Files
worklane/src/completions.rs
T
Ansho ReiandClaude Fable 5 5bf0fd1bfe (451fb83) Add a discarded status for work that will not be done, and bump to 0.31.0
The flow ended at done, so a ticket that turned out to be a duplicate or
obsolete had nowhere to go: it sat in the backlog or was deleted by hand,
which destroyed the record and left every relation to it dangling. Discarded
is a fifth status with its own directory, keeping the invariant that the
filename, the Status field and the directory agree, and keeping done meaning
the work landed.

Entering it is allowed from any status without -f, because discarding is not
a stage of work. Discarded sits last in Status::ALL, which is what makes the
rest fall out rather than needing cases: nothing follows it, so a discard
clears no stamp and a ticket abandoned from reviewing keeps the journey it
made, and every move out is a backward move, so un-discarding needs -f.

The two dependency guards now look opposite ways. check_blockers asks what a
ticket waits on and gates work moves; check_dependents asks who waits on it
and gates the discard, because those plans depended on work about to be
declared never-happening. Both skip targets that are themselves terminal.

Status::FLOW is the four stages the board draws, so a discarded column on
screen stays a136d13's ticket and done stays the last drawn column. Every
tree walker keeps using Status::ALL, or a discarded ticket would be
unaddressable.

util::is_open replaces three spellings of "not done and not discarded",
across mv, ls and the campaign standing. The board's blocked mark goes
through it as well: that path was only accidentally correct, since a
discarded blocker is absent from the FLOW columns, and drawing the column
later would have diverged it from what mv allows.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-24 19:44:14 +02:00

617 lines
26 KiB
Rust

//! Shell completions.
//!
//! Two halves, because neither alone is enough.
//!
//! `clap_complete` generates the structure (subcommands, flags, their arguments)
//! straight from the clap definition, so a new flag never needs remembering here. But it
//! can only offer values it knows at compile time, and the value worth completing for
//! worklane is the ticket, which lives in the board the command is run above.
//!
//! So the fish output gets a hand-written block appended that asks worklane itself. That
//! costs one `worklane ls --ids` per keypress, which measured at 2.5ms in release on a
//! board of seventy-five tickets, comfortably below anything a person notices.
//!
//! The four statuses are half clap's: `Status` is a `ValueEnum`, so it emits them for
//! `mv --status`, but the fish generator does not emit possible values for a *positional*
//! and `ls <status>` gets nothing. Same for `completions <shell>`. So the block below
//! lists those two, from the enums rather than by hand.
use std::io;
use clap::{CommandFactory, ValueEnum};
use clap_complete::Shell;
use crate::install::SKILLS;
use crate::util::Status;
/// Subcommands whose first positional is a ticket.
///
/// Kept honest by a test rather than by memory: a new ticket-taking subcommand that is
/// missing here silently loses completion, which nothing else would catch.
pub const TICKET_COMMANDS: [&str; 8] = [
"show", "mv", "comment", "set", "append", "relate", "check", "uncheck",
];
/// Subcommands whose *second* positional is the ticket, because they name two things.
///
/// `chown alice <ticket>` and `assign "v1 release" <ticket>`, which put the ticket
/// second deliberately, see AGENTS.md. The first argument of each is a user or a
/// campaign, and neither is completed.
pub const TICKET_SECOND_COMMANDS: [&str; 2] = ["chown", "assign"];
/// Options whose value is a ticket, as `relate` takes three of.
pub const TICKET_FLAGS: [&str; 3] = ["blocked-by", "regression-of", "relates-to"];
/// Subcommands whose first positional is a campaign.
///
/// `assign` alone, which takes the campaign first and the ticket second. Kept honest by
/// the same test the ticket lists are.
pub const CAMPAIGN_COMMANDS: [&str; 1] = ["assign"];
/// Nested subcommands whose first argument is a campaign, as `parent child`.
///
/// `campaign show`, `campaign set` and `campaign append`, and they are why the fish half
/// needs a third position helper: their argument is the *third* token, where
/// `__worklane_first_arg_of` reads the second and `__worklane_second_arg_of` counts a
/// top-level command's arguments. Neither can tell `campaign show <TAB>` from
/// `chown alice <TAB>` by position.
pub const CAMPAIGN_NESTED_COMMANDS: [(&str, &str); 3] = [
("campaign", "show"),
("campaign", "set"),
("campaign", "append"),
];
/// Options whose value is a campaign.
pub const CAMPAIGN_FLAGS: [&str; 1] = ["campaign"];
/// Options whose value is a tag.
///
/// Its own list rather than a line in the block below, because the values come from a
/// different command: a tag is not an object with an id, so `worklane tags` answers this
/// one where `ls --ids` answers the tickets.
pub const TAG_FLAGS: [&str; 1] = ["tag"];
/// Nested commands whose first argument is a tag the board already has.
///
/// `tags set` changes what a declared tag means, so the values it takes are exactly the
/// ones `worklane tags` prints. `tags add` is deliberately absent: the tag it takes is one
/// the board does not have yet, so offering the ones it does would offer exactly the values
/// it refuses. Which of the two a new tag positional belongs to is what
/// `no_tag_taking_option_is_missing` makes somebody decide.
pub const TAG_NESTED_COMMANDS: [(&str, &str); 1] = [("tags", "set")];
pub fn generate(shell: Shell, out: &mut dyn io::Write) -> io::Result<()> {
let mut command = crate::command::Cli::command();
clap_complete::generate(shell, &mut command, "worklane", out);
if shell == Shell::Fish {
write!(out, "{}", fish_dynamic())?;
}
Ok(())
}
/// The dynamic half, appended to what clap generated.
///
/// Fish merges completions from every `complete` call, so these add to the generated
/// rules rather than replacing them.
fn fish_dynamic() -> String {
let ticket_commands = TICKET_COMMANDS.join(" ");
let ticket_second_commands = TICKET_SECOND_COMMANDS.join(" ");
let ticket_flags = flag_list(&TICKET_FLAGS);
let campaign_commands = CAMPAIGN_COMMANDS.join(" ");
let campaign_nested = CAMPAIGN_NESTED_COMMANDS
.iter()
.map(|(parent, child)| format!("__worklane_nested_arg_of {parent} {child}"))
.collect::<Vec<_>>()
.join("; or ");
let campaign_flags = flag_list(&CAMPAIGN_FLAGS);
let tag_flags = flag_list(&TAG_FLAGS);
let tag_nested = TAG_NESTED_COMMANDS
.iter()
.map(|(parent, child)| format!("__worklane_nested_arg_of {parent} {child}"))
.collect::<Vec<_>>()
.join("; or ");
let statuses = possible_values(Status::value_variants());
let shells = possible_values(Shell::value_variants());
let skills = SKILLS
.iter()
.map(|skill| skill.name)
.collect::<Vec<_>>()
.join(" ");
format!(
r#"
# ---------------------------------------------------------------------------
# Dynamic values, appended by `worklane completions fish`. Fish accumulates
# completions, so these add to the generated rules above.
# ---------------------------------------------------------------------------
# Almost every argument worklane takes names something on the board rather than
# something on disk, so stop fish falling back to the directory listing. Without this,
# `worklane show 1b78a24 <TAB>` offers Cargo.toml and src/.
complete -c worklane -f
# The exception, and the reason the line above is not the whole story: `tui` takes
# project paths, as many as you like. Directories only, since each is resolved to the
# first `.worklane` at or above it.
complete -c worklane -n '__worklane_command_is tui' \
-a '(__fish_complete_directories)' -d project
# Position is counted rather than inferred from what has been typed: matching against
# the ticket list would run worklane again inside the condition, and would also keep
# offering tickets after one had been chosen.
#
# The subcommand is the second token, so its first argument is being typed once two
# tokens are there. `$tokens[2]` rather than `__fish_seen_subcommand_from`, because
# `campaign ls` and `campaign show` are also an `ls` and a `show` that fish has seen,
# and their arguments are not what these complete.
function __worklane_first_arg_of
set -l tokens (commandline -opc)
test (count $tokens) -eq 2; and contains -- $tokens[2] $argv
end
function __worklane_second_arg_of
set -l tokens (commandline -opc)
test (count $tokens) -eq 3; and contains -- $tokens[2] $argv
end
# The first argument of a nested subcommand, `worklane campaign show <TAB>`. Three tokens
# with both names matched, rather than the token counting the two above do: `campaign show`
# puts its argument where a top-level command puts its second, so counting alone would
# offer campaigns to `chown alice <TAB>` and tickets here.
function __worklane_nested_arg_of
set -l tokens (commandline -opc)
test (count $tokens) -eq 3; and test "$tokens[2]" = "$argv[1]"; and test "$tokens[3]" = "$argv[2]"
end
# Which command this is, at any position in it. `$tokens[2]` again rather than
# `__fish_seen_subcommand_from`, and for the reason spelled out above: `campaign ls` is an
# `ls` that fish has seen, so scoping `--tag` with fish's own helper would offer it under a
# command that does not take it.
#
# A rule carrying `-l` needs one of these. Without it fish takes the option as one the
# whole tool accepts, and `worklane init --<TAB>` offered `--blocked-by`, `--tag` and four
# others that only `relate` and `ls` have.
function __worklane_command_is
set -l tokens (commandline -opc)
test (count $tokens) -ge 2; and contains -- $tokens[2] $argv
end
# No `-d`: `ls --ids` prints the id and the title separated by a tab, which is fish's
# own form for a candidate and its description, so every ticket carries its own.
complete -c worklane \
-n '__worklane_first_arg_of {ticket_commands}; or __worklane_second_arg_of {ticket_second_commands}' \
-a '(worklane ls --ids)'
# `relate` takes its targets as option values, so they are completed by flag rather
# than by position.
complete -c worklane -n '__worklane_command_is relate' {ticket_flags} \
-x -a '(worklane ls --ids)'
# Campaigns, from `campaign ls --ids`, in the three places one is named: `assign`'s first
# argument, `campaign show`'s only one, and `ls --campaign`.
complete -c worklane \
-n '__worklane_first_arg_of {campaign_commands}; or {campaign_nested}' \
-a '(worklane campaign ls --ids)'
complete -c worklane -n '__worklane_command_is ls' {campaign_flags} \
-x -a '(worklane campaign ls --ids)'
# Tags come from `worklane tags` rather than from a listing of objects, since a tag has no
# id. Three fields now, the tag, its count and its annotation, and fish takes everything
# after the *first* tab as one description, so the count is dropped here rather than landing
# in the middle of the prose. `cut` and not `string replace` because the columns are what is
# being picked, and `cut -f1` is how a script reads this listing anyway.
#
# What is left is the annotation, and a tag the board carries but has never declared has
# none: it is offered with no description at all, which is the near-duplicate the file was
# added to catch. The count used to be the description and said the same thing more weakly.
complete -c worklane -n '__worklane_command_is ls' {tag_flags} \
-x -a '(worklane tags | cut -f1,3)'
# And `tags set` names one that is already declared, so it reads the same listing. `tags
# add` has no rule on purpose: it takes a tag the board does not have, so the candidates
# here are the values it refuses.
complete -c worklane -n '{tag_nested}' -a '(worklane tags | cut -f1,3)'
# `chown --clear` names the ticket in the flag too, since with no user to give there is
# no second positional to put it in. Scoped to `chown`, unlike the rule above: `set`
# has a `--clear` of its own that takes a field name, and clap already offers those.
complete -c worklane -n '__worklane_command_is chown' -l clear -x -a '(worklane ls --ids)'
# clap emits the four for `mv --status`, which is an option, but the fish generator
# emits nothing for the possible values of a positional. So the one place a status is
# positional needs them listed, and the same goes for the shell below.
complete -c worklane -n '__worklane_first_arg_of ls' -a '{statuses}' -d status
complete -c worklane -n '__worklane_first_arg_of completions' -a '{shells}' -d shell
# Shipped with the binary, so this list comes from `install::SKILLS` and cannot fall
# behind it.
complete -c worklane -n '__worklane_first_arg_of install-skill' -a '{skills}' -d skill
"#
)
}
/// Option names as fish wants them on a `complete` line, `-l one -l two`.
fn flag_list(flags: &[&str]) -> String {
flags
.iter()
.map(|flag| format!("-l {flag}"))
.collect::<Vec<_>>()
.join(" ")
}
/// The names a `ValueEnum` accepts, for a fish `-a` list.
///
/// Read off the enum rather than written out, so a status or a shell added to either one
/// is offered without anybody remembering this file. A variant marked hidden is left out,
/// which is what `to_possible_value` returning `None` means.
fn possible_values<T: ValueEnum>(variants: &[T]) -> String {
variants
.iter()
.filter_map(|variant| variant.to_possible_value())
.map(|value| value.get_name().to_string())
.collect::<Vec<_>>()
.join(" ")
}
#[cfg(test)]
mod tests {
use super::*;
/// Built, not just derived: an index is assigned when the command is built, and the
/// positionals of an unbuilt one all report `None`.
fn command() -> clap::Command {
let mut cli = crate::command::Cli::command();
cli.build();
cli
}
/// Every argument in the tool, as the path that reaches it and the argument itself.
///
/// One walker for the honesty tests below, rather than each one repeating the
/// traversal: they differ only in which id they are looking for. Two levels, which is
/// as deep as the tool goes, and `help` is skipped because clap generates it and it
/// takes no object of ours.
fn every_argument() -> Vec<(Vec<String>, clap::Arg)> {
let cli = command();
let mut found = Vec::new();
for subcommand in cli.get_subcommands() {
let name = subcommand.get_name().to_string();
if name == "help" {
continue;
}
for arg in subcommand.get_arguments() {
found.push((vec![name.clone()], arg.clone()));
}
for nested in subcommand.get_subcommands() {
let child = nested.get_name().to_string();
if child == "help" {
continue;
}
for arg in nested.get_arguments() {
found.push((vec![name.clone(), child.clone()], arg.clone()));
}
}
}
found
}
/// Every subcommand that takes a ticket must be listed, at the position it takes it,
/// or it silently loses completion.
///
/// Nested subcommands are walked now that `__worklane_nested_arg_of` exists, but
/// nothing under `campaign` takes a ticket, so a nested one is a panic rather than a
/// list: it would need a rule of its own in the fish block, and a name in a list here
/// would look like it already had one.
#[test]
fn no_ticket_taking_subcommand_is_missing() {
for (path, arg) in every_argument() {
// `show` calls it `id`; everything else calls it `ticket`.
if !matches!(arg.get_id().as_str(), "ticket" | "id") {
continue;
}
let name = path.join(" ");
if arg.is_positional() {
assert_eq!(
path.len(),
1,
"`worklane {name}` takes a nested ticket argument, \
which needs a rule in the fish block"
);
let top = path[0].as_str();
match arg.get_index() {
Some(1) => assert!(
TICKET_COMMANDS.contains(&top),
"`worklane {name}` takes a ticket first but is not in TICKET_COMMANDS"
),
Some(2) => assert!(
TICKET_SECOND_COMMANDS.contains(&top),
"`worklane {name}` takes a ticket second but is not in \
TICKET_SECOND_COMMANDS"
),
index => panic!("`worklane {name}` takes a ticket at position {index:?}"),
}
}
}
}
/// The same for campaigns, which is why the lists above are three rather than one:
/// a campaign is named as a first positional, as a nested subcommand's only argument,
/// and as an option value, and each needs a different rule in the fish block.
#[test]
fn no_campaign_taking_argument_is_missing() {
let mut seen = 0;
for (path, arg) in every_argument() {
if arg.get_id().as_str() != "campaign" {
continue;
}
seen += 1;
let name = path.join(" ");
if let Some(long) = arg.get_long() {
assert!(
CAMPAIGN_FLAGS.contains(&long),
"`worklane {name} --{long}` takes a campaign but is not in CAMPAIGN_FLAGS"
);
continue;
}
assert_eq!(
arg.get_index(),
Some(1),
"`worklane {name}` takes a campaign somewhere the fish block has no rule for"
);
match path.as_slice() {
[top] => assert!(
CAMPAIGN_COMMANDS.contains(&top.as_str()),
"`worklane {name}` takes a campaign first but is not in CAMPAIGN_COMMANDS"
),
[parent, child] => assert!(
CAMPAIGN_NESTED_COMMANDS.contains(&(parent.as_str(), child.as_str())),
"`worklane {name}` takes a campaign but is not in CAMPAIGN_NESTED_COMMANDS"
),
_ => panic!("`worklane {name}` is nested deeper than the fish block goes"),
}
}
// A guard on the walker itself: if it stopped finding arguments, every assertion
// above would pass by never running.
assert_eq!(
seen, 5,
"expected assign, campaign show, campaign set, campaign append and ls --campaign"
);
}
/// Commands whose tag positional names a tag that does not exist yet.
///
/// `tags add` declares a word the board has never carried, so completing it from
/// `worklane tags` would offer exactly the values the command refuses. Recorded rather
/// than skipped, so the *next* tag positional has to say which it is: `tags set` takes
/// a tag that is already declared and wants a rule in the fish block, and the test
/// below fails until it gets one or lands here.
///
/// In the test module because nothing outside it reads the list; the rule it encodes is
/// about what the fish block is missing, which is a test's question.
const TAG_NEW_COMMANDS: [(&str, &str); 1] = [("tags", "add")];
/// Tags have their own list because their values come from `worklane tags` rather than
/// from a listing of objects, so a second tag-taking option would need that rule and
/// not the campaign one.
#[test]
fn no_tag_taking_option_is_missing() {
let mut seen = 0;
for (path, arg) in every_argument() {
if arg.get_id().as_str() != "tag" {
continue;
}
seen += 1;
let name = path.join(" ");
let Some(long) = arg.get_long() else {
let known = match path.as_slice() {
[parent, child] => {
let pair = (parent.as_str(), child.as_str());
TAG_NESTED_COMMANDS.contains(&pair) || TAG_NEW_COMMANDS.contains(&pair)
}
_ => false,
};
assert!(
known,
"`worklane {name}` takes a tag as a positional, so it belongs in \
TAG_NESTED_COMMANDS if the tag already exists and TAG_NEW_COMMANDS if it \
does not"
);
continue;
};
assert!(
TAG_FLAGS.contains(&long),
"`worklane {name} --{long}` takes a tag but is not in TAG_FLAGS"
);
}
assert_eq!(seen, 3, "expected ls --tag, tags add and tags set");
}
/// **Every rule naming an option has to say which command it belongs to.** A bare
/// `complete -c worklane -l tag` tells fish the whole tool takes `--tag`, so
/// `worklane init --<TAB>` offered six options that only `relate` and `ls` have. This
/// is what stops the next rule being written the same way.
#[test]
fn every_rule_naming_an_option_is_scoped_to_a_command() {
let script = fish_dynamic();
// Rules are one logical line, continued with a trailing backslash.
for rule in script.replace("\\\n", " ").lines() {
if !rule.starts_with("complete") || !rule.contains("-l ") {
continue;
}
assert!(
rule.contains("__worklane_command_is"),
"this rule declares an option for the whole tool: {rule}"
);
}
}
#[test]
fn every_listed_command_is_a_real_subcommand() {
let cli = command();
let known: Vec<&str> = cli.get_subcommands().map(|c| c.get_name()).collect();
for name in TICKET_COMMANDS.iter().chain(TICKET_SECOND_COMMANDS.iter()) {
assert!(known.contains(name), "no such subcommand: {name}");
}
}
/// `chown --clear` takes a ticket, so it is completed like one, and only under
/// `chown`: `set --clear` is the same flag name carrying a field.
#[test]
fn chowns_clear_is_completed_with_tickets_and_sets_is_not() {
let script = fish_dynamic();
let rule = script
.lines()
.find(|line| line.starts_with("complete") && line.contains("-l clear"))
.expect("chown --clear has a rule of its own");
assert!(rule.contains("__worklane_command_is chown"), "{rule}");
assert!(rule.contains("worklane ls --ids"), "{rule}");
let chown = command().find_subcommand("chown").expect("chown exists").clone();
assert!(
chown.get_arguments().any(|arg| arg.get_long() == Some("clear")),
"`chown` has no --clear for the rule to complete"
);
assert!(
!TICKET_FLAGS.contains(&"clear"),
"--clear in TICKET_FLAGS would offer tickets to `set --clear`, which takes a field"
);
}
/// The flags completed as tickets have to be `relate`'s actual flags, or the rule
/// completes an option nothing accepts.
#[test]
fn every_ticket_flag_is_a_real_option() {
let relate = command()
.find_subcommand("relate")
.expect("relate exists")
.clone();
let known: Vec<String> = relate
.get_arguments()
.filter_map(|arg| arg.get_long().map(str::to_string))
.collect();
for flag in TICKET_FLAGS {
assert!(known.contains(&flag.to_string()), "`relate` has no --{flag}");
}
}
#[test]
fn the_fish_script_asks_worklane_for_its_tickets() {
let script = fish_dynamic();
assert!(script.contains("worklane ls --ids"));
// Both positions, in one rule.
assert!(script.contains("__worklane_first_arg_of show"));
assert!(script.contains("__worklane_second_arg_of chown assign"));
// Nothing here takes a file, bar the one that does.
assert!(script.contains("complete -c worklane -f\n"));
assert!(script.contains("__fish_complete_directories"));
// Generated from `install::SKILLS`, so every shipped skill is in it.
assert!(script.contains("worklane-review"));
}
/// The two listings the campaign and tag rules read, which are separate commands
/// because a campaign has an id and a tag does not.
#[test]
fn the_fish_script_asks_worklane_for_its_campaigns_and_tags() {
let script = fish_dynamic();
assert!(script.contains("worklane campaign ls --ids"), "{script}");
assert!(script.contains("worklane tags"), "{script}");
// The nested position, which neither of the two counting helpers can reach.
assert!(script.contains("__worklane_nested_arg_of campaign show"), "{script}");
assert!(script.contains("__worklane_first_arg_of assign"), "{script}");
}
/// Every rule that reads the tag listing drops the count, because `worklane tags` prints
/// three fields and fish takes everything after the first tab as the description. Left
/// whole, a tag would be described as its count, a tab and its annotation.
///
/// Pinned per rule rather than once over the script: a rule added later that reads
/// `worklane tags` raw is the bug this catches, and it would not show up in a single
/// `contains`.
#[test]
fn every_tag_rule_takes_the_annotation_as_its_description() {
let script = fish_dynamic();
let reading: Vec<&str> = script
.lines()
.filter(|line| line.contains("(worklane tags"))
.collect();
assert_eq!(
reading.len(),
2,
"expected the ls --tag rule and the tags set one: {reading:?}"
);
for line in reading {
assert!(line.contains("(worklane tags | cut -f1,3)"), "{line}");
}
}
/// The one status positional and the one shell positional, which clap does not emit
/// possible values for. Read off the enums, so a fifth status would appear here on
/// its own.
#[test]
fn the_fish_script_lists_the_statuses_and_the_shells() {
let script = fish_dynamic();
for status in Status::ALL {
assert!(
script.contains(status.as_str()),
"the ls positional is never offered {}",
status.as_str()
);
}
assert!(script.contains("-a 'backlog committed reviewing done discarded' -d status"));
for shell in ["fish", "bash", "zsh", "elvish", "powershell"] {
assert!(script.contains(shell), "the shell positional is never offered {shell}");
}
}
/// The four for `mv --status` are clap's, which is half of why `Status` is a
/// `ValueEnum`. If clap ever stops emitting them, the fish half has to carry them
/// too, and this is what says so.
#[test]
fn clap_emits_the_statuses_for_the_status_flag() {
let mut out = Vec::new();
generate(Shell::Fish, &mut out).expect("generates");
let script = String::from_utf8(out).expect("utf8");
let flag = script
.lines()
.find(|line| line.contains("using_subcommand mv") && line.contains("-l status"))
.expect("mv --status has a rule of its own");
assert!(flag.contains("backlog"), "{flag}");
}
#[test]
fn generating_fish_produces_both_halves() {
let mut out = Vec::new();
generate(Shell::Fish, &mut out).expect("generates");
let script = String::from_utf8(out).expect("utf8");
// clap's half.
assert!(script.contains("complete -c worklane"));
assert!(script.contains("install-skill"));
// Ours.
assert!(script.contains("__worklane_first_arg_of"));
}
#[test]
fn other_shells_get_only_the_generated_half() {
for shell in [Shell::Bash, Shell::Zsh] {
let mut out = Vec::new();
generate(shell, &mut out).expect("generates");
let script = String::from_utf8(out).expect("utf8");
assert!(
!script.contains("__worklane_first_arg_of"),
"fish-only helpers leaked into {shell}"
);
}
}
}