Compare commits

..
4 Commits
Author SHA1 Message Date
Ansho Rei 96cddc14ad (53ffb94) Run the command the prompt parsed, and redraw the board on what it wrote, and bump to 0.32.16
The prompt parsed a line and named what it resolved to. This is the half
that does it: the parsed command goes to the same cmd_* handler main.rs
hands it to, against the selected tab's own .worklane, with a Sink over a
buffer collecting what the handler says instead of stdout.

One dispatch table, not two. main.rs's match on Command is lifted into
`dispatch`, over the command, the start path and the sink, and both
callers go through it. A second copy in tui.rs would drift the first time
a command was added and would drift silently, since both halves go on
compiling; the match is exhaustive so a new verb cannot fall into an `_`
arm either. Only `tui` is matched outside it, in main, because the board
is that table's caller rather than an entry in it, and reached from the
prompt it is refused in those words.

Two things the board has to do that the CLI never had to. It reloads
unconditionally after a command returns rather than waiting for the
digest, so a reader does not watch their own mv take up to a poll
interval to appear, and either way, since a handler that failed part way
through has still written. And the selection is restored by uuid across
that reload: mv puts the card in another column, set --title renames its
file, and a filter can drop it out of view, so the slot is not the
ticket. Where the board cannot follow it, the notice says so and which
key shows it.

What the handler says is what the board reports, word for word. Notice
therefore holds as many lines as the command used, and notice_on_row is
the one rule for where they go: one line on the footer row where a copy's
notice has always gone, anything else as a box at the foot of the
columns. A box rather than more rows of footer, because a footer that
grew and shrank as notices came and went would reflow the columns on
every keypress. A handler that errors leaves the prompt open with the
line intact, the usual cause being an ambiguous id.

`without_escapes` in util.rs is the guard at the seam. A handler's own
output is plain because the sink is, but an error message is built with
color_text, which asks the process about its stdout and under the board
answers yes about a terminal the message is not going to; a ratatui
paragraph prints an escape rather than obeying it. That is 6d327df's
fault on a new path.

The merge tab is refused for now, resolved by index rather than by the
tab's name, which is a label two projects can share. Routing by the card
a command names is ebc5755.

Not done here, and both tracked: check with no criterion still reaches
ask_which_criterion and reads a raw-mode stdin, which is 6d37042's
criterion 4 and commented there; and check_blockers warns about a
dangling blocker with eprintln!, which reaches the alternate screen
because it never passes through dispatch's result, filed as c680e24.
2026-08-27 23:38:02 +02:00
Ansho ReiandClaude Opus 5 f16e2aba0c (3686dd6) Open a command prompt on the board with : and parse what is typed, and bump to 0.32.15
The board reads. Changing anything meant quitting it and retyping an id
that was already selected. So `:` opens a one line editor in the footer
and the line is a worklane command, spelled exactly as the shell spells
it.

This is the mode and the grammar check, not the doing. Enter tokenises
the line with split_line, prepends the binary's own name and hands the
arguments to the same Cli definition the shell reaches, so a flag added
to a command is typeable here the day it is added and no list of verbs
lives on the board to go stale. A line that resolves is named and the
prompt closes; a line that does not keeps the prompt open with the line
intact and clap's own message, cut back to what fits a row. Running it
is 53ffb94, and the notice and the `↵ check` key both say so rather than
reading like a move that happened.

The prompt owns the keyboard while it is open, the contract search_key
and picker_key already keep, and declines ctrl-c for the same reason
they do. A line that ends inside a quote never reaches clap, which is
what the splitter's Unclosed state is for. --help is caught, or clap's
page lands on the alternate screen the way an unsunk println! would.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-27 21:07:26 +02:00
Ansho ReiandClaude Opus 5 bf7d8e05d8 (54fde0b) Split a typed command line into arguments the way a shell does, and bump to 0.32.14
The command prompt reads one string and clap reads an argv, so a line has
to be tokenised before anything can parse it: `comment 1b78a24 the offset
was never passed` split on whitespace is five arguments and `comment`
takes two.

One pure function over the line and the cursor, in a module of its own
rather than in tui.rs. The rules are the shell's, because that is what
the prompt promises. Two answers a shell's own splitter does not owe
anybody come out of the same scan: what a half-typed line is still
inside, as a state to report rather than an error to raise, and which
argument the cursor is in with whether it begins one.

Nothing calls it yet. The prompt that will is 3686dd6, and the dead_code
allow on the module goes when it lands.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-27 20:13:01 +02:00
Ansho Rei 30a2b3429e Specify in skill that set overwrites and to use append when that is not desired 2026-08-27 14:11:59 +02:00
8 changed files with 2135 additions and 79 deletions
+30 -1
View File
@@ -1,7 +1,7 @@
# AGENTS.md
## Project Structure
- `src/main.rs` - dispatch, the SIGPIPE reset, and the TUI event loop
- `src/main.rs` - `dispatch`, the one mapping from a parsed command to its handler, the SIGPIPE reset, and the TUI event loop
- `src/command.rs` - CLI argument definitions using clap derive
- `src/completions.rs` - the generated completion script and the fish half that asks the board
- `src/handlers.rs` - one `cmd_*` per subcommand
@@ -9,6 +9,7 @@
- `src/campaign.rs` - the campaign file format, its resolution, its index and its members
- `src/history.rs` - the history log: the path, the kinds, the reader, the writer, the card's counter and the block `show` and the board both render
- `src/tui.rs` - board state, the change digest, and the ratatui draw
- `src/split.rs` - one pure function turning a typed line into arguments the way a shell does, plus what the half-typed line is still inside and where the cursor sits among them
- `src/clipboard.rs` - the OSC 52 write and the base64 it carries, the board's only way out to the clipboard
- `src/install.rs` - the embedded agent assets and the compare-and-refuse install
- `assets/SKILL.md` - the agent skill for the CLI itself, embedded with `include_str!`
@@ -350,6 +351,34 @@ and this file is for what someone changing worklane's own code needs.
- Both closed footers name `/`, since the jump works with a ticket open as well: it moves the selection and the overlay follows
- Out of scope on purpose, each its own ticket once the input mode exists: `n` and `N` for the next and previous match, searching the description, tags or owner, regex, and filtering the board down to matches rather than selecting one
### The Command Prompt
- **`prompt: Option<Prompt>` is the whole mode**, the shape `search` and `detail` use: open and content cannot disagree, and closing it leaves nothing to reset. `Prompt` carries the line, the cursor as a byte offset into it, and `refused`, which is `Search::missed` with something to say
- **`Board::prompt_key` is third in the pre-dispatch**, beside `search_key` and `picker_key` and on the same contract: while a command is being typed `q` is a letter, `d` leaves the columns alone and `tab` stays out of the project tabs, and `ctrl-c` is the one key declined. It takes a `Size` for enter's sake: running a command reloads the board and then has to bring the ticket it moved back into view, which is the measurement `follow_selection` takes. Only one of the two prompts can be open at once: `:` is reachable from the action arms alone, and while either has the keyboard the other's key is a character in a line
- **`esc` and `enter` are the mode's own keys and are answered in `prompt_key`; everything else belongs to the line and goes to `Prompt::key`.** That is also what keeps the borrow straight, since closing the prompt takes the whole board and editing takes only the prompt. Left, right, home, end, backspace and `ctrl-w` behave as they do in a shell, and `ctrl-w` eats the separators behind the cursor before the word behind those, so one press at the end of `mv 1b78a24 ` takes the id. There is deliberately no `tab`: suggesting the selected ticket is 00a14e7, and until then it is a key that must not reach the tabs
- **The grammar is clap's, with nothing restated.** The line is tokenised by `split_line`, the binary's own name goes in front, and the arguments go to the same `Cli` definition the shell reaches, so a flag added to a command is typeable here the day it is added. Through `Cli::command().try_get_matches_from_mut` rather than `try_parse_from`, for what the matches give back: `command_name` walks the subcommands to name what resolved (`campaign show`, not `campaign`) instead of a `match` over every variant, which would be a second list of every verb the CLI has and one that goes stale silently
- **`refused` is cleared by an edit and not by a movement.** The message is about the line, and moving the cursor does not change the line. `Prompt::key` clears it in the three arms that write to the line
- **A line that ends inside a quote never reaches clap.** That is what `Unclosed` is for: a grammar complaint about a sentence nobody finished describes the wrong thing. Enter on an empty line, or one of nothing but separators, closes the prompt with nothing said, the way `esc` does — the bare form of the CLI opens a board and there is one on the screen
- **Enter runs it, through `main.rs`'s `dispatch` and never through a handler this file names.** That match is the one mapping from a parsed command to the function that performs it: `main` calls it with its own stdout and the working directory, `Board::run` calls it with a buffer and the selected tab's root. A second copy inside `tui.rs` would drift the first time a command was added, and it would drift *silently*, since both halves go on compiling. The match is exhaustive so a new verb cannot fall into an `_` arm and do nothing, and `main_and_the_prompt_run_one_dispatch_table` reads both files as text to pin it. Only `tui` is matched outside the table, in `main`, because the board is that table's caller rather than an entry in it; reached from the prompt it is refused in those words
- **`from_arg_matches` on the same matches gives `run` the typed values.** One parse, two readings: `command_name` walks the subcommands to name what resolved and `from_arg_matches` reconstructs the `Cli`. A `match` over every `Command` variant to get one or the other would be a second list of every verb the CLI has
- **The handler's own words are what the board reports, word for word.** Those messages are written for a person already, and a second set kept here would go stale the first time one of them changed. The sink is a buffer, so the words come back as text and the board styles them; it is plain for the same reason `Detail` holds spans, since a ratatui paragraph prints an escape rather than obeying it. An error never goes through a sink at all — `color_text` asks the *process* about its stdout, which under the board answers yes about a terminal the message is not going to — so `error_lines` runs `without_escapes` over it and puts each cause on a line of its own
- **A handler that errors leaves the prompt open with the line intact.** The usual cause is an ambiguous id, and the fix is two more characters rather than the whole line retyped from `:`. Ambiguity is an `Err` carrying the candidates, not a print and an `exit`, which is what makes it safe to reach from here at all
- **The reload after a command is unconditional, and runs whether it failed or not.** `tick` reloads on the digest, which is right for a change made in another pane and needless latency for one the board just made: the reader would watch their own `mv` take up to a poll interval to appear. Either way, because a handler that failed part way through has still written
- **The selection is restored by uuid across that reload, never by index.** `mv` puts the card in another column, `set --title` renames its file, and a filter can drop it out of view entirely, so the slot is not the ticket. `Board::reselect` has three answers: the card is in a drawn column and the selection goes to it; it is on the board but in a column the rotation is hiding, which is where `--status done` lands by default, so the notice says where it went and which key shows it; or the filter no longer holds it, which is said too, since a card vanishing from under the highlight reads as a bug in the board
- **The merge tab is refused, for now.** It is the one tab that is not a board — its cards came from several `.worklane` directories and it is none of them — so `Board::current_project` answers `None` there and the command is refused rather than landing on whichever project came first. Routing by the card a command names is ebc5755. The project is resolved by index and not by the tab's name, which is a label two projects can share: e1d79c3 is exactly that, and a write routed by name would be that bug with a file changing at the end of it
- **`Notice` holds as many lines as the command used, and `Board::notice_on_row` decides where they go.** One line with nothing else holding the row draws on the footer, where a copy's notice has always drawn. Anything else — several lines, or any answer at all while the prompt is still open on an error — draws as a box at the foot of the columns, cut to what the columns leave with the dropped count on its last row. A box rather than more rows of footer, and that is the whole reason it is an overlay: a footer that grew and shrank as notices came and went would reflow the columns on every keypress, moving every card the reader is looking at. `draw` and `draw_footer` both ask `notice_on_row`, so the two cannot come to disagree and draw the message twice or not at all
- **What a refused line shows is clap's own message with the usage block cut off**, which is also what drops the `--help` line after it. `--help` is the exception: it parses as an error of the kind that means *print this* and what it would print is a page, so it is caught and answered in a row's worth of words, or clap's help lands on the alternate screen the way an unsunk `println!` would. `--version` goes through the ordinary path, because what it prints is one line by construction. `ColorChoice::Never` on the command, so nothing that reaches the row carries an escape
- **The keys' width comes off the row before the line is windowed**, so a long command does not push them off the end; a message takes the row from the keys instead, since the reader has just been told something and `esc` and `enter` still do what they did. `prompt_window` anchors the window on the cursor's right edge, which costs the text *after* the cursor on a long line: keeping both ends needs a scroll offset in state, and the draw is where the width is known and the state is not writable
- `: command` is named in every footer state, because the key works in every one of them, and the prompt's own key is drawn as `↵ run`
### Splitting a Typed Line
- **`split_line` is the whole module, and it knows nothing about `Board`.** The command prompt reads one string and clap reads an argv, so `comment 1b78a24 the offset was never passed` has to be tokenised before anything can parse it: split on whitespace it is five arguments and `comment` takes two. A pure function over the line and the cursor is what makes every rule reachable from a unit test, and it is in a file of its own rather than in `tui.rs`, which is eleven thousand lines already
- **The rules are the shell's, because that is what the prompt promises**: the line is spelled there exactly as it is spelled in the shell. Runs of whitespace separate arguments, single and double quotes group one, a backslash escapes the next character *except* inside single quotes where it is a backslash, and an explicitly empty `''` is an argument that exists and is empty — which is what `--title ""` being a usage error rather than a silent clear depends on. A leading hyphen is not special here: whether `--foo` is a flag or the text of a criterion is clap's question, and `PROSE` already answers it
- **Written rather than pulled in, for the two answers a shell's own splitter does not owe anybody.** A line being typed is half finished, so an unterminated quote is a state to report and not an error to raise: `Unclosed` says which quote is open, or that the line ends on a backslash with nothing yet to escape, and the arguments up to that point are handed back anyway, because that is what a hint or a completion has to work from. And completion needs the cursor: which argument it is in counting from the verb, and whether nothing of that argument lies before it, which is the difference between completing an argument and starting one. Both fall out of the one scan and neither is reachable afterwards from the arguments alone
- **`Some("")` is an argument that has been opened and is still empty**, which is the only thing that tells `''` from nothing typed at all. The cursor's answer is taken *before* the last argument is pushed, so a cursor at the end of the line is in the argument being typed rather than one past it
- The cursor is a byte offset, because that is what the prompt's own cursor is: a `String` inserts and removes at byte indices. Past the end reads as the end and inside a multi-byte character reads as after it, so nothing panics and no caller has to clamp first
- `Unclosed::Escape` wins over an open quote, being the inner of the two, and either answer refuses the line
- The binary's own name is not among the arguments: prepending it before `Cli::try_parse_from` is the caller's, since only the caller knows what it is calling
### The Copy Key
- **`y` copies the selected card's short id, over OSC 52 written by hand.** The board is where a ticket is found and the CLI is where it is acted on, and the only thing joining them was seven characters retyped off the screen. `y` reads as yank to the readers a keyboard board is for, and was the letter left: `q`, `esc`, `enter`, `space`, `hjkl`, the arrows, `pgup`, `pgdn`, `tab`, `backtab`, `d` and `/` were taken
- **The mechanism was chosen against ssh**, which is where a board is often read. An escape sequence is the only one of the three ways that crosses the connection, since it is the terminal at the reader's end that owns the clipboard. A crate (`arboard` and its kind) wants X11 and Wayland libraries at build time, and on X11 the selection belongs to the process that set it: copy an id, quit, and the paste buffer is empty, which is precisely this flow. Shelling out to `wl-copy`, `xclip` or `pbcopy` is platform sniffing plus a process spawn, and does nothing on a machine where none of them is installed or on the far end of a connection. crossterm 0.28 is what ratatui 0.29 pins and it ships no clipboard command, and the backend types have to match whatever ratatui pins, so a newer crossterm cannot be picked up on its own
Generated
+1 -1
View File
@@ -1503,7 +1503,7 @@ dependencies = [
[[package]]
name = "worklane"
version = "0.32.13"
version = "0.32.16"
dependencies = [
"anyhow",
"assert_cmd",
+1 -1
View File
@@ -1,6 +1,6 @@
[package]
name = "worklane"
version = "0.32.13"
version = "0.32.16"
edition = "2024"
[dependencies]
+6 -4
View File
@@ -173,10 +173,12 @@ worklane chown alice 1b78a24 # hand the ticket to someone
worklane chown --clear 1b78a24 # hand it back to nobody
```
`set` replaces, `append` adds, `--clear` empties. Neither `set` nor `append` has
`--owner` or `--status`: owner is `chown`'s and status is `mv`'s, so each field has
exactly one writer. A field set to the value it already holds is reported unchanged and
nothing is written, and so is a field cleared that was already unset.
`set` replaces, `append` adds, `--clear` empties. `set --tags` overwrites the whole
field: to add a tag to a ticket that already carries one, replace the value with every
tag separated by commas. Neither `set` nor `append` has `--owner` or `--status`: owner
is `chown`'s and status is `mv`'s, so each field has exactly one writer. A field set to
the value it already holds is reported unchanged and nothing is written, and so is a
field cleared that was already unset.
**`--clear` takes the field as its value**, one of `due`, `tags`, `priority`,
`description`, `acceptance` or `plan`, and repeats. It writes back exactly what `add`
+115 -66
View File
@@ -1,3 +1,5 @@
use std::path::Path;
use anyhow::Result;
use clap::Parser;
@@ -8,6 +10,11 @@ mod completions;
mod handlers;
mod history;
mod install;
/// The command prompt reads `split_line` and `unclosed`; `cursor_arg` and `at_new_arg` are
/// what completing an argument needs and nothing reads them yet, so the allow stays until
/// 00a14e7 and 952a9f8 land. The tests are what drive them in the meantime.
#[cfg_attr(not(test), allow(dead_code))]
mod split;
mod tag;
mod tui;
mod util;
@@ -45,76 +52,16 @@ fn main() -> Result<()> {
let mut sink = Sink::std();
let result = match cli.command {
// The board is the one thing `dispatch` does not run, and it is matched here
// rather than there because it is the *caller* of that table rather than an entry
// in it: a `tui` reached from the command prompt would enter the alternate screen
// a second time from inside the draw loop. `dispatch` refuses it in those words.
//
// No subcommand opens the board, the way a bare `frig` does, and with no path
// it opens the board above the working directory.
None => run_tui(&[], &start),
Some(Command::Tui { paths }) => run_tui(&paths, &start),
Some(Command::Init) => cmd_init(&start, &mut sink),
Some(Command::Add(args)) => cmd_add(&args, &start, &mut sink),
Some(Command::Ls(args)) => {
let worklane_dir = find_worklane_dir(&start)?;
cmd_ls(&args, &worklane_dir, &mut sink)
}
Some(Command::Show { id }) => {
let worklane_dir = find_worklane_dir(&start)?;
cmd_show(&id, &worklane_dir, &mut sink)
}
Some(Command::Mv(args)) => cmd_mv(&args, &start, &mut sink),
Some(Command::Set(args)) => cmd_set(&args, &start, &mut sink),
Some(Command::Append(args)) => cmd_append(&args, &start, &mut sink),
Some(Command::Relate(args)) => cmd_relate(&args, &start, &mut sink),
Some(Command::Chown(args)) => cmd_chown(&args, &start, &mut sink),
Some(Command::Check(args)) => cmd_check(&args, true, &start, &mut sink),
Some(Command::Uncheck(args)) => cmd_check(&args, false, &start, &mut sink),
Some(Command::Comment {
ticket,
body,
actor,
actor_from_git,
}) => cmd_comment(&ticket, &body, &actor, actor_from_git, &start, &mut sink),
Some(Command::Campaign { command }) => match command {
CampaignCommand::Add(args) => cmd_campaign_add(&args, &start, &mut sink),
CampaignCommand::Ls { ids } => {
let worklane_dir = find_worklane_dir(&start)?;
cmd_campaign_ls(&worklane_dir, ids, &mut sink)
}
CampaignCommand::Set(args) => cmd_campaign_set(&args, &start, &mut sink),
CampaignCommand::Append(args) => cmd_campaign_append(&args, &start, &mut sink),
CampaignCommand::Show { campaign } => {
let worklane_dir = find_worklane_dir(&start)?;
cmd_campaign_show(&campaign, &worklane_dir, &mut sink)
}
},
Some(Command::Assign {
campaign,
ticket,
actor,
actor_from_git,
}) => cmd_assign(&campaign, &ticket, &actor, actor_from_git, &start, &mut sink),
// The bare form and `ls` are one arm: `worklane tags` was the listing before the
// group existed and the fish completion still calls it that way, so the two
// spellings cannot be allowed to drift into two behaviours.
Some(Command::Tags { command }) => match command {
None | Some(TagCommand::Ls) => {
let worklane_dir = find_worklane_dir(&start)?;
cmd_tags(&worklane_dir, &mut sink)
}
Some(TagCommand::Add(args)) => cmd_tag_add(&args, &start, &mut sink),
Some(TagCommand::Set(args)) => cmd_tag_set(&args, &start, &mut sink),
},
// No board either, for the reason `install-skill` has none: a completion script
// is generated from the clap definition and is the same wherever it is run.
Some(Command::Completions { shell }) => {
completions::generate(shell, sink.as_write()).map_err(Into::into)
}
// No board is looked for: the point of the skill is to teach an agent worklane
// before there is a `.worklane` anywhere.
Some(Command::InstallSkill { skill, force, print }) => {
cmd_install_skill(skill.as_deref(), force, print, &mut sink)
}
// No board either, and one agent rather than both: opencode loads the tool file,
// Claude Code has no path for one.
Some(Command::InstallTool { force, print }) => cmd_install_tool(force, print, &mut sink),
Some(command) => dispatch(&command, &start, &mut sink),
};
// The board (and every handler) own their own writing, so flush what the last one
@@ -123,6 +70,93 @@ fn main() -> Result<()> {
result
}
/// Run one parsed command against the board at or above `start`, saying what it did
/// through `sink`.
///
/// **The one mapping from a parsed command to the function that performs it.** `main`
/// calls it with the process's stdout and the working directory; the command prompt on
/// the board calls it with a buffer and the selected tab's own root, and reads back what
/// the handler said. A second copy of this match inside `tui.rs` would be a copy that
/// drifted the first time a command was added, and it would drift *silently*, since both
/// halves would go on compiling.
///
/// A command added to [`Command`] therefore lands here once and is runnable from both.
/// The match is exhaustive for exactly that reason: there is no `_` arm for a new verb to
/// fall into and do nothing.
pub fn dispatch(command: &Command, start: &Path, sink: &mut Sink) -> Result<()> {
match command {
// Reachable only from the prompt, since `main` takes this arm before it gets
// here. A board is not something a board can open.
Command::Tui { .. } => Err(anyhow::anyhow!("`tui` opens a board, and you are on one")),
Command::Init => cmd_init(start, sink),
Command::Add(args) => cmd_add(args, start, sink),
Command::Ls(args) => {
let worklane_dir = find_worklane_dir(start)?;
cmd_ls(args, &worklane_dir, sink)
}
Command::Show { id } => {
let worklane_dir = find_worklane_dir(start)?;
cmd_show(id, &worklane_dir, sink)
}
Command::Mv(args) => cmd_mv(args, start, sink),
Command::Set(args) => cmd_set(args, start, sink),
Command::Append(args) => cmd_append(args, start, sink),
Command::Relate(args) => cmd_relate(args, start, sink),
Command::Chown(args) => cmd_chown(args, start, sink),
Command::Check(args) => cmd_check(args, true, start, sink),
Command::Uncheck(args) => cmd_check(args, false, start, sink),
Command::Comment {
ticket,
body,
actor,
actor_from_git,
} => cmd_comment(ticket, body, actor, *actor_from_git, start, sink),
Command::Campaign { command } => match command {
CampaignCommand::Add(args) => cmd_campaign_add(args, start, sink),
CampaignCommand::Ls { ids } => {
let worklane_dir = find_worklane_dir(start)?;
cmd_campaign_ls(&worklane_dir, *ids, sink)
}
CampaignCommand::Set(args) => cmd_campaign_set(args, start, sink),
CampaignCommand::Append(args) => cmd_campaign_append(args, start, sink),
CampaignCommand::Show { campaign } => {
let worklane_dir = find_worklane_dir(start)?;
cmd_campaign_show(campaign, &worklane_dir, sink)
}
},
Command::Assign {
campaign,
ticket,
actor,
actor_from_git,
} => cmd_assign(campaign, ticket, actor, *actor_from_git, start, sink),
// The bare form and `ls` are one arm: `worklane tags` was the listing before the
// group existed and the fish completion still calls it that way, so the two
// spellings cannot be allowed to drift into two behaviours.
Command::Tags { command } => match command {
None | Some(TagCommand::Ls) => {
let worklane_dir = find_worklane_dir(start)?;
cmd_tags(&worklane_dir, sink)
}
Some(TagCommand::Add(args)) => cmd_tag_add(args, start, sink),
Some(TagCommand::Set(args)) => cmd_tag_set(args, start, sink),
},
// No board either, for the reason `install-skill` has none: a completion script
// is generated from the clap definition and is the same wherever it is run.
Command::Completions { shell } => {
completions::generate(*shell, sink.as_write()).map_err(Into::into)
}
// No board is looked for: the point of the skill is to teach an agent worklane
// before there is a `.worklane` anywhere.
Command::InstallSkill { skill, force, print } => {
cmd_install_skill(skill.as_deref(), *force, *print, sink)
}
// No board either, and one agent rather than both: opencode loads the tool file,
// Claude Code has no path for one.
Command::InstallTool { force, print } => cmd_install_tool(*force, *print, sink),
}
}
// ---------------------------------------------------------------------- tui --
/// Restore the terminal whatever happens.
@@ -215,6 +249,15 @@ fn run_tui(paths: &[std::path::PathBuf], start: &std::path::Path) -> Result<()>
continue;
}
// And the command prompt, third, on the same contract: while a command is being
// typed `q` is a letter, `d` leaves the columns alone and `tab` stays out of the
// project tabs. It takes the screen because enter runs the command and then has
// to bring the ticket it moved back into view. The rule is `Board::prompt_key`,
// where a test can reach it.
if board.prompt_key(key, screen) {
continue;
}
match key.code {
// One arm, because the two keys mean one thing: close what is open, and
// quit when there is nothing left to close. Two arms carrying one rule
@@ -258,6 +301,12 @@ fn run_tui(paths: &[std::path::PathBuf], start: &std::path::Path) -> Result<()>
// at an open picker is taken by `picker_key` to open the picker's own
// query, so this arm is only ever the way in to the board's prompt.
KeyCode::Char('/') => board.open_search(),
// The way in to the command prompt, and the same single-shot arm `/` is:
// from the next keypress on, `prompt_key` above has the keyboard. A `:`
// typed into either open prompt is a character there and one pressed at an
// open picker is swallowed by it, so this arm is only ever the way in from
// the board itself. Free of every existing binding.
KeyCode::Char(':') => board.open_prompt(),
// `y` as in yank, for the readers this board is for, and the only letter
// left that a vim hand would reach for. The clipboard is handed in rather
// than reached for inside `Board`, so the suite can drive both answers.
+453
View File
@@ -0,0 +1,453 @@
//! Splitting a typed line into the arguments a parser wants.
//!
//! The prompt on the board takes one string and clap takes an argv, so something has to
//! turn `comment 1b78a24 the offset was never passed` into a verb, a ticket and one body
//! with spaces in it. Splitting on whitespace hands clap five arguments where `comment`
//! takes two, which is not a grammar error to report but a tokenising step nobody ran.
//!
//! The rules are the shell's, because that is what the prompt promises: a line is spelled
//! there exactly as it is spelled in the shell. Runs of whitespace separate arguments,
//! single and double quotes group one, a backslash escapes the next character except
//! inside single quotes where it is a backslash, and an explicitly empty `''` is an
//! argument that exists and is empty — which is what `--title ""` being a usage error
//! rather than a silent clear depends on. A leading hyphen is not special here: whether
//! `--foo` is a flag or the text of a criterion is clap's question, and the prose
//! arguments already accept one.
//!
//! Two things a shell's own splitter does not have to answer, and both are why this is
//! written rather than pulled in:
//!
//! - **A line being typed is half finished.** An unterminated quote is a state to report,
//! not an error to raise, because the prompt has to keep taking characters and refuse
//! only enter. See [`Unclosed`].
//! - **Completion needs to know where the cursor is.** Which argument it sits in, counting
//! from the verb, and whether nothing of that argument lies before it. Both fall out of
//! the scan for free and neither is reachable afterwards from the arguments alone.
//!
//! One pure function over the line and the cursor, and nothing here knows about `Board`.
/// What a line is still inside when it ends.
///
/// A half-typed line is not a line that failed: `comment 1b78a24 "the offset` is a
/// sentence somebody is in the middle of, and the arguments before the quote are already
/// right. So this is a state the caller reports, and the reason enter is refused while it
/// is anything but [`Unclosed::Closed`] — the arguments up to that point are still handed
/// back, because that is what a completion or a hint has to work from.
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum Unclosed {
/// Nothing. The arguments are whole and the line can go to a parser.
Closed,
/// A `'` with no `'` after it.
Single,
/// A `"` with no `"` after it.
Double,
/// A trailing `\`, which has nothing yet to escape.
///
/// Not a quote, and reported the same way for the same reason: the character it will
/// escape is the next one typed. It wins over an open quote, being the inner of the
/// two, and either answer refuses the line.
Escape,
}
/// Which quote the scan is inside.
///
/// Private, because what a caller needs is [`Unclosed`], which also carries the state no
/// quote can. Two public spellings of one fact is how they come to disagree.
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
enum Quote {
Single,
Double,
}
/// One scan of a line: its arguments, where the cursor sits among them, and what the line
/// is still inside.
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct Split {
/// The arguments, in order, with the quotes and escapes taken off. The binary's own
/// name is not among them: prepending it is the caller's, since only the caller knows
/// what it is calling.
pub args: Vec<String>,
/// Which argument the cursor is in, counting from the verb at 0. Equal to
/// `args.len()` when the cursor is past the last one, which is the index an argument
/// typed next would take.
pub cursor_arg: usize,
/// Whether nothing of `cursor_arg` lies before the cursor, because the line up to it
/// ends in a separator or has nothing in it at all. That is the difference between
/// completing an argument and starting one.
pub at_new_arg: bool,
/// What the line is still inside, and therefore whether it can be run at all.
pub unclosed: Unclosed,
}
/// Split `line` into arguments, reading the cursor at `cursor` as a byte offset into it.
///
/// Bytes because that is what the prompt's own cursor is: a `String` inserts and removes
/// at byte indices. A cursor past the end of the line reads as the end, and one inside a
/// multi-byte character reads as being after that character, so neither can panic and
/// neither needs the caller to have clamped it first.
pub fn split_line(line: &str, cursor: usize) -> Split {
let mut args: Vec<String> = Vec::new();
// `Some("")` is an argument that has been opened and is still empty, which is what
// tells `''` from nothing at all: a bare `take()` cannot make that distinction.
let mut arg: Option<String> = None;
let mut quote: Option<Quote> = None;
let mut escape = false;
// The cursor's answer, taken at the character it falls on and never revised.
let mut at: Option<(usize, bool)> = None;
for (offset, c) in line.char_indices() {
if at.is_none() && offset >= cursor {
at = Some((args.len(), arg.is_none()));
}
if escape {
arg.get_or_insert_default().push(c);
escape = false;
continue;
}
match (quote, c) {
// The shell's exception, and the reason a criterion reading `a\b` survives
// being typed as `'a\b'`: single quotes are literal, backslash included.
(Some(Quote::Single), '\\') => arg.get_or_insert_default().push(c),
(_, '\\') => escape = true,
(None, '\'') => {
quote = Some(Quote::Single);
arg.get_or_insert_default();
}
(None, '"') => {
quote = Some(Quote::Double);
arg.get_or_insert_default();
}
(Some(Quote::Single), '\'') | (Some(Quote::Double), '"') => quote = None,
// A run of separators closes the open argument once and then finds none.
(None, ws) if ws.is_whitespace() => {
if let Some(done) = arg.take() {
args.push(done);
}
}
_ => arg.get_or_insert_default().push(c),
}
}
// Before the last argument is pushed, so a cursor at the end of the line is still in
// the argument it is typing rather than one past it.
let (cursor_arg, at_new_arg) = at.unwrap_or((args.len(), arg.is_none()));
if let Some(last) = arg {
args.push(last);
}
Split {
args,
cursor_arg,
at_new_arg,
unclosed: match (escape, quote) {
(true, _) => Unclosed::Escape,
(false, Some(Quote::Single)) => Unclosed::Single,
(false, Some(Quote::Double)) => Unclosed::Double,
(false, None) => Unclosed::Closed,
},
}
}
#[cfg(test)]
mod tests {
use super::*;
/// One line and everything the scan is supposed to say about it.
///
/// The defaults are the empty line's own answers, so a row states the rule it is about
/// and stays quiet about the rest.
struct Case {
line: &'static str,
/// Where the cursor is, or the end of the line, which is where it sits while a
/// line is being typed. Only the cursor rows move it.
cursor: Option<usize>,
args: &'static [&'static str],
cursor_arg: usize,
at_new_arg: bool,
unclosed: Unclosed,
}
impl Default for Case {
fn default() -> Self {
Case {
line: "",
cursor: None,
args: &[],
cursor_arg: 0,
at_new_arg: true,
unclosed: Unclosed::Closed,
}
}
}
/// Every field of every row, each named in its own failure: a table that reports only
/// the first mismatch makes the next run tell a different story than this one.
fn check(cases: &[Case]) {
for case in cases {
let cursor = case.cursor.unwrap_or(case.line.len());
let split = split_line(case.line, cursor);
let args: Vec<&str> = split.args.iter().map(String::as_str).collect();
let at = format!("{:?} with the cursor at {cursor}", case.line);
assert_eq!(args, case.args, "arguments of {at}");
assert_eq!(split.cursor_arg, case.cursor_arg, "cursor argument of {at}");
assert_eq!(split.at_new_arg, case.at_new_arg, "new argument of {at}");
assert_eq!(split.unclosed, case.unclosed, "unclosed state of {at}");
}
}
#[test]
fn a_line_splits_into_arguments_the_way_a_shell_splits_one() {
check(&[
// Nothing typed yet, which is the state the prompt opens in.
Case::default(),
Case {
line: "tags",
args: &["tags"],
at_new_arg: false,
..Case::default()
},
Case {
line: "mv 1b78a24 --status committed",
args: &["mv", "1b78a24", "--status", "committed"],
cursor_arg: 3,
at_new_arg: false,
..Case::default()
},
// Runs of separators collapse, at either end and in the middle, and a tab is
// one of them.
Case {
line: " mv 1b78a24 ",
args: &["mv", "1b78a24"],
cursor_arg: 2,
..Case::default()
},
Case {
line: "mv\t1b78a24",
args: &["mv", "1b78a24"],
cursor_arg: 1,
at_new_arg: false,
..Case::default()
},
// What the ticket for this was written about: one body, spaces and all.
Case {
line: "comment 1b78a24 \"the offset was never passed\"",
args: &["comment", "1b78a24", "the offset was never passed"],
cursor_arg: 2,
at_new_arg: false,
..Case::default()
},
// Single quotes group the same, and a run of spaces inside one is not a
// separator to collapse.
Case {
line: "comment 1b78a24 'two spaces'",
args: &["comment", "1b78a24", "two spaces"],
cursor_arg: 2,
at_new_arg: false,
..Case::default()
},
// A quote in the middle of an argument groups part of it, which is how a
// `--flag=value` carrying a space is typed.
Case {
line: "set 1b78a24 --title=\"a b\"",
args: &["set", "1b78a24", "--title=a b"],
cursor_arg: 2,
at_new_arg: false,
..Case::default()
},
// A backslash escapes the next character: a separator, a quote, itself.
Case {
line: "comment 1b78a24 a\\ b",
args: &["comment", "1b78a24", "a b"],
cursor_arg: 2,
at_new_arg: false,
..Case::default()
},
Case {
line: "comment 1b78a24 \"she said \\\"no\\\"\"",
args: &["comment", "1b78a24", "she said \"no\""],
cursor_arg: 2,
at_new_arg: false,
..Case::default()
},
Case {
line: "comment 1b78a24 a\\\\b",
args: &["comment", "1b78a24", "a\\b"],
cursor_arg: 2,
at_new_arg: false,
..Case::default()
},
// Except inside single quotes, where it is a character like any other.
Case {
line: "comment 1b78a24 'a\\b'",
args: &["comment", "1b78a24", "a\\b"],
cursor_arg: 2,
at_new_arg: false,
..Case::default()
},
// A leading hyphen is clap's business, not the splitter's.
Case {
line: "add --title --print is supported",
args: &["add", "--title", "--print", "is", "supported"],
cursor_arg: 4,
at_new_arg: false,
..Case::default()
},
// An empty quoted value is an argument that exists and is empty. Losing it
// here would turn a usage error into a silent clear.
Case {
line: "set 1b78a24 --title \"\"",
args: &["set", "1b78a24", "--title", ""],
cursor_arg: 3,
at_new_arg: false,
..Case::default()
},
Case {
line: "''",
args: &[""],
at_new_arg: false,
..Case::default()
},
]);
}
#[test]
fn a_half_typed_line_says_what_it_is_still_inside_and_hands_back_the_rest() {
check(&[
// The arguments before the open quote are already right, and the open one is
// as far as it has been typed.
Case {
line: "comment 1b78a24 \"the offset",
args: &["comment", "1b78a24", "the offset"],
cursor_arg: 2,
at_new_arg: false,
unclosed: Unclosed::Double,
..Case::default()
},
Case {
line: "comment 1b78a24 'the offset",
args: &["comment", "1b78a24", "the offset"],
cursor_arg: 2,
at_new_arg: false,
unclosed: Unclosed::Single,
..Case::default()
},
// A quote just opened is an argument that exists and is empty, the same shape
// a closed `''` leaves.
Case {
line: "comment 1b78a24 \"",
args: &["comment", "1b78a24", ""],
cursor_arg: 2,
at_new_arg: false,
unclosed: Unclosed::Double,
..Case::default()
},
// A trailing backslash has nothing yet to escape. It opened no argument, so
// the cursor is starting one.
Case {
line: "mv 1b78a24 \\",
args: &["mv", "1b78a24"],
cursor_arg: 2,
unclosed: Unclosed::Escape,
..Case::default()
},
// Inside a quote the escape is the inner of the two, and it is what the line
// is waiting on.
Case {
line: "comment 1b78a24 \"a\\",
args: &["comment", "1b78a24", "a"],
cursor_arg: 2,
at_new_arg: false,
unclosed: Unclosed::Escape,
..Case::default()
},
// A closed quote reopened is open again: the state is where the scan ended,
// not how many quotes were typed.
Case {
line: "comment 1b78a24 'a' 'b",
args: &["comment", "1b78a24", "a", "b"],
cursor_arg: 3,
at_new_arg: false,
unclosed: Unclosed::Single,
..Case::default()
},
]);
}
#[test]
fn the_split_says_which_argument_the_cursor_is_in_and_whether_it_begins_one() {
check(&[
// The line is split whole whatever the cursor is doing: only the two cursor
// answers move.
Case {
line: "mv 1b78a24",
cursor: Some(0),
args: &["mv", "1b78a24"],
..Case::default()
},
Case {
line: "mv 1b78a24",
cursor: Some(1),
args: &["mv", "1b78a24"],
at_new_arg: false,
..Case::default()
},
// At the end of the verb, still in the verb: the cursor is completing it.
Case {
line: "mv 1b78a24",
cursor: Some(2),
args: &["mv", "1b78a24"],
at_new_arg: false,
..Case::default()
},
// Past the separator, so the argument at 1 has nothing before the cursor even
// though the line goes on to hold one.
Case {
line: "mv 1b78a24",
cursor: Some(3),
args: &["mv", "1b78a24"],
cursor_arg: 1,
..Case::default()
},
// A cursor past the end of the line reads as the end of it.
Case {
line: "mv",
cursor: Some(99),
args: &["mv"],
at_new_arg: false,
..Case::default()
},
// Inside a quoted argument the cursor is in that argument, not in a fourth
// one the quotes might look like they opened.
Case {
line: "comment 1b78a24 \"the offset\"",
cursor: Some(20),
args: &["comment", "1b78a24", "the offset"],
cursor_arg: 2,
at_new_arg: false,
..Case::default()
},
// Byte offsets, so a multi-byte character before the cursor counts for its
// width and not for one: 8 is where `é` starts and 9 is inside it.
Case {
line: "comment é",
cursor: Some(8),
args: &["comment", "é"],
cursor_arg: 1,
..Case::default()
},
// A cursor inside a character reads as being after it, which is the answer
// that cannot claim an argument has not started when it has.
Case {
line: "comment é",
cursor: Some(9),
args: &["comment", "é"],
cursor_arg: 1,
at_new_arg: false,
..Case::default()
},
]);
}
}
+1489 -5
View File
File diff suppressed because it is too large Load Diff
+40 -1
View File
@@ -303,6 +303,42 @@ pub fn color_text(text: &str, color: Color) -> String {
paint(text, color, should_use_colors())
}
/// `text` with every ANSI escape sequence taken out of it.
///
/// The inverse of what [`paint`] writes, and the guard at the seam where a handler's
/// words become something the board draws: ratatui prints an escape as literal bytes
/// rather than obeying it, so a painted status reaches the screen as `\x1b[33m` and
/// the word after it.
///
/// A [`Sink`] the board owns is plain, so this is not about a handler's ordinary
/// output. It is about the strings that never go through one: an error message is
/// built with `color_text`, which asks the *process* about its stdout, and under the
/// board that answers yes about a terminal the message is not going to.
pub fn without_escapes(text: &str) -> String {
let mut out = String::with_capacity(text.len());
let mut chars = text.chars();
while let Some(c) = chars.next() {
if c != '\x1b' {
out.push(c);
continue;
}
// A CSI sequence is `\x1b[`, then parameter bytes, closed by a final byte in
// `@` to `~`. That is every sequence `Color` writes. Anything else is a short
// escape whose one following character goes with it, which `next` has taken.
if chars.next() == Some('[') {
for c in chars.by_ref() {
if ('\x40'..='\x7e').contains(&c) {
break;
}
}
}
}
out
}
/// A destination for a command's output, chosen by the caller.
///
/// `main` hands a handler its own stdout and nothing about the CLI changes; the
@@ -332,7 +368,10 @@ impl<'a> Sink<'a> {
/// An in-memory buffer, plain, for a test or a caller that colours the text
/// itself.
#[allow(dead_code)]
///
/// What the board's command prompt passes: the handler's words come back as text
/// and the board styles them with ratatui, which is also why it must not be handed
/// a coloured one — a paragraph prints an escape rather than obeying it.
pub fn buffer(target: &'a mut Vec<u8>) -> Sink<'a> {
Sink {
writer: Box::new(target),