Skip to content

Worktrunk is a CLI for Git worktree management, designed for parallel AI agent workflows

feat(list): add git.branch.* template namespace for custom columns (#3319)

Maximilian Roos4388defff530
diff --git a/dev/config.example.toml b/dev/config.example.toml
index c83c9b373..ed0040447 100644
--- a/dev/config.example.toml
+++ b/dev/config.example.toml
@@ -154,19 +154,29 @@
# # lower = kept longer (default: 9, the URL band)
#
# Templates may reference `{{ branch }}`, `{{ worktree_path }}`,
-# `{{ worktree_name }}` (empty for branch-only rows), and `{{ vars.* }}` —
-# per-branch values stored with
-# `wt config state vars set` (https://worktrunk.dev/config/#wt-config-state-vars).
+# `{{ worktree_name }}` (empty for branch-only rows), and two per-branch
+# namespaces:
+#
+# - `{{ vars.* }}` — values stored with
+# `wt config state vars set` (https://worktrunk.dev/config/#wt-config-state-vars).
+# - `{{ git.branch.* }}` — the branch's own git config under `branch.<name>.*`,
+# read straight from `git config` (e.g. `{{ git.branch.jira }}` for a key you
+# set yourself, or the git-native `description`). Git lowercases config variable
+# names, so `branch.<name>.nvciShelf` reads as `{{ git.branch.nvcishelf }}`.
+#
# All standard filters work (`sanitize`, `hash_port`, `codename`, …). A row
-# where the template renders empty (e.g. a branch without the vars key) shows an
+# where the template renders empty (e.g. a branch without the key) shows an
# empty cell; a column that is empty for every row is dropped from the table.
# `wt list --format json` includes the rendered values under `columns`.
#
-# A `Note` column showing free-form descriptions, set per branch with
-# `wt config state vars set note "Bug fix for production fire"`:
+# A `Jira` column reading a key kept in git config, and a `Summary` column
+# showing just the first line of the git-native branch description:
+#
+# [list.custom-columns.Jira]
+# template = "{{ git.branch.jira }}"
#
-# [list.custom-columns.Note]
-# template = "{{ vars.note }}"
+# [list.custom-columns.Summary]
+# template = "{{ git.branch.description | lines | first }}"
#
# ### Commit
#
diff --git a/docs/content/config.md b/docs/content/config.md
index 0543abd38..9b534f317 100644
--- a/docs/content/config.md
+++ b/docs/content/config.md
@@ -246,20 +246,30 @@ priority = 9 # Optional drop order when the terminal narrows
```
Templates may reference `{{ branch }}`, `{{ worktree_path }}`,
-`{{ worktree_name }}` (empty for branch-only rows), and `{{ vars.* }}` —
-per-branch values stored with
-[`wt config state vars set`](@/config.md#wt-config-state-vars).
+`{{ worktree_name }}` (empty for branch-only rows), and two per-branch
+namespaces:
+
+- `{{ vars.* }}` — values stored with
+ [`wt config state vars set`](@/config.md#wt-config-state-vars).
+- `{{ git.branch.* }}` — the branch's own git config under `branch.<name>.*`,
+ read straight from `git config` (e.g. `{{ git.branch.jira }}` for a key you
+ set yourself, or the git-native `description`). Git lowercases config variable
+ names, so `branch.<name>.nvciShelf` reads as `{{ git.branch.nvcishelf }}`.
+
All standard filters work (`sanitize`, `hash_port`, `codename`, …). A row
-where the template renders empty (e.g. a branch without the vars key) shows an
+where the template renders empty (e.g. a branch without the key) shows an
empty cell; a column that is empty for every row is dropped from the table.
`wt list --format json` includes the rendered values under `columns`.
-A `Note` column showing free-form descriptions, set per branch with
-`wt config state vars set note "Bug fix for production fire"`:
+A `Jira` column reading a key kept in git config, and a `Summary` column
+showing just the first line of the git-native branch description:
```toml
-[list.custom-columns.Note]
-template = "{{ vars.note }}"
+[list.custom-columns.Jira]
+template = "{{ git.branch.jira }}"
+
+[list.custom-columns.Summary]
+template = "{{ git.branch.description | lines | first }}"
```
### Commit
diff --git a/docs/content/list.md b/docs/content/list.md
index 90ca1c5d3..a09cfbbff 100644
--- a/docs/content/list.md
+++ b/docs/content/list.md
@@ -137,7 +137,7 @@ Reuses the [`commit.generation`](@/config.md#commit) command — the same LLM th
<span class="badge-experimental"></span>
-Each `[list.custom-columns]` entry in user config adds a column: the key is the header, the template renders each row's cell. Templates can reference per-branch `{{ vars.* }}` stored with [`wt config state vars set`](@/config.md#wt-config-state-vars) — useful for tracking what each of many (often agent-driven) branches is for:
+Each `[list.custom-columns]` entry in user config adds a column: the key is the header, the template renders each row's cell. Templates read two per-branch namespaces — `{{ vars.* }}`, stored with [`wt config state vars set`](@/config.md#wt-config-state-vars), and `{{ git.branch.* }}`, the branch's own git config under `branch.<name>.*` (a `jira` key you set yourself, or the git-native `description`) — useful for tracking what each of many (often agent-driven) branches is for:
```toml
[list.custom-columns.Ticket]
diff --git a/skills/worktrunk/reference/config.md b/skills/worktrunk/reference/config.md
index 7af583c39..f25f6d287 100644
--- a/skills/worktrunk/reference/config.md
+++ b/skills/worktrunk/reference/config.md
@@ -243,20 +243,30 @@ priority = 9 # Optional drop order when the terminal narrows
```
Templates may reference `{{ branch }}`, `{{ worktree_path }}`,
-`{{ worktree_name }}` (empty for branch-only rows), and `{{ vars.* }}` —
-per-branch values stored with
-[`wt config state vars set`](https://worktrunk.dev/config/#wt-config-state-vars).
+`{{ worktree_name }}` (empty for branch-only rows), and two per-branch
+namespaces:
+
+- `{{ vars.* }}` — values stored with
+ [`wt config state vars set`](https://worktrunk.dev/config/#wt-config-state-vars).
+- `{{ git.branch.* }}` — the branch's own git config under `branch.<name>.*`,
+ read straight from `git config` (e.g. `{{ git.branch.jira }}` for a key you
+ set yourself, or the git-native `description`). Git lowercases config variable
+ names, so `branch.<name>.nvciShelf` reads as `{{ git.branch.nvcishelf }}`.
+
All standard filters work (`sanitize`, `hash_port`, `codename`, …). A row
-where the template renders empty (e.g. a branch without the vars key) shows an
+where the template renders empty (e.g. a branch without the key) shows an
empty cell; a column that is empty for every row is dropped from the table.
`wt list --format json` includes the rendered values under `columns`.
-A `Note` column showing free-form descriptions, set per branch with
-`wt config state vars set note "Bug fix for production fire"`:
+A `Jira` column reading a key kept in git config, and a `Summary` column
+showing just the first line of the git-native branch description:
```toml
-[list.custom-columns.Note]
-template = "{{ vars.note }}"
+[list.custom-columns.Jira]
+template = "{{ git.branch.jira }}"
+
+[list.custom-columns.Summary]
+template = "{{ git.branch.description | lines | first }}"
```
### Commit
diff --git a/skills/worktrunk/reference/list.md b/skills/worktrunk/reference/list.md
index 4f5db1ce0..2094d3f63 100644
--- a/skills/worktrunk/reference/list.md
+++ b/skills/worktrunk/reference/list.md
@@ -122,7 +122,7 @@ Reuses the [`commit.generation`](https://worktrunk.dev/config/#commit) command
### Custom columns [experimental]
-Each `[list.custom-columns]` entry in user config adds a column: the key is the header, the template renders each row's cell. Templates can reference per-branch `{{ vars.* }}` stored with [`wt config state vars set`](https://worktrunk.dev/config/#wt-config-state-vars) — useful for tracking what each of many (often agent-driven) branches is for:
+Each `[list.custom-columns]` entry in user config adds a column: the key is the header, the template renders each row's cell. Templates read two per-branch namespaces — `{{ vars.* }}`, stored with [`wt config state vars set`](https://worktrunk.dev/config/#wt-config-state-vars), and `{{ git.branch.* }}`, the branch's own git config under `branch.<name>.*` (a `jira` key you set yourself, or the git-native `description`) — useful for tracking what each of many (often agent-driven) branches is for:
```toml
[list.custom-columns.Ticket]
diff --git a/src/cli/mod.rs b/src/cli/mod.rs
index 5fcc28930..4525a798a 100644
--- a/src/cli/mod.rs
+++ b/src/cli/mod.rs
@@ -865,7 +865,7 @@ Reuses the [`commit.generation`](@/config.md#commit) command — the same LLM th
### Custom columns [experimental]
-Each `[list.custom-columns]` entry in user config adds a column: the key is the header, the template renders each row's cell. Templates can reference per-branch `{{ vars.* }}` stored with [`wt config state vars set`](@/config.md#wt-config-state-vars) — useful for tracking what each of many (often agent-driven) branches is for:
+Each `[list.custom-columns]` entry in user config adds a column: the key is the header, the template renders each row's cell. Templates read two per-branch namespaces — `{{ vars.* }}`, stored with [`wt config state vars set`](@/config.md#wt-config-state-vars), and `{{ git.branch.* }}`, the branch's own git config under `branch.<name>.*` (a `jira` key you set yourself, or the git-native `description`) — useful for tracking what each of many (often agent-driven) branches is for:
```toml
[list.custom-columns.Ticket]
@@ -1970,20 +1970,30 @@ priority = 9 # Optional drop order when the terminal narrows
```
Templates may reference `{{ branch }}`, `{{ worktree_path }}`,
-`{{ worktree_name }}` (empty for branch-only rows), and `{{ vars.* }}` —
-per-branch values stored with
-[`wt config state vars set`](@/config.md#wt-config-state-vars).
+`{{ worktree_name }}` (empty for branch-only rows), and two per-branch
+namespaces:
+
+- `{{ vars.* }}` — values stored with
+ [`wt config state vars set`](@/config.md#wt-config-state-vars).
+- `{{ git.branch.* }}` — the branch's own git config under `branch.<name>.*`,
+ read straight from `git config` (e.g. `{{ git.branch.jira }}` for a key you
+ set yourself, or the git-native `description`). Git lowercases config variable
+ names, so `branch.<name>.nvciShelf` reads as `{{ git.branch.nvcishelf }}`.
+
All standard filters work (`sanitize`, `hash_port`, `codename`, …). A row
-where the template renders empty (e.g. a branch without the vars key) shows an
+where the template renders empty (e.g. a branch without the key) shows an
empty cell; a column that is empty for every row is dropped from the table.
`wt list --format json` includes the rendered values under `columns`.
-A `Note` column showing free-form descriptions, set per branch with
-`wt config state vars set note "Bug fix for production fire"`:
+A `Jira` column reading a key kept in git config, and a `Summary` column
+showing just the first line of the git-native branch description:
```toml
-[list.custom-columns.Note]
-template = "{{ vars.note }}"
+[list.custom-columns.Jira]
+template = "{{ git.branch.jira }}"
+
+[list.custom-columns.Summary]
+template = "{{ git.branch.description | lines | first }}"
```
### Commit
diff --git a/src/commands/list/collect/mod.rs b/src/commands/list/collect/mod.rs
index e38e37a4f..6e018cc4b 100644
--- a/src/commands/list/collect/mod.rs
+++ b/src/commands/list/collect/mod.rs
@@ -1145,9 +1145,10 @@ pub fn collect(
let llm_command = config.commit_generation.command.clone();
// Custom [list.custom-columns] values expand before layout: their inputs
- // (branch, worktree identity, vars from the bulk config snapshot) are
- // already in memory, so cells paint with the skeleton and column widths
- // are measured from content like Branch/Path. Pure CPU — no subprocess.
+ // (branch, worktree identity, vars and branch_config from the bulk config
+ // snapshot) are already in memory, so cells paint with the skeleton and
+ // column widths are measured from content like Branch/Path. Pure CPU — no
+ // subprocess.
//
// A broken column definition aborts `wt list` with the error. The picker
// shares this path but runs collect on a background thread while skim
@@ -1164,10 +1165,12 @@ pub fn collect(
};
if !custom_columns.is_empty() {
let all_vars = repo.all_vars_from_snapshot()?;
+ let all_branch_config = repo.all_branch_config_from_snapshot()?;
super::custom_columns::expand_custom_columns(
&custom_columns,
&mut all_items,
&all_vars,
+ &all_branch_config,
repo,
);
}
diff --git a/src/commands/list/custom_columns.rs b/src/commands/list/custom_columns.rs
index 5e82254b0..8e00dce93 100644
--- a/src/commands/list/custom_columns.rs
+++ b/src/commands/list/custom_columns.rs
@@ -2,15 +2,16 @@
//!
//! Resolved once per invocation, then expanded eagerly for every row before
//! the table skeleton renders: the template inputs (branch, worktree
-//! identity, per-branch vars from the bulk config snapshot) are all in memory
-//! by then, so cells paint with the skeleton, widths are measured from
-//! content like Branch/Path, and a column that renders empty for every row is
-//! dropped by the regular empty-column penalty. Layout participation lives in
-//! `columns.rs` / `layout.rs` via `ColumnKind::Custom`.
+//! identity, per-branch `vars` and `git.branch` config from the bulk config
+//! snapshot) are all in memory by then, so cells paint with the skeleton,
+//! widths are measured from content like Branch/Path, and a column that
+//! renders empty for every row is dropped by the regular empty-column penalty.
+//! Layout participation lives in `columns.rs` / `layout.rs` via
+//! `ColumnKind::Custom`.
//!
//! Expansion stays off the skeleton's critical costs: one minijinja
-//! environment and one parse per column per invocation, one vars conversion
-//! per branch, zero subprocesses per cell.
+//! environment and one parse per column per invocation, one value conversion
+//! per branch per namespace, zero subprocesses per cell.
use std::collections::{BTreeMap, HashMap};
@@ -88,15 +89,18 @@ pub fn resolve_custom_columns(
/// Expand every custom column for every item, populating
/// `ListItem::custom_values`.
///
-/// `all_vars` is the pre-fetched vars map for all branches (one snapshot
-/// read covers the whole table — no subprocess per cell). Render errors
-/// produce an empty cell, logged at `-vv`: a vars key absent on this branch
-/// is the expected sparse-column shape, and config-level typos were already
-/// rejected at resolution.
+/// `all_vars` and `all_branch_config` are the pre-fetched per-branch maps for
+/// the whole table (one snapshot read each — no subprocess per cell); the
+/// former backs `{{ vars.* }}` (worktrunk's `wt config state vars`), the
+/// latter `{{ git.branch.* }}` (the branch's own `branch.<name>.*` git
+/// config). Render errors produce an empty cell, logged at `-vv`: a key absent
+/// on this branch is the expected sparse-column shape, and config-level typos
+/// were already rejected at resolution.
pub fn expand_custom_columns(
columns: &[ResolvedCustomColumn],
items: &mut [ListItem],
all_vars: &HashMap<String, BTreeMap<String, String>>,
+ all_branch_config: &HashMap<String, BTreeMap<String, String>>,
repo: &Repository,
) {
let env = template_environment(repo);
@@ -120,12 +124,17 @@ pub fn expand_custom_columns(
)
.collect();
- // Convert each branch's vars to a template value once, not per cell
+ // Convert each branch's vars / branch_config to a template value once, not
+ // per cell. Both namespaces share the same empty value for unknown branches.
let branch_values: HashMap<&str, Value> = all_vars
.iter()
.map(|(branch, entries)| (branch.as_str(), vars_map_to_value(entries)))
.collect();
- let empty_vars = vars_map_to_value(&BTreeMap::new());
+ let branch_config_values: HashMap<&str, Value> = all_branch_config
+ .iter()
+ .map(|(branch, entries)| (branch.as_str(), vars_map_to_value(entries)))
+ .collect();
+ let empty = vars_map_to_value(&BTreeMap::new());
// Worktree identity is only computed when a template references it:
// `to_posix_path` shells out to cygpath per call on Windows (Git Bash),
@@ -158,10 +167,15 @@ pub fn expand_custom_columns(
context.insert("branch".to_string(), Value::from(branch));
context.insert("worktree_path".to_string(), Value::from(worktree_path));
context.insert("worktree_name".to_string(), Value::from(worktree_name));
- // Always inject vars (empty for unknown branches) so
- // `{{ vars.key | default(...) }}` works in SemiStrict mode
- let vars = branch_values.get(branch).unwrap_or(&empty_vars).clone();
+ // Always inject vars / git.branch (empty for unknown branches) so
+ // `{{ vars.key | default(...) }}` works in SemiStrict mode. The branch's
+ // git config nests under `git` so the namespace can grow (git.remote,
+ // git.upstream, …) without claiming a new top-level name per field.
+ let vars = branch_values.get(branch).unwrap_or(&empty).clone();
context.insert("vars".to_string(), vars);
+ let branch_config = branch_config_values.get(branch).unwrap_or(&empty).clone();
+ let git = Value::from_object(HashMap::from([("branch".to_string(), branch_config)]));
+ context.insert("git".to_string(), git);
let context = Value::from_object(context);
item.custom_values = templates
@@ -322,7 +336,7 @@ mod tests {
ListItem::new_branch("abc12345".to_string(), "feature".to_string()),
ListItem::new_branch("abc12345".to_string(), "other".to_string()),
];
- expand_custom_columns(&columns, &mut items, &all_vars, &test.repo);
+ expand_custom_columns(&columns, &mut items, &all_vars, &HashMap::new(), &test.repo);
// Vars-backed cells are sanitized to one line; nested JSON access
// works; identity vars expand
@@ -333,4 +347,54 @@ mod tests {
// A branch without the vars keys renders empty cells, not errors
assert_eq!(items[1].custom_values, ["", "", "other!"]);
}
+
+ #[test]
+ fn test_expand_custom_columns_git_branch() {
+ let test = TestRepo::new();
+ let columns = vec![
+ ResolvedCustomColumn {
+ name: "Ticket".to_string(),
+ template: "{{ git.branch.jira }}".to_string(),
+ max_width: 40,
+ priority: 9,
+ },
+ // The git-native description is multi-line; `lines | first` keeps
+ // just the summary line the use case wants.
+ ResolvedCustomColumn {
+ name: "Summary".to_string(),
+ template: "{{ git.branch.description | lines | first }}".to_string(),
+ max_width: 40,
+ priority: 9,
+ },
+ ];
+ let mut all_branch_config = HashMap::new();
+ all_branch_config.insert(
+ "feature".to_string(),
+ BTreeMap::from([
+ ("jira".to_string(), "PROJ-1".to_string()),
+ (
+ "description".to_string(),
+ "Add telemetry\n\n# Details\nlong body".to_string(),
+ ),
+ ]),
+ );
+
+ let mut items = vec![
+ ListItem::new_branch("abc12345".to_string(), "feature".to_string()),
+ ListItem::new_branch("abc12345".to_string(), "other".to_string()),
+ ];
+ expand_custom_columns(
+ &columns,
+ &mut items,
+ &HashMap::new(),
+ &all_branch_config,
+ &test.repo,
+ );
+
+ // The branch's own git config surfaces; the multi-line description is
+ // reduced to its first line.
+ assert_eq!(items[0].custom_values, ["PROJ-1", "Add telemetry"]);
+ // A branch without the keys renders empty cells, not errors.
+ assert_eq!(items[1].custom_values, ["", ""]);
+ }
}
diff --git a/src/config/expansion.rs b/src/config/expansion.rs
index 28dfb4720..e0ce6f5e1 100644
--- a/src/config/expansion.rs
+++ b/src/config/expansion.rs
@@ -1073,13 +1073,15 @@ pub fn vars_map_to_value(entries: &std::collections::BTreeMap<String, String>) -
Value::from_serialize(&vars_map)
}
-/// A vars stand-in that answers every key with a placeholder string.
+/// A stand-in for the `vars` / `git.branch` maps that answers every key
+/// with a placeholder string.
///
/// Used only by [`validate_list_column_template`]'s trial render: list
-/// columns lean on `{{ vars.key }}` for keys that only some branches set, so
-/// validating against an empty map (as [`validate_template`] does) would
-/// wrongly reject the dominant pattern. Answering every key keeps the trial
-/// render exercising filters without constraining which keys exist.
+/// columns lean on `{{ vars.key }}` / `{{ git.branch.key }}` for keys that
+/// only some branches set, so validating against an empty map (as
+/// [`validate_template`] does) would wrongly reject the dominant pattern.
+/// Answering every key keeps the trial render exercising filters without
+/// constraining which keys exist.
#[derive(Debug)]
struct AnyKeyVars;
@@ -1092,10 +1094,10 @@ impl Object for AnyKeyVars {
/// Validate a `wt list` custom-column template.
///
/// Checks syntax, restricts top-level variables to `LIST_COLUMN_VARS` ∪
-/// `vars`, and trial-renders with placeholder values so filter errors (e.g. a
-/// misspelled filter name) surface at config resolution rather than as
-/// silently empty cells. `vars.*` access is unconstrained — any key
-/// validates, since which keys exist is per-branch runtime state.
+/// `{vars, git}`, and trial-renders with placeholder values so filter errors
+/// (e.g. a misspelled filter name) surface at config resolution rather than as
+/// silently empty cells. `vars.*` and `git.branch.*` access is unconstrained —
+/// any key validates, since which keys exist is per-branch runtime state.
pub fn validate_list_column_template(
template: &str,
repo: &Repository,
@@ -1109,12 +1111,13 @@ pub fn validate_list_column_template(
let mut undefined: Vec<String> = tmpl
.undeclared_variables(false)
.into_iter()
- .filter(|var| var != "vars" && !LIST_COLUMN_VARS.contains(&var.as_str()))
+ .filter(|var| var != "vars" && var != "git" && !LIST_COLUMN_VARS.contains(&var.as_str()))
.collect();
undefined.sort();
if !undefined.is_empty() {
let mut available = LIST_COLUMN_VARS.to_vec();
available.push("vars.<key>");
+ available.push("git.branch.<key>");
return Err(build_undefined_vars_error(
name,
&undefined,
@@ -1127,6 +1130,10 @@ pub fn validate_list_column_template(
.map(|&k| (k.to_string(), Value::from("PLACEHOLDER")))
.collect();
context.insert("vars".to_string(), Value::from_object(AnyKeyVars));
+ // git.branch.<key> resolves to a placeholder; nesting AnyKeyVars under
+ // `branch` lets the trial render exercise `{{ git.branch.* }}` filters.
+ let git = HashMap::from([("branch".to_string(), Value::from_object(AnyKeyVars))]);
+ context.insert("git".to_string(), Value::from_object(git));
match tmpl.render(Value::from_object(context)) {
Ok(_) => Ok(()),
// At runtime an undefined value renders as an empty cell by design
@@ -3016,13 +3023,27 @@ mod tests {
// Nested vars access (JSON values at runtime) must not be rejected
// even though the trial render's placeholder vars are flat strings
validate_list_column_template("{{ vars.config.port }}", &test.repo, "t").unwrap();
+ // git.branch.* mirrors vars.*: any key validates, filters apply
+ validate_list_column_template("{{ git.branch.jira }}", &test.repo, "t").unwrap();
+ validate_list_column_template(
+ "{{ git.branch.description | lines | first }}",
+ &test.repo,
+ "t",
+ )
+ .unwrap();
- // Unknown top-level variables list the available set, vars included
+ // Unknown top-level variables list the available set, both namespaces included
let err = validate_list_column_template("{{ branhc }}", &test.repo, "t").unwrap_err();
assert!(err.message.contains("branhc"), "got: {}", err.message);
assert_eq!(
err.available_vars,
- vec!["branch", "vars.<key>", "worktree_name", "worktree_path"]
+ vec![
+ "branch",
+ "git.branch.<key>",
+ "vars.<key>",
+ "worktree_name",
+ "worktree_path"
+ ]
);
// Syntax errors fail
diff --git a/src/git/repository/config.rs b/src/git/repository/config.rs
index 627da75e3..c9c8a0045 100644
--- a/src/git/repository/config.rs
+++ b/src/git/repository/config.rs
@@ -197,6 +197,36 @@ impl Repository {
pub fn all_vars_from_snapshot(
&self,
) -> anyhow::Result<std::collections::HashMap<String, std::collections::BTreeMap<String, String>>>
+ {
+ self.subsection_map_from_snapshot(parse_vars_config_key)
+ }
+
+ /// Get all `branch.<name>.*` git config entries across all branches from
+ /// the bulk config snapshot.
+ ///
+ /// Returns a map of branch → (key → value), reading the in-memory
+ /// `Repository::all_config` map with no subprocess (see
+ /// [`Self::all_vars_from_snapshot`] for the snapshot's coherence model).
+ /// This surfaces the keys a user already stores under `branch.<name>.*` in
+ /// git config — both convention keys (`branch.<name>.jira`) and the
+ /// git-native `branch.<name>.description` — in `wt list` custom columns via
+ /// `{{ git.branch.* }}`, the parallel namespace to `{{ vars.* }}`.
+ pub fn all_branch_config_from_snapshot(
+ &self,
+ ) -> anyhow::Result<std::collections::HashMap<String, std::collections::BTreeMap<String, String>>>
+ {
+ self.subsection_map_from_snapshot(parse_branch_config_key)
+ }
+
+ /// Build a branch → (key → value) map from the bulk config snapshot,
+ /// splitting each config key with `parse` (which returns `None` to skip a
+ /// key). Backs both [`Self::all_vars_from_snapshot`] and
+ /// [`Self::all_branch_config_from_snapshot`]. `parse_config_list_z`
+ /// guarantees at least one value per key, so the last one is authoritative.
+ fn subsection_map_from_snapshot(
+ &self,
+ parse: fn(&str) -> Option<(&str, &str)>,
+ ) -> anyhow::Result<std::collections::HashMap<String, std::collections::BTreeMap<String, String>>>
{
let guard = self.all_config()?.read().unwrap();
let mut result: std::collections::HashMap<
@@ -204,16 +234,14 @@ impl Repository {
std::collections::BTreeMap<String, String>,
> = std::collections::HashMap::new();
for (config_key, values) in guard.iter() {
- let Some((branch, key)) = parse_vars_config_key(config_key) else {
- continue;
- };
- let Some(value) = values.last() else {
+ let Some((branch, key)) = parse(config_key) else {
continue;
};
+ let value = values.last().cloned().unwrap_or_default();
result
.entry(branch.to_string())
.or_default()
- .insert(key.to_string(), value.clone());
+ .insert(key.to_string(), value);
}
Ok(result)
}
@@ -692,6 +720,18 @@ fn parse_vars_config_key(config_key: &str) -> Option<(&str, &str)> {
.rsplit_once(".vars.")
}
+/// Split a `branch.<name>.<key>` config key into `(branch, key)`.
+///
+/// Uses `rsplit_once`: git config variable names cannot contain dots, so the
+/// last `.` always separates the branch subsection (which may itself contain
+/// dots or slashes, e.g. `feature.foo`, `feature/bar`) from the variable name.
+/// Git's own section-level branch settings have no subsection and so flatten to
+/// two segments (`branch.sort`, `branch.autoSetupMerge`); after stripping the
+/// prefix they hold no `.`, so `rsplit_once` yields `None` and they're skipped.
+fn parse_branch_config_key(config_key: &str) -> Option<(&str, &str)> {
+ config_key.strip_prefix("branch.")?.rsplit_once('.')
+}
+
#[cfg(test)]
mod tests {
use super::*;
@@ -801,4 +841,34 @@ mod tests {
assert_eq!(snapshot["feature"]["note"], "a note");
assert_eq!(snapshot["weird.vars.branch"]["key"], "v");
}
+
+ /// The branch-config snapshot read parses `branch.<name>.*` keys, including
+ /// branch names containing dots or slashes, lowercases the variable name as
+ /// git does, and skips git's own section-level branch settings
+ /// (`branch.sort`), which have no subsection.
+ #[test]
+ fn test_all_branch_config_from_snapshot() {
+ let test = TestRepo::with_initial_commit();
+ let repo = Repository::at(test.root_path()).unwrap();
+
+ repo.set_config("branch.feature.jira", "PROJ-1").unwrap();
+ repo.set_config("branch.feature.foo.jira", "PROJ-2")
+ .unwrap();
+ repo.set_config("branch.feature/bar.jira", "PROJ-3")
+ .unwrap();
+ // Git lowercases variable names; a mixed-case key reads back lowered.
+ repo.set_config("branch.feature.nvciShelf", "64645277")
+ .unwrap();
+ // A section-level branch setting has no subsection — must be skipped.
+ repo.set_config("branch.sort", "-committerdate").unwrap();
+
+ let snapshot = repo.all_branch_config_from_snapshot().unwrap();
+ assert_eq!(snapshot["feature"]["jira"], "PROJ-1");
+ assert_eq!(snapshot["feature"]["nvcishelf"], "64645277");
+ // Dotted and slashed branch names keep their full subsection.
+ assert_eq!(snapshot["feature.foo"]["jira"], "PROJ-2");
+ assert_eq!(snapshot["feature/bar"]["jira"], "PROJ-3");
+ // `branch.sort` is git's own key, not a per-branch entry.
+ assert!(!snapshot.contains_key("sort"));
+ }
}
diff --git a/tests/integration_tests/list_config.rs b/tests/integration_tests/list_config.rs
index 8e4e78e5d..a00cfa184 100644
--- a/tests/integration_tests/list_config.rs
+++ b/tests/integration_tests/list_config.rs
@@ -827,6 +827,51 @@ template = "{{ vars.ticket }}"
assert!(by_branch("feature-b")["columns"].is_null());
}
+#[rstest]
+fn test_list_custom_column_git_branch(repo: TestRepo) {
+ // A git.branch column reads the branch's own git config under
+ // branch.<name>.*, with no `wt config state vars set` round-trip. The
+ // git-native multi-line description is reduced to its first line.
+ fs::write(
+ repo.test_config_path(),
+ r#"[list.custom-columns.Jira]
+template = "{{ git.branch.jira }}"
+
+[list.custom-columns.Summary]
+template = "{{ git.branch.description | lines | first }}"
+"#,
+ )
+ .unwrap();
+ repo.run_git(&["config", "branch.feature-a.jira", "HWINFCI-2810"]);
+ repo.run_git(&[
+ "config",
+ "branch.feature-a.description",
+ "Add telemetry\n\n# Description\nlong body",
+ ]);
+
+ let mut cmd = wt_command();
+ repo.configure_wt_cmd(&mut cmd);
+ cmd.args(["list", "--format=json"])
+ .current_dir(repo.root_path());
+
+ let output = cmd.output().unwrap();
+ let stdout = String::from_utf8_lossy(&output.stdout);
+ let json: serde_json::Value = serde_json::from_str(&stdout).unwrap();
+ let items = json.as_array().unwrap();
+ let by_branch = |branch: &str| {
+ items
+ .iter()
+ .find(|item| item["branch"] == branch)
+ .unwrap_or_else(|| panic!("no item for branch {branch}"))
+ };
+
+ let cols = &by_branch("feature-a")["columns"];
+ assert_eq!(cols["Jira"], "HWINFCI-2810");
+ assert_eq!(cols["Summary"], "Add telemetry");
+ // A branch without any branch.<name>.* keys renders no custom cells.
+ assert!(by_branch("feature-b")["columns"].is_null());
+}
+
#[rstest]
fn test_list_custom_column_invalid_template(repo: TestRepo) {
// An unknown top-level variable aborts wt list with the available set
diff --git a/tests/snapshots/integration__integration_tests__help__help_config_create.snap b/tests/snapshots/integration__integration_tests__help__help_config_create.snap
index 7db60d636..b9f3d63c4 100644
--- a/tests/snapshots/integration__integration_tests__help__help_config_create.snap
+++ b/tests/snapshots/integration__integration_tests__help__help_config_create.snap
@@ -220,19 +220,29 @@ Creates ~/.config/worktrunk/config.toml with the following content:
  # # lower = kept longer (default: 9, the URL band)
  #
  # Templates may reference `{{ branch }}`, `{{ worktree_path }}`,
-  # `{{ worktree_name }}` (empty for branch-only rows), and `{{ vars.* }}` —
-  # per-branch values stored with
-  # `wt config state vars set` (https://worktrunk.dev/config/#wt-config-state-vars).
+  # `{{ worktree_name }}` (empty for branch-only rows), and two per-branch
+  # namespaces:
+  #
+  # - `{{ vars.* }}` — values stored with
+  # `wt config state vars set` (https://worktrunk.dev/config/#wt-config-state-vars).
+  # - `{{ git.branch.* }}` — the branch's own git config under `branch.<name>.*`,
+  # read straight from `git config` (e.g. `{{ git.branch.jira }}` for a key you
+  # set yourself, or the git-native `description`). Git lowercases config variable
+  # names, so `branch.<name>.nvciShelf` reads as `{{ git.branch.nvcishelf }}`.
+  #
  # All standard filters work (`sanitize`, `hash_port`, `codename`, …). A row
-  # where the template renders empty (e.g. a branch without the vars key) shows an
+  # where the template renders empty (e.g. a branch without the key) shows an
  # empty cell; a column that is empty for every row is dropped from the table.
  # `wt list --format json` includes the rendered values under `columns`.
  #
-  # A `Note` column showing free-form descriptions, set per branch with
-  # `wt config state vars set note "Bug fix for production fire"`:
+  # A `Jira` column reading a key kept in git config, and a `Summary` column
+  # showing just the first line of the git-native branch description:
+  #
+  # [list.custom-columns.Jira]
+  # template = "{{ git.branch.jira }}"
  #
-  # [list.custom-columns.Note]
-  # template = "{{ vars.note }}"
+  # [list.custom-columns.Summary]
+  # template = "{{ git.branch.description | lines | first }}"
  #
  # ### Commit
  #
diff --git a/tests/snapshots/integration__integration_tests__help__help_config_long.snap b/tests/snapshots/integration__integration_tests__help__help_config_long.snap
index e60dfa98b..3d6dc850d 100644
--- a/tests/snapshots/integration__integration_tests__help__help_config_long.snap
+++ b/tests/snapshots/integration__integration_tests__help__help_config_long.snap
@@ -268,19 +268,29 @@ renders each row's cell.
   # lower = kept longer (default: 9, the URL band)
Templates may reference {{ branch }}, {{ worktree_path }},
-{{ worktree_name }} (empty for branch-only rows), and {{ vars.* }} —
-per-branch values stored with
-wt config state vars set.
+{{ worktree_name }} (empty for branch-only rows), and two per-branch
+namespaces:
+
+- {{ vars.* }} — values stored with
+ wt config state vars set.
+- {{ git.branch.* }} — the branch's own git config under branch.<name>.*,
+ read straight from git config (e.g. {{ git.branch.jira }} for a key you
+ set yourself, or the git-native description). Git lowercases config variable
+ names, so branch.<name>.nvciShelf reads as {{ git.branch.nvcishelf }}.
+
All standard filters work (sanitize, hash_port, codename, …). A row
-where the template renders empty (e.g. a branch without the vars key) shows an
+where the template renders empty (e.g. a branch without the key) shows an
empty cell; a column that is empty for every row is dropped from the table.
wt list --format json includes the rendered values under columns.
-A Note column showing free-form descriptions, set per branch with
-wt config state vars set note "Bug fix for production fire":
+A Jira column reading a key kept in git config, and a Summary column
+showing just the first line of the git-native branch description:
-  [list.custom-columns.Note]
-  template = "{{ vars.note }}"
+  [list.custom-columns.Jira]
+  template = "{{ git.branch.jira }}"
+ 
+  [list.custom-columns.Summary]
+  template = "{{ git.branch.description | lines | first }}"
Commit
diff --git a/tests/snapshots/integration__integration_tests__help__help_list_long.snap b/tests/snapshots/integration__integration_tests__help__help_list_long.snap
index 42405d12b..af7479155 100644
--- a/tests/snapshots/integration__integration_tests__help__help_list_long.snap
+++ b/tests/snapshots/integration__integration_tests__help__help_list_long.snap
@@ -193,7 +193,7 @@ Reuses the commit.generation command — the same LLM that generates com
Custom columns [experimental]
-Each [list.custom-columns] entry in user config adds a column: the key is the header, the template renders each row's cell. Templates can reference per-branch {{ vars.* }} stored with wt config state vars set — useful for tracking what each of many (often agent-driven) branches is for:
+Each [list.custom-columns] entry in user config adds a column: the key is the header, the template renders each row's cell. Templates read two per-branch namespaces — {{ vars.* }}, stored with wt config state vars set, and {{ git.branch.* }}, the branch's own git config under branch.<name>.* (a jira key you set yourself, or the git-native description) — useful for tracking what each of many (often agent-driven) branches is for:
  [list.custom-columns.Ticket]
  template = "{{ vars.ticket }}"
diff --git a/tests/snapshots/integration__integration_tests__help__help_list_narrow_80.snap b/tests/snapshots/integration__integration_tests__help__help_list_narrow_80.snap
index 8dce57bce..befc89d3e 100644
--- a/tests/snapshots/integration__integration_tests__help__help_list_narrow_80.snap
+++ b/tests/snapshots/integration__integration_tests__help__help_list_narrow_80.snap
@@ -231,9 +231,11 @@ are cached until the branch's diff changes.
Custom columns [experimental]
Each [list.custom-columns] entry in user config adds a column: the key is the
-header, the template renders each row's cell. Templates can reference per-branch
- {{ vars.* }} stored with wt config state vars set — useful for tracking what
-each of many (often agent-driven) branches is for:
+header, the template renders each row's cell. Templates read two per-branch
+namespaces — {{ vars.* }}, stored with wt config state vars set, and {{
+git.branch.* }}, the branch's own git config under branch.<name>.* (a jira key
+you set yourself, or the git-native description) — useful for tracking what each
+ of many (often agent-driven) branches is for:
  [list.custom-columns.Ticket]
  template = "{{ vars.ticket }}"
diff --git a/tests/snapshots/integration__integration_tests__list_config__list_custom_column_invalid_template.snap b/tests/snapshots/integration__integration_tests__list_config__list_custom_column_invalid_template.snap
index b7412f7ec..35d25e11e 100644
--- a/tests/snapshots/integration__integration_tests__list_config__list_custom_column_invalid_template.snap
+++ b/tests/snapshots/integration__integration_tests__list_config__list_custom_column_invalid_template.snap
@@ -6,6 +6,7 @@ info:
- list
env:
APPDATA: "[TEST_CONFIG_HOME]"
+ CLAUDE_CONFIG_DIR: "[TEST_CLAUDE_CONFIG]"
CLICOLOR_FORCE: "1"
COLUMNS: "500"
GIT_AUTHOR_DATE: "2025-01-01T00:00:00Z"
@@ -46,4 +47,4 @@ exit_code: 1
----- stderr -----
✗ Failed to expand list.custom-columns.Ticket: undefined value: `branhc`
-↳ Available variables: branch, vars.<key>, worktree_name, worktree_path
+↳ Available variables: branch, git.branch.<key>, vars.<key>, worktree_name, worktree_path