scroll/page_* rule for rule, but hxdiff ran pardes at 80x24, where its chrome leaves the pane 21 rows to helix's 22 (half page 10 to 11, page 21 to 22); the harness runs the core at 80x25 so the pane is helix's 22 and refuses a case whose pane is not, and helix-keys.md says why: hxdiff's twelve page-scroll cases and gb-after-zz pass (16 to 3), hxparity's five page waivers match again (5 to 0) Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* helix: hxparity's pty run parked its cursor at (0,0) but left the view where ↵Gabriel Schneider14 days
| | | | | | | tty mode had it, at the live grid's bottom, so zj there scrolled nothing and scrolloff snapped the cursor to row 13 where the file run's went to 4; the harness now lets the view follow the parked cursor (ensureCursorVisible, the state a real gg reaches: the same 738 results), zj-snaps-cursor passes, and the seven viewport waivers that only recorded the stranded view go (zj-snaps-cursor, zj-noop-in-band, zk-bottom-snap, sticky-view-scroll, sticky-view-then-move, sticky-view-ignores, view-z-not-sticky); helix-keys.md's parity waivers say so Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* helix: zz and zc put the cursor's line (rows - 1) / 2 rows down the view, ↵Gabriel Schneider14 days
| | | | | | | helix's align_view (helix-view/src/lib.rs:62-67), where pardes put it rows / 2 - 1 down, a row too high on an odd view: hxdiff's four gt-after-zz cases at 80x24 (21 body rows) pass, gb-after-zz passed only by the view's own missing row and now waits for it; a behaviour test centres l15 on the 21-row view Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* fix: Follow's cursor past a page could hop into a collapsed page of the ↵Gabriel Schneider14 days
| | | | | | | chain, which shows nothing (a collapsed page counted as an empty page): the chain flows past collapsed pages, its hops and ends skip them, and each keeps its scroll until it opens; reference.typ's Follow says what it does not do (a line longer than a page shows its start only, as in any pane; a hop leaves one cursor; two views ordering the pages differently); a behaviour test collapses the last of three pages Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* fix: a session whose runtime directory was not the user's alone ran with no ↵Gabriel Schneider14 days
| | | | | | | 9P socket and no word, which read as a detached session parked before its listener (a bash recipe's mkdir -m 700 -p dir/sub leaves dir at the umask's 755; smoke.py chmods it): the socket check (9p_io.noSocket, the long path's too) says no socket: <dir> must be a directory of yours alone (mode 755, owner 1000): chmod 700 it, to the log, pane 0's message and a --detach start's stderr, and --9p's help says the directory must be mode 700; a unit test opens the directory to others Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* fix: a build with the addons compiled in (any off Linux, the static step) ↵Gabriel Schneider14 days
| | | | | | | and -Dpdf=false did not compile, Addon.zig importing a pdf_addon the build had not made: the builtin_addons options say whether the pdf addon is built, and without it the image addon is compiled in alone; an aarch64-macos -Dpdf=false compile no longer names pdf_addon (its one error left is daemon.zig's darwin CMSG_CLOEXEC, a break of its own) Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* fix: the Debug box's frame timings blanked once their numbers passed ↵Gabriel Schneider14 days
| | | | | | | draw.zig's 96-byte buffer (FrameStats.summary's bufPrint caught to ""): the widest summary is 132 bytes, FrameStats.summary_max, summary takes a buffer of that size and cannot fail, and draw's is one; a test fills the ring with the widest numbers Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* fix: bytes for a terminal's program past the 4 MiB that may wait behind the ↵Gabriel Schneider14 days
| | | | | | | effect ring were dropped with no word: parkPendingWrite says so in the pane's message, write: N bytes for the program dropped, past the 4 MiB that may wait for it; a core test writes past the ring and the cap Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* fix: the ledger's recent.merge leak on an out-of-memory appendSlice is not ↵Gabriel Schneider14 days
| | | | | | | one: merge's defer frees every path still in opens on that return; a test fails each of a merge's allocations in turn and testing.allocator finds nothing left, kept as the pin Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* fix: a pipe's answer past max_stdout_bytes or the job's total, or with cuts ↵Gabriel Schneider14 days
| | | | | | | that overlap, left the text as it was and said nothing: each says so in the pane's message, the text left as it was; a test answers a two-selection pipe with an output one byte past the cap Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* fix: Recent's one-time import from the files before the state file read ↵Gabriel Schneider14 days
| | | | | | | nothing, and said nothing, for a directory whose path passed its room (a bufPrint caught to ""): a directory past max_path_bytes is said in the message, Recent: the files before the state file were not read: their directory's path is N bytes, too long; the import test gains such a window Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* fix: upgrade.load's unlink of an older build's dump file took its path ↵Gabriel Schneider14 days
| | | | | | | through a catch "" that a path past 4096 bytes would have hit, unlinking nothing: such a path never loads (fs.readFile refuses it, PathTooLong), and the session starts afresh saying so with the dump kept for a Restore, so the unlink 0-terminates the path where it runs and says why it fits; a test loads a record naming a 4096-byte path Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* fix: a window handover under a runtime directory too long for a window's 9P ↵Gabriel Schneider14 days
| | | | | | | socket made the window and waited 15 s on a registry entry that could never answer (past ~490 bytes its path was empty): it says so first, no new window: XDG_RUNTIME_DIR is N bytes, too long for its 9P socket; daemon.find's reason for a name past its room is cut, never empty (Zig 0.17's bufPrint writes what fits), and a test pins that Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* fix: a raw terminal's key whose text passed forwardKey's 256-byte buffer (an ↵Gabriel Schneider14 days
| | | | | | | input method's commit, a paste as one key) was dropped with no word: the encoding grows past the stack's room onto the heap and goes out through emitWrite, cut into writes in order; a failure to grow is reported; a pane test forwards 1000 bytes of key text and reads them back whole Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* fix: a session whose runtime directory makes its 9P socket's path pass ↵Gabriel Schneider14 days
| | | | | | | sun_path said so only to std.log, off by default, so a detached one ran with no socket and no word: the check (9p_io.tooLong) gives the words, 9p: the socket path is N bytes, 108 is the most: set XDG_RUNTIME_DIR shorter, to the log, to pane 0's message and, at a --detach start, to stderr, and the session runs on; a unit test pins the check and the message Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* fix: test/daemon.py --upgrade --gui waited for a second view's frame in the ↵Gabriel Schneider14 days
| | | | | | | session's own captures (caps/view-1), which no build writes since a session shows one frontend of its own: the second window is a replica client, opened as window_clients' second is (its own process and captures), and after the upgrade the session's window is drawn again while the old build's client window is let go and told to upgrade too Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* fix: test/daemon.py --upgrade's fds level counted the second terminal as a ↵Gabriel Schneider14 days
| | | | | | | view of its own (its terminal and its stub's pipe, two fds) where it joins as a replica client today (one connection, let go at the exec and told to upgrade too): the session lost exactly its three replica connections (ss -xp maps each to its peer), so nothing leaked; the level takes one for a client and two for a view, and the redraw check accepts the client's word Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* gui: a fine wheel's or a touchpad's turns that stop at a fraction settle ↵Gabriel Schneider2026-11-07
| | | | | | | onto a whole row: two loop steps with no turn (SDL tells no gesture's end), scroll_lag eases onto the core's row on the glide's law, which is the next row the way the turns went, so the settle never moves against them, and a turn mid-settle goes on from where it got; the picture no longer rests a fraction of a row off its cells; a gui test ends a burst at a quarter row and turns again mid-settle Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* gui: a body scrolling under held rows (TreeContext's sticky context) glides ↵Gabriel Schneider2026-11-07
| | | | | | | without trash: the glide measured and kept its rows from the body's top, so the gap showed a held row or none, its rows that left were drawn above the held band where the scissor hid them (the layer's first row showed through twice, 0.9 rows apart), and a body layer's rows past the grid (what the compact context rows save at TaglineSize < 100) were left behind at the bottom; the glide now takes each body's held rows (Glide.held), the gap rows go under them, the touch path's edge row too, and the layer's extra rows move with the rest; a glide test pins the kept row under held rows Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* gui: the caret is drawn with its line while a wheel's fraction holds the ↵Gabriel Schneider2026-11-07
| | | | | | | text off its cells: a bar (and the caret's quad while it eases) was drawn on its cell, up to a row off its text, which a high-resolution wheel leaves resting at a fraction nearly all the time (caretRide now carries scroll_lag as it carries the glide); a gui test pins the bar on its line at a resting half row Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* gui: a wheel's fraction that moves the picture owes a frame even when the ↵Gabriel Schneider2026-11-07
| | | | | | | core does not scroll: a high-resolution wheel's turns (fractions, the touchpad path) that took the picture back onto the core's row drew nothing, so the window rested on a frame a fraction of a row off ("halfway") and the next turn's frame jumped from there, against the turn when it reversed (the bounce); two gui tests turn the wheel as SDL does and hold the drawn picture to the state and to the turn's way Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* gui: a kind's pane (an image, a PDF) is watched again and shows its file as ↵Gabriel Schneider2026-11-07
| | | | | | | it is now: Kind.open and a restored kind's pane emit its watch, which kinds #14 dropped with pdf_view, and a reload counts into its pictures' keys, since its kind numbers them afresh from each open and a window keeps a texture by key; the T1 session watches files as a host does (Session.filesChanged) and two behaviour cases pin it Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* perf.zig: --op <label> measures one gesture alone, and the three baselines ↵Gabriel Schneider2026-11-07
| | | | | | | (tui and gui Debug, tui ReleaseFast) are re-recorded on an idle machine at this head, re-keyed to the harness: the old files carried the pre-rewrite's numbers, and every cell is now the rewrite's own (render 90→81 µs Debug, 77→73 ReleaseFast; edits 2.6x, body-appends 100x faster) Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* perf: recolorSyntax paints a grapheme's cells in a loop per painter again ↵Gabriel Schneider2026-11-07
| | | | | | | (its syntax colour, a diff's emphasis ground, a result's emphasis, a +Pager's colours), each run only when it paints, where qoqpsnqx's one loop asked all four on every cell: that cost the medium file's frame 185k instructions (1.15M to 1.33M, +16%), about 6 µs of every render, key, page and wheel step, and the frame is back to 1.14M, what it was before the refactor Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* follow: `Follow` (SPC c f) is Emacs's follow-mode, one text across a chain ↵Gabriel Schneider2026-11-07
| | | | | | | of Zerox twins shown in column order as consecutive pages (File.followChain, which syncTwins runs only for a twin group holding a chain): a scroll of any page reflows the others from it, the keyboard's page leaves its inner edges to the cursor, which goes on into the next page rather than scrolling it, the chain's ends scroll it as one pane, `Unfollow` on a follower's tag (or Del) takes a pane out, and the chain is a twin's `follow` flag with each view's `follow_scroll`, so a dump, Restore and replicas carry it; test/behaviour/follow.zig and follow.snap pin it, the reference, guide, changelog and builtins-json say it, and the builtins, leader and ttyhelp goldens list SPC c f Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* themes: a theme may colour the tagline (the workspace row) on its own, ↵Gabriel Schneider2026-11-07
| | | | | | | tagline_bg and tagline_fg, an editor's status bar, resolved to the tag's colours without them (ChromeTheme); dark_plus takes VS Code's statusBar.background #007acc with white text, cited; the chrome floors test holds the pair to 4.5:1 and the default to the tag's; the theme and themesel snapshots move by that row Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* gui: a loaded session asks its host for its font again, as it does its ↵Gabriel Schneider2026-11-06
| | | | | | | ThemeFile: an Upgrade's process runs no startup config and the dump's font was already acknowledged, so the new window kept its built-in face at 27 px (Adwaita Mono, 1.5x the user's Go-Mono:18) and told the core that was the font in effect Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* kinds #51, builtins-json lists the shipped addons' words, read off the Zig ↵Gabriel Schneider2026-11-06
| | | | | | | kinds' `Word` lines (image.zig and pdf.zig, handed to the generator and read by the staleness test), each `"addon"` first so the docs test's search for a builtin never meets one and an addon word the docs name is checked against the sources (renaming PdfFit fails it); the glossary says whose addon a word is, and the changelog Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* kinds #50, macOS by pattern: off Linux every core module has the addons ↵Gabriel Schneider2026-11-06
| | | | | | | compiled in (Addon.zig's loader is a Linux memfd, now refused elsewhere at compile time), the native shell registers them at start as the tui and gui do, and `-Dmacos-sdk=<path>` hands SDL's build the SDK's paths, so `-Dplatform=gui -Dtarget=aarch64-macos` configures (and `zig build core-test -Dtarget=aarch64-macos` compiles its test binary, MuPDF and both kinds in, without running it) Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* kinds #49, the static builds: `zig build static -Dplatform=tui ↵Gabriel Schneider2026-11-06
| | | | | | | -Dtarget=x86_64-linux-musl` and `zig build static -Dplatform=gui` (dynamic, both addons in) build and open a PNG and a PDF over 9P with `/addons` saying builtin; `Addon load` in a pardes with its addons compiled in is refused in words (`not on this platform: its addons are compiled in`) rather than by a failed dlopen, and building.typ and shared-libs.md name the gui's static step Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* kinds #48, the export rule finished: every Zig kind declares `pub const ↵Gabriel Schneider2026-11-06
| | | | | | | pardes_addon = addon.ops(T)` and calls `addon.exportLib(&pardes_addon)` in a comptime block, which exports it only when the output is a library (Zig 0.17 analyses no `pub const` of a library's root and no comptime block of a type made in a generic function, so the export cannot live in `ops` alone: one comptime line in the kind is the spelling that works), and a pardes with kinds compiled in reads the declarations; image.zig's `pub export`, snake's and life's `export` go, no rename anywhere, and fake.c stays the loader's test, never compiled in Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* kinds #47, MuPDF and FreeType lazy: Zig 0.17's configure is handed the -D ↵Gabriel Schneider2026-11-06
| | | | | | | options and never the steps asked for (the configurer's argv, and the configure cache keyed on it), so no step can fetch for itself; the guard is `-Dpdf`, on by itself once MuPDF is fetched (std.Build.package_map's `available`) and off Linux, which asks `b.lazyDependency` for MuPDF, and FreeType (with the zlib and libpng it brings) is lazy too and asked by it or a window: a fresh checkout's `zig build -Dplatform=tui` fetches none of them, and without `-Dpdf` the pdf, addons, static, mupdf-check, pdf-harness and bench steps say to add it; building.typ, install.typ and shared-libs.md say so Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* kinds #46, the docs say pdf is Zig over MuPDF: reference.typ's Addons ↵Gabriel Schneider2026-11-06
| | | | | | | section (the module's init and held body, pdf.zig through pdf_bridge.c, its jobs, one place shown by each view, the text layer's lines and the glide), shared-libs.md (the library and the static build), the changelog Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* kinds #45, the PDF benches drive pdf.zig, take another build's library ↵Gabriel Schneider2026-11-06
| | | | | | | (--addon, so pdf.c's from the base runs beside it under one core) and a touchpad's turns (--turn F: whole notches glide in pdf.zig, so the like-for-like fling is turns), and the scroll bench's first frame is as a host shows it, the frame after the first page lands (settled_ns, the page ahead landed too, the number it printed before) Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* kinds #44, what kinds 5 could not restore: as text the place names a line of ↵Gabriel Schneider2026-11-06
| | | | | | | the page (L:F, F its share of the page's lines), the wheel, gj, gk, Ctrl-e and Ctrl-y move a line and a page at the window's edge (j, k and the window keys still turn a page, as the keys test pins); with pictures a wheel's whole notch glides there, half the way each frame drawn and the rest by its fourth tick, a touchpad's fraction moving at once; pdf.script shows the lines, and the touchpad pane case waits out a notch's glide Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* kinds #43, the PDF kind in Zig: addons/pdf/pdf.zig replaces pdf.c over the ↵Gabriel Schneider2026-11-06
| | | | | | | same pdf_bridge.c (its header translated by build.zig, Zig 0.17 having no @cImport), one place a pane (a page and how far down it, which where says and address takes) laid out by each view its own way (a view with pictures its own strip, rasters and thumb by its id, a view of glyphs the page's text layer), the view of the last input the one whose window's top the place follows; the pages' sizes, the page's text and its hits, and the search over every page on the lanes as jobs (rows streamed through found, a stale search cancelled, its first hit gone to as a Look of its first row went), the tick asking and laying out only, open's one parse and close kept on the loop (0.4 and 0.8 ms; a truncated file stays a refused Look); a band at most PA_RASTER_MAX; the static build reads the module's pardes_addon, exported only from the library Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* kinds #42, what a Zig kind over a C library needs of the module and the ↵Gabriel Schneider2026-11-06
| | | | | | | core: addon.zig's `init` at load (a library's own setup, MuPDF's allocator) and a body the kind holds (`body(s) []const u8`, any length: a PDF page's text is past body_max), the Debug budget counting only the kind's own time (the file's hash before `open` and pa_api.heap's allocations are the core's: a region's first took 23 ms of a PDF's open, which was the breach), `current` the loop's thread's (a job's heap call is in no call), an Output's arg set from itself (a kind's rows found after its search refill its +Search with its own arg: the @memcpy aliased), and the docs test reading the words of Zig kinds (`.{ .name = "PdfFit"`) beside C's Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* kinds #41, the soak's stale tty screen (7a's two blank rows of a PDF's text ↵Gabriel Schneider2026-11-06
| | | | | | | layer, seed 2): a kind's frame whose row count changed sent the rows it kept at their place as hashes the peer held, while the peer's Shadow of the new size started from blank cells, so those rows stayed blank until they moved; a row is held now only at the same size, and a test sizes a frame up and down and finds its first row each time Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* kinds #40, rasters as bounded messages on the wire (§17.3 C): a placement ↵Gabriel Schneider2026-11-06
| | | | | | | 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
* kinds #39, paint writes nothing a step reads and the digest is every Slot ↵Gabriel Schneider2026-11-06
| | | | | | | 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
* kinds #38, a kind's ticks and views as units of the session's origin (§17.2 ↵Gabriel Schneider2026-11-06
| | | | | | | 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
* Build: a word that prints this build into a +Build pane (SPC h b): the ↵Gabriel Schneider2026-11-06
| | | | | | | version line, the commit, the jj change and the build time (stamped as the commit is: release builds and the ~/.local install, else "not stamped"), the zig that built it, its mode, target, platform and grammars, whether the addons are compiled in, the wire's semantics hash and the dependencies; build.zig stamps the change (jj log -r git_head()) and the time (date -u) through one command helper beside gitCommit; a T1 case reads the pane Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* kinds #37, tests and docs for ABI 3: the image dump tests reopen from the ↵Gabriel Schneider2026-11-06
| | | | | | | path (through the same /n/os mount, and a file gone waits saying so until a watch or a Look opens it), the PDF view tests assert the art words in every view and each view told, /addons says 3.0 and a 2.2 fake is refused, reference.typ's Addons section and the changelog say ABI 3 Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
* kinds #36, the five kinds on ABI 3: fake.c with every function (view ↵Gabriel Schneider2026-11-06
| | | | | | | painting each view's size, search through found, FakeJob cancel) and PA_PICTURES, snake, life and image over the module's prefix and open(path, bytes), image without a record (a Restore reads its file), and pdf.c ported mechanically (tick into pa_tick with its [P/N] prefix, a table of the views it is told, its strip laid out for the view its last key or click came from, the text layer painted for a window of glyphs and read on the tick while some view shows text, a body read keeping nothing, search through found, the record its fit, tint and place, PA_PICTURES); the goldens moved once, said in the report: the art words from the first frame, fake's view line, a pane ticked when opened or loaded, a view told after a frame re-wrapped its tag Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge