summaryrefslogtreecommitdiff
path: root/docs
Commit message (Collapse)AuthorAge
* The documentation site's landing page has four short silent clips in ↵Gabriel Schneider14 hours
| | | | | | | | | | | | | docs/site/media, recorded from a hidden test window by test/record_demos.py text-is-the-interface, terminals-are-panes, scripting-over-9p and diff-review: each a webm (VP9), an mp4 (H.264, faststart) and a poster PNG at 960 px, with captions.txt. The script drives a PARDES_TEST session on the virtual clock through stdin mouse events, keys and 9P, keeps the settled frame after each step, draws the pointer and loops each clip with a dissolve back to its first frame. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The docs become a website: a page per chapter, a landing page with ↵Gabriel Schneider14 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | recordings, one Typst bundle run docs/site/site.typ writes every page in one bundle compile: the landing page (the tagline, the recordings listed in media/clips.toml, links to start from), a page per entry of `chapters` in style.typ, and the tutor. Every page shares a tag-line header with the chapter words; #doc links resolve to the right page and anchor; a chapter with four or more sections gets an "On this page" list. Links are relative and pages flat, so it reads from file:// or any static server. `zig build site` installs it to <prefix>/share/doc/pardes/site and fails plainly without typst. The look is 0x4200.cafe's lapis layout (framed panels, hard shadows, a tag-line trail, Departure Mono headings over Crimson Pro) in orchard's greens and amber, forestbones_light when the reader prefers light; both fonts are OFL and shipped. In HTML a key is one pill with its + and ›, and a mouse button is an inline SVG in currentColor. The tagline lives once in style.typ for the sheet, the book and the site, and the single page html.typ gives way to the site. The landing leads with one plain sentence (columns of panes, click to run or open, helix keys, scripts through files), then the tagline and a line on where it runs; a hero image is the first ready clip's poster; "Start: the guide" leads the links, "Hands on: the tutor" next, the rest a plain row; the captions use no undefined words. docs/typ/install.typ holds the install lines once: setup.typ has an Install section that includes it, and the landing includes it too. Building says up front it is for contributors and points at setup to install. In HTML the guide's glossary is a closed <details> (#glossary, the same heading and table on paper), each drawn mouse carries hidden text (left-click, hold left, click middle) for screen readers, w3m and copying, and the footer takes a source link and a licence from `project` in style.typ once decided. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The docs give back the day-one lines the trim cut, the tutor becomes ↵Gabriel Schneider14 hours
| | | | | | exercises on a +Tutor copy, the README builds the GUI and says what a terminal needs, and Shift-Esc says it needs the kitty keyboard protocol Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The docs trimmed to their path: a guide with Words you'll see and one Where ↵Gabriel Schneider14 hours
| | | | | | commands run and panes go table, scripting's first tag word Fmt and eight traps, setup as the one home of the pager and the servers, a scannable cheatsheet, an honest README, the edge cases moved to the reference, and the pager's colours Co-Authored-By: Claude Opus 5.5 <[email protected]>
* SPC q runs Exit, refused once over unsaved panes as the tag word is, as ↵Gabriel Schneider15 hours
| | | | | | helix's :q Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A command line runs when its newline comes, or at its open's close, never ↵Gabriel Schneider16 hours
| | | | | | | | | | | | | | | | because a write was short: a long line cut by a mount is one line A write shorter than its message, and not a multiple of 4096, was taken as whole, its tail a line. 9ns cuts a FUSE write of up to 512 KiB into 65512-byte Twrites, the last one short, so a 600 KiB `Edit ,c/.../` arrived cut and its tail ran as an unknown control message; the 1 MiB refusal failed the same way. Each open now holds a partial line until its newline, or its release, which ends the last line; an Edit block is whole when it closes, as before. A name write with a NUL still fails its write at once. selfmount writes a 600 KiB Edit line and a 1.1 MiB line through a real 9ns mount; building.typ says how lines are cut. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Grep's miss says `Grep: text not found`: it is a literal search, and ↵Gabriel Schneider17 hours
| | | | | | | | | "pattern" suggested a regular expression Still ENOENT through a mount ("not found" is in 9ns's table, and the unit tests' errno-words check holds it so). The guide's quote follows. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Shift-Esc always backs out: in every mode and pane it goes to the previous ↵Gabriel Schneider17 hours
| | | | | | | | | | | | | | | pane, as Esc does from normal, and a raw program never sees it; Ctrl-b alone toggles raw Shift-Esc switched a terminal in normal or insert mode into raw, so the one key meant to get out of anything got a script or a person into a program instead. It is now handled before any mode: an open prompt is closed, insert mode left, a pending count dropped, and Last runs, the pane keeping its mode (a raw terminal stays raw). Ctrl-b is the only switch between raw and normal. The keymap's two chords become one, back_out; the cheatsheet row, the guide's mode table and the tutor say so, and the tty and builtins goldens are re-recorded. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The docs take round 36: empty names, get as an undo step, what a Dump keeps ↵Gabriel Schneider17 hours
| | | | | | and Restore reads, Find and Grep caps and notes, the ctl read's collapsed and fit/tint, PDF search hits, ./x misses, 1 MiB lines, DumpDir, bad dials, /pager files, rename previews, Lspinfo, and --wait from an agent Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The docs take round 35: cloud9's errno table, Restore's parser words, insert ↵Gabriel Schneider19 hours
| | | | | | after a match, LookWord list rows, the last-run awk without exit, command pane directories, rename previews, listings that keep the keyboard, set-value toggles, pty/run one line a write, /pager resolving, PDF :N:H, skipped directories and Lspinfo Co-Authored-By: Claude Opus 5.5 <[email protected]>
* addr, dot and limit on a pane with no text are refused `invalid address: ↵Gabriel Schneider21 hours
| | | | | | | | | this pane has no text to address`, which a mount reads as EINVAL, not EIO 9ns maps a refusal back to its errno by its words; "this pane has no text to address" held none it knows. The reference quotes the new words. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* /pager's refusals say invalid, so a mount's shell hears EINVAL: `pager: ↵Gabriel Schneider21 hours
| | | | | | | | | | | invalid directory "relative": must be absolute` 9ns maps a refusal back to an errno by its words, and "the directory must be absolute" and "one directory a write" held none it knows, so a shell through the mount got EIO for both. They now say `invalid`, as every EINVAL here does, and the reference quotes the new words. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Recipes work through a pane of their own, its look and exec, and keep a ↵Gabriel Schneider21 hours
| | | | | | command pane while their open lasts; the docs give cmd panes, the err grep, alt-screen prompts, /pager's refusals, Tty+ on a ctl, --, and theme failures in words Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Recipes read look and exec on the open they wrote, the reference says a ↵Gabriel Schneider22 hours
| | | | | | fresh open reads anyone's last answer, and /pager is documented Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The docs say Grep is a literal search, that a write's own status comes ↵Gabriel Schneider22 hours
| | | | | | before the log's last err, and that Delcol opens no +Unsaved Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The docs take round 32: petscii words in terminal and image tags, open ↵Gabriel Schneider23 hours
| | | | | | records and EMFILE, ThemeFile failures and replays, trailing-colon looks, pty/run after clear, no-text addr refusals, held log reads at Restore, pardes - in one open, and a detached pty in pixels Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Scripting shows a bash event helper that owns two tag words, writes other ↵Gabriel Schneider23 hours
| | | | | | clicks back and lets event go, and warns that cat event | while read swallows the next click Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The docs take round 31: one +Pager per directory and how pardes - cleans its ↵Gabriel Schneider25 hours
| | | | | | text, SYSTEMD_PAGER, a paged diff drawn as one, event reads in pieces, sel, write-back origins, +PdfSections rows, a PDF body by lines, bare :N looks and quoted bad values Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Every scripting and setup recipe runs as written: fmt-here finds the session ↵Gabriel Schneider25 hours
| | | | | | itself, the recipes share one stated preamble, run a command waits for its tag, and setup says man reads MANPAGER first Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A terminal's kitty graphics draw as glyph art where pixels cannot, and by ↵Gabriel Schneider26 hours
| | | | | | | | | | | | | | | | | | | | choice: Petscii on the pane, TermImages for new ones Under an outer terminal without kitty graphics a tty pardes cannot pass a terminal pane's images on, so it draws them itself as an image pane's Petscii does: each row of an image is matched to C64-palette glyph art over the cells it covers (GlyphArt.renderRect, the exact-grid half of GlyphArt.render), cached while it is on screen. yazi is still told the kitty protocol works, so it keeps sending its previews. The same art is a choice anywhere: the Petscii builtin now takes a terminal pane too (PaneKind.picture), flipping its images between pixels and glyphs, and a terminal holding images, or drawing them as glyphs by choice, says petscii:on|off in its tag as an image pane does. TermImages real|petscii (default real) is what a new terminal starts with; it is in DumpConfig, /ctl and the reference. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The docs describe the pager that landed: pardes - and its +Pager pane, a ↵Gabriel Schneider26 hours
| | | | | | terminal's PAGER, command panes' cat, Pager pardes|off, and pty/run no longer waiting in one Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The book and the HTML gain "Setting up your environment" after the guideGabriel Schneider26 hours
| | | | | | | | | chapters lists setup.typ second and #doc takes a setup id. Shell lines (#cmd) follow the density too: roomy in the book and the HTML, with a tint and space around them, tight on the A4 sheet; the book sets code blocks on the same tint. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Setting up your environment: a 9ns mount, pardes as the editor, pager and ↵Gabriel Schneider26 hours
| | | | | | yazi opener, the shells with prompt marks, language servers, fonts and plan9port, each snippet tried Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The docs take the 9P agent's round 30: log stats 0, names escape every ↵Gabriel Schneider26 hours
| | | | | | control byte, $winid and the pager defaults, the one client rule, and Shift-Esc said one way in the guide and the cheatsheet Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The book and the HTML set #pairs tables roomy, the A4 sheet tightGabriel Schneider26 hours
| | | | | | | | | A pairs-density state, set to tight by cheatsheet-a4.typ and roomy otherwise, gives book and HTML tables more row padding and room between the key and the meaning. The book sets its body at 10pt with more space around sections and between paragraphs and list items. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The book and the HTML hold every topic, and #doc links to its chapter or sectionGabriel Schneider26 hours
| | | | | | | | | | | | | chapters now names guide, cheatsheet, scripting, reference, themes and building. #doc takes the chapter ids, keeps the old ones mapped onto them (tags and fs to the guide and the reference, config, keys and detached to their guide sections), and links in the book and the HTML, where a doc-links state says every chapter is there; on the A4 sheet it names the chapter and section. The reference points at the scripting and building chapters by #doc instead of in words, and the guide heading for pardes FILE and --wait keeps its two dashes. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* docs/typ gets book and HTML wrappers over one chapter listGabriel Schneider26 hours
| | | | | | | | | | | style.typ lists the chapters in reading order (guide, cheatsheet, scripting, reference, building), each with its content file or none while unwritten. book.typ sets them as an A4 book with a title page and contents, the tutor read from src/tutor.txt last; html.typ sets them as one HTML page with a nav of anchors. An unwritten chapter shows a placeholder, so both compile before the topics exist. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The cheatsheet draws mouse buttons as a mouse, keys as chips, and pairs as ↵Gabriel Schneider26 hours
| | | | | | | | | | | | | | | | | | | | | | | | tables #btn and #chord draw a mouse (left button, wheel, right button) with the pressed part filled; the wheel is a narrow slot on a paper halo, hollow or filled, so it never reads as a button. A chord is the held button, then for each click the held one and that one down, with arrows between; #btn(.., shift: true) adds a ⇧. HTML gets a span with an aria-label. A key is one soft pill with no border: a combination (Ctrl-w) reads Ctrl + w and a sequence (g g, SPC f c) g › g, separators muted inside it; #keys("h", "j", ...) sets alternatives as separate pills. The drift test reads every key a #keys call names. #pairs sets key | meaning content as a two-column table on alternating tints; split: 2 sets the leader map as two such tables side by side. The mouse chords get their own table: cut, paste, copy (B1-B2 then B1-B3, one B1 hold, as mouse.zig says), 2-1 with an argument, 2-3 cancelling. The A4 page draws its own title and drops the content's level-1 heading. docs/typ/README is the function contract for content files. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Themes become a short Typst topic: choosing one, the native palettes and ↵Gabriel Schneider26 hours
| | | | | | ports, and a theme file's roles Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Building gathers the platforms, options, tests, release gates and how the ↵Gabriel Schneider26 hours
| | | | | | core, threads, detached sessions and 9P fit, from the old design notes checked against the code Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The reference is fs.md's per-file semantics, errors and limits in Typst, ↵Gabriel Schneider26 hours
| | | | | | with the settings table, and says what the pane ctl's Left, Right, Up and Down do Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Scripting over 9P has one chapter, and the served README and the 9P skill ↵Gabriel Schneider26 hours
| | | | | | shrink to a summary, the traps and a pointer to it Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The guide walks a newcomer through pardes day to day, and the cheatsheet ↵Gabriel Schneider26 hours
| | | | | | says what Esc and Shift-Esc really do Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Docs say how a look resolves a relative path: the looking pane's directory, ↵Gabriel Schneider26 hours
| | | | | | | | | | | then each jump-list pane's, most recent first fs.md's look section and tags.md said nothing of where a relative path is looked for, so a look that found a file in a directory visited an hour ago, or not in an open pane never visited, was a surprise. The order is written as look.zig's resolveLookTarget has it. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Docs give a look miss's real words, that event is one reader's, the real cap ↵Gabriel Schneider26 hours
| | | | | | | | | | | | | | on held reads, Dump's Restore word, a reused command pane's runs, and the one-write trap fs.md's look misses now read as the editor says them (<path>:99 has no line 99, has no page N); event says a second reading open fails file in use; held reads say the 64 state-holding opens are the real cap, under the 128 a connection may park; Dump says it adds a Restore word to the workspace tag; a reused command pane's body is said to keep each run under a % line, with how to take the last. The 9P skill gains the trap that bash's printf to a mounted ctl arrives one write per line. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Edit's w and the other left-out sam commands fail EIO in words, never ↵Gabriel Schneider26 hours
| | | | | | | | | | | | | EOPNOTSUPP, and /os refuses a rename or a chmod as not permitted, EACCES "w is not supported in pardes" read through a mount as EOPNOTSUPP, as if the file system lacked an operation, and a chmod under /os was answered "invalid truncate", EINVAL. An Edit command pardes leaves out now says so in words 9ns reads as EIO, and a rename or mode change of an /os file is "permission denied", EACCES. fs.md's Failure section and /os entry say both. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* name takes any byte a file name can hold, controls, edge blanks and bytes ↵Gabriel Schneider26 hours
| | | | | | | | | | | | not UTF-8 included, so what name reads writes back as it was A pane opened on a file whose name held a tab, a control byte or bytes not UTF-8 read that name from its name file, and the same bytes written back were refused. Only what no file name holds is refused now: a second line and a NUL. The ctl word name still refuses a second blank after the word, as a line's words are split at blanks. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Views generated by each read stat 0, as acme's do: data, xdata, index, ↵Gabriel Schneider26 hours
| | | | | | | | | | | | layout, recent and listeners; fs.md says which files have real lengths A stat of index, layout, recent or listeners built the whole view to count it, and data and xdata reported the body's length though a read answers what addr selects. Those now stat 0, as screen and pane/new already did, and fs.md lists the files whose length is real and says to read the rest to the end. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A shell command from the workspace's or a column's tag runs as a command ↵Gabriel Schneider26 hours
| | | | | | | | | | | | | | | pane in the session's directory, never typed into the terminal with the keyboard A command clicked in a header tag, or written to /tagexec or a column's exec, was typed into the active terminal when it sat at an empty prompt, so it ran in that pane's directory and left no run or exit record. It now always gets a command pane of its own in the session's directory, as acme's row and column tags have no directory but the session's. tags.md writes the rule down: pane words act on the column's active pane, a column's words on the column, shell commands as their own panes, and over 9P a pane's word is refused. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Release Pardes 0.13Gabriel Schneider26 hours
|\ | | | | | | Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * A one-page A4 cheatsheet, the first file of a Typst docs corpus in docs/typGabriel Schneider26 hours
| | | | | | | | | | | | | | | | | | | | | | | | docs/typ/cheatsheet.typ holds the facts as semantic calls (#key, #btn, #chord, #word, #tag, #addr, #file, #cmd, #doc); docs/typ/style.typ defines those once, drawing them for paged output and emitting HTML elements under --features html; docs/typ/cheatsheet-a4.typ is the page. `zig build cheatsheet` renders it when typst is on PATH. A config.zig test reads every .typ in docs/typ and fails when one names a builtin, key binding or default tag pardes no longer has. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | A diff line looks up the path and line it names, as any look doesGabriel Schneider26 hours
|/ | | | | | | | | | | | | | | | | | | | | | | | The look on a diff line was its own resolver: a walk up to .git/.jj from the diff's directory, its own "no such file here" and "deleted by this diff". A review tool's cached patch, in no repository, missed every file though the session's terminals sat in the checkout. Now a diff line only expands and transforms. Its expansion (what the hover shows) is the whole line for a header, a `@@` line, or a hunk line pointed at on its prefix, and its text is the look address the line names (look.diffAddress): `path` for `diff --git`/`---`/`+++`, `path:c` for `@@ -a,b +c,d @@`, `path:N` for a hunk line's new line (a removed line's, the new line now where it was). That text goes through the regular look exactly as if selected by hand: same resolution, placement and errors. A 9P look of a whole diff line takes the same transform. Kept: the line classification, hunk arithmetic, git prefix stripping and timestamp cutting (diff.zig). Gone: the repo-root walk, its messages, Target.old_side and deletesAhead, and the click's row stash (Pardes.diff_look). A test compares a B3 on a cached patch's line with a look of the hand-selected `entry/src/entry.rs:2` in the same session. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Release Pardes 0.12Gabriel Schneider26 hours
|\ | | | | | | Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * Over a program that tracks the mouse, B2 and B3 are pardes's execute and ↵Gabriel Schneider26 hours
| | | | | | | | | | | | | | | | | | | | look, as in any pane, and Shift sends them to the program as its buttons 2 and 3; B1 and the wheel stay the program's, Shift keeping them pardes's The Shift that sends B2 or B3 is left out of their reports and of the drags they hold: it only chose the target. A B2 or B3 sweep of pardes's moves nothing to the program under 1003. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | Recent and /recent keep PDFs and images as they keep files, and an open ↵Gabriel Schneider26 hours
| | | | | | | | | | | | | | | | | | | | | | | | file's row is where its dot is now Only text files were recorded, so a PDF read yesterday was not there to find again, and an open file's row showed the place it last closed at, stale by however far it had moved since. A PDF is kept at its page and reopened there from its row; an image by its path; an open one's row reads its pane's place at the time of the listing. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | DumpConfig writes a setting this shell cannot show as a "# Word unsupported" ↵Gabriel Schneider26 hours
| | | | | | | | | | | | | | | | | | | | | | comment, GripWidth among them on the tty, so the whole report writes back to /ctl The report said it could be written back as is, but "Lift unsupported" and the tty's "GripWidth 100" were each refused by the root ctl, and "DumpDir (none)" would have named a directory "(none)". Those lines are now comments, and a test writes every line of the report back. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | Docs say what Rename, a look at a missing or unreadable file, Restore, a ↵Gabriel Schneider26 hours
|/ | | | | | | | | | | | | PDF's body and one open's undo step do, and the PDF tag in its order fs.md said Rename lists the places; it says its count and lists none. It now also says a look at no such file is a logged miss while an unreadable one fails the write, that Restore brings no jump list back, that a PDF's body is read-only page text no address reaches, and that one open's writes are one undo step only while nothing else edits the pane between them. tags.md listed the PDF tag's words out of order. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A diff look names a file as the repository does: git's side prefix dropped ↵Gabriel Schneider26 hours
| | | | | | | | | | | | | | | | | | by git's rules, a plain diff's names as written In a git section (one with a `diff --git` line) the one-letter side prefix git writes, `a/`/`b/` or with diff.mnemonicPrefix `c/`, `i/`, `w/`, `o/`, is dropped from its paths only when both paths of that line carry one, different ones: `--no-prefix` writes none, so its `a/x a/x` is a real directory `a` and is kept. A plain `diff -u` never strips: its names are used as written, past the timestamp. `/dev/null` stays itself. The walker does this once (diff.Walk.side, diff.gitHeader), so the path a look resolves, and the one a miss says, are the repository's name: `Look: src/x.zig: no such file here`, `Look: lib/old.zig: deleted by this diff`. The look no longer tries a stripped name beside the written one. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* fs.md says a body write to a terminal is typed input, never a bracketed pasteGabriel Schneider26 hours
| | | | | | | | | | | A write to a terminal's body goes to its child as typed keys: no bracketed-paste marks around it, even when the program set 2004, so a newline in it is Enter and a shell runs each line. fs.md did not say which, and a script cannot guess. It does now, and a test writes to a terminal whose program asked for bracketed paste and checks the bytes sent are only the ones written. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Save path logs a save record naming the path it wroteGabriel Schneider26 hours
| | | | | | | | | | | Save of a pane's own file logged `save <serial> <name>`, and so did a scratch's `Save path` (the pane takes that name). A `Save path` that writes a copy of a file pane, or a terminal's scrollback, elsewhere logged nothing, so a follower could not tell the file was written. It now logs `save <serial> <path>` with the path written, once the host has written it (events.noteLogAs). fs.md's log table says so. Co-Authored-By: Claude Opus 5.5 <[email protected]>
generated by cgit v1.3 (git 2.53.0) at 2026-10-02 05:36:21 +0000