feat(init): add ctrl-n keybinding to create new worktree from fzf picker#141
Merged
feat(init): add ctrl-n keybinding to create new worktree from fzf picker#141
Conversation
WalkthroughAdds ctrl-n support to the interactive picker across Bash, Zsh, and Fish: picker exposes a ctrl-n:new action, scripts prompt for a branch name when triggered, invoke Changes
Sequence DiagramsequenceDiagram
actor User
participant Shell as "Shell\n(runner)"
participant fzf as "fzf\n(picker)"
participant git_gtr as "git gtr\n(new)"
User->>Shell: run interactive command
Shell->>fzf: launch picker with --expect=ctrl-n (header includes "ctrl-n:new")
User->>fzf: press ctrl-n
fzf-->>Shell: return "ctrl-n" selection
Shell->>User: prompt "Branch name: "
User->>Shell: enter branch name
Shell->>git_gtr: run `git gtr new "<branch>"`
git_gtr-->>Shell: exit status
Shell->>User: display result
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~22 minutes Poem
🚥 Pre-merge checks | ✅ 2 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches
🧪 Generate unit tests (beta)
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
🧹 Nitpick comments (2)
lib/commands/init.sh (1)
115-117: Prefer|| return $?over barereturn $?to satisfy theset -eguard guideline.
command git gtr new "$_gtr_branch"followed byreturn $?on the next line is dead code whenset -eis active—ifgit gtr newexits non-zero, the function exits before thereturn. The|| return $?form makes the propagation explicit and safe regardless ofset -estate.♻️ Proposed fix (bash)
- command git gtr new "$_gtr_branch" - return $? + command git gtr new "$_gtr_branch" || return $? + return 0The same pattern applies to the zsh block (lines 233–234).
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@lib/commands/init.sh` around lines 115 - 117, Replace the pattern that runs the git gtr command then returns its exit code with an explicit short-circuit so failures propagate under set -e: change the block using the variable _gtr_branch and the call command git gtr new "$_gtr_branch" followed by return $? to a single line command git gtr new "$_gtr_branch" || return $?; apply the same replacement in the zsh block that likewise invokes command git gtr new "$_gtr_branch".tests/init.bats (1)
208-280: New ctrl-n test block LGTM; consider adding split-logic coverage.The 12 tests correctly verify structural presence of
--expect=ctrl-n, the header label, thegit gtr newinvocation, and the prompt string across all three shells. The fish assertion at line 261 (*'git gtr new "$_gtr_branch"'*) correctly matches the actualcommand git gtr new "$_gtr_branch"via the leading*wildcard.One minor gap: the output-parsing lines (
head -1/sed -n '2p'for bash/zsh,$_gtr_selection[1]/$_gtr_selection[2]for fish) that implement the two-line fzf split are untested. Given that the rest of the test file uses the same static-pattern style, adding a couple of assertions would catch any future regression in the parsing logic:+@test "bash output splits fzf output into key and line" { + run cmd_init bash + [ "$status" -eq 0 ] + [[ "$output" == *'head -1 <<< "$_gtr_selection"'* ]] + [[ "$output" == *"sed -n '2p' <<< \"\$_gtr_selection\""* ]] +} + +@test "fish output splits fzf output into key and line" { + run cmd_init fish + [ "$status" -eq 0 ] + [[ "$output" == *'_gtr_selection[1]'* ]] + [[ "$output" == *'_gtr_selection[2]'* ]] +}🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@tests/init.bats` around lines 208 - 280, Add tests exercising the two-line fzf split/parse logic so regressions in the selection parsing are caught: extend the new ctrl-n test block to assert that for bash/zsh the parsing lines (head -1 and sed -n '2p') result in the expected branch and description (e.g. check the first line appears where head -1 is used and the second line where sed -n '2p' is used), and for fish assert $_gtr_selection[1] holds the branch and $_gtr_selection[2] holds the description; reference the existing tests invoking cmd_init and the variables/_parsing expressions ($_gtr_selection, head -1, sed -n '2p') so the new assertions match the existing output-check style.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Nitpick comments:
In `@lib/commands/init.sh`:
- Around line 115-117: Replace the pattern that runs the git gtr command then
returns its exit code with an explicit short-circuit so failures propagate under
set -e: change the block using the variable _gtr_branch and the call command git
gtr new "$_gtr_branch" followed by return $? to a single line command git gtr
new "$_gtr_branch" || return $?; apply the same replacement in the zsh block
that likewise invokes command git gtr new "$_gtr_branch".
In `@tests/init.bats`:
- Around line 208-280: Add tests exercising the two-line fzf split/parse logic
so regressions in the selection parsing are caught: extend the new ctrl-n test
block to assert that for bash/zsh the parsing lines (head -1 and sed -n '2p')
result in the expected branch and description (e.g. check the first line appears
where head -1 is used and the second line where sed -n '2p' is used), and for
fish assert $_gtr_selection[1] holds the branch and $_gtr_selection[2] holds the
description; reference the existing tests invoking cmd_init and the
variables/_parsing expressions ($_gtr_selection, head -1, sed -n '2p') so the
new assertions match the existing output-check style.
ℹ️ Review info
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Disabled knowledge base sources:
- Linear integration is disabled
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (2)
lib/commands/init.shtests/init.bats
helizaga
force-pushed
the
new-branch-in-fzf
branch
from
2c6e5de to
3202292
Compare
helizaga
force-pushed
the
new-branch-in-fzf
branch
from
3202292 to
b132162
Compare
There was a problem hiding this comment.
Actionable comments posted: 2
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
lib/commands/init.sh (1)
360-372:⚠️ Potential issue | 🟠 MajorFix Fish key parsing to handle special keys in single-line
--expectoutput.When
count $_gtr_selectionis1, the code assumes Enter and sets_gtr_keyto empty. However, if a special key likectrl-nis pressed without a selection (or if fzf outputs only the key), Fish command substitution may collapse it to a single item, causing the key to be misplaced into_gtr_line. This bypasses the checks forctrl-n,ctrl-a, andctrl-eat lines 367–380.Suggested patch
+ set -l _gtr_key "" + set -l _gtr_line "" if test (count $_gtr_selection) -eq 1 - set -l _gtr_key "" - set -l _gtr_line "$_gtr_selection[1]" + if contains -- "$_gtr_selection[1]" ctrl-n ctrl-a ctrl-e + set _gtr_key "$_gtr_selection[1]" + else + set _gtr_line "$_gtr_selection[1]" + end else - set -l _gtr_key "$_gtr_selection[1]" - set -l _gtr_line "$_gtr_selection[2]" + set _gtr_key "$_gtr_selection[1]" + set _gtr_line "$_gtr_selection[2]" end🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@lib/commands/init.sh` around lines 360 - 372, When parsing fzf output in init.sh, the branch that handles count $_gtr_selection -eq 1 incorrectly assumes the single item is always the selected line; instead detect if that single item is a special key and populate _gtr_key (and clear _gtr_line) or otherwise populate _gtr_line (and clear _gtr_key). Update the block that sets _gtr_key/_gtr_line so that when count $_gtr_selection is 1 you test the value of $_gtr_selection[1] against the known special keys (e.g., "ctrl-n","ctrl-a","ctrl-e" or a pattern like "ctrl-") and set _gtr_key="that_value" and _gtr_line="" for keys, otherwise set _gtr_key="" and _gtr_line="that_value"; keep the existing downstream checks that compare $_gtr_key to "ctrl-n" etc.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@lib/commands/init.sh`:
- Around line 110-116: The unguarded read -r _gtr_branch can cause the script to
exit under set -e when the user sends EOF (Ctrl-D); update both occurrences of
read -r _gtr_branch so EOF doesn't propagate a nonzero exit: either guard the
call (e.g. use "read -r _gtr_branch || true" or an if/! read check) and treat
EOF as an empty branch name, preserving the existing empty-check and
returning/canceling gracefully before invoking command git gtr new; ensure you
change both occurrences that reference _gtr_branch.
In `@tests/init.bats`:
- Around line 411-483: Tests that assert the exact fzf --expect string are stale
(they expect "--expect=ctrl-a,ctrl-e") and must accept the new
"--expect=ctrl-n,ctrl-a,ctrl-e" ordering; locate assertions in tests/init.bats
that check for the exact "--expect=ctrl-a,ctrl-e" (look for the pattern string
used with run cmd_init) and change them to assert presence of the individual
keys instead (e.g. assert output contains "ctrl-a" and contains "ctrl-e" or
assert output contains "--expect=" and contains "ctrl-a" and "ctrl-e"), or use a
glob/regex that allows additional keys like "ctrl-n" to be present.
---
Outside diff comments:
In `@lib/commands/init.sh`:
- Around line 360-372: When parsing fzf output in init.sh, the branch that
handles count $_gtr_selection -eq 1 incorrectly assumes the single item is
always the selected line; instead detect if that single item is a special key
and populate _gtr_key (and clear _gtr_line) or otherwise populate _gtr_line (and
clear _gtr_key). Update the block that sets _gtr_key/_gtr_line so that when
count $_gtr_selection is 1 you test the value of $_gtr_selection[1] against the
known special keys (e.g., "ctrl-n","ctrl-a","ctrl-e" or a pattern like "ctrl-")
and set _gtr_key="that_value" and _gtr_line="" for keys, otherwise set
_gtr_key="" and _gtr_line="that_value"; keep the existing downstream checks that
compare $_gtr_key to "ctrl-n" etc.
ℹ️ Review info
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Disabled knowledge base sources:
- Linear integration is disabled
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (2)
lib/commands/init.shtests/init.bats
There was a problem hiding this comment.
Actionable comments posted: 1
♻️ Duplicate comments (1)
lib/commands/init.sh (1)
233-239:⚠️ Potential issue | 🟡 MinorSame unguarded
readissue in Zsh handler.Consistent with the Bash handler, the
read -r _gtr_branchat line 235 should be guarded to handle EOF gracefully.Proposed fix
if [ "$_gtr_key" = "ctrl-n" ]; then printf "Branch name: " >&2 - read -r _gtr_branch + read -r _gtr_branch || true [ -z "$_gtr_branch" ] && return 0 command git gtr new "$_gtr_branch" return $? fi🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@lib/commands/init.sh` around lines 233 - 239, The Zsh key handler uses an unguarded read for _gtr_branch (when _gtr_key == "ctrl-n") which can fail on EOF; modify the block around the read -r _gtr_branch so the read is checked and handles failure gracefully (e.g., use a conditional/read return check such as if ! read -r _gtr_branch; then return 0; fi) before calling command git gtr new "$_gtr_branch", keeping the existing early-return when _gtr_branch is empty.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@lib/commands/init.sh`:
- Around line 92-116: The branch-name prompt's unguarded read (read -r
_gtr_branch) can fail on EOF/ctrl-D and, under set -e, abort before the
subsequent empty-string check; update the ctrl-n branch handling to check the
return status of read -r _gtr_branch and return 0 if the read fails (EOF) before
proceeding to the [ -z "$_gtr_branch" ] check, so that functions like git gtr
new are only called when a valid branch name is obtained.
---
Duplicate comments:
In `@lib/commands/init.sh`:
- Around line 233-239: The Zsh key handler uses an unguarded read for
_gtr_branch (when _gtr_key == "ctrl-n") which can fail on EOF; modify the block
around the read -r _gtr_branch so the read is checked and handles failure
gracefully (e.g., use a conditional/read return check such as if ! read -r
_gtr_branch; then return 0; fi) before calling command git gtr new
"$_gtr_branch", keeping the existing early-return when _gtr_branch is empty.
ℹ️ Review info
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Disabled knowledge base sources:
- Linear integration is disabled
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (2)
lib/commands/init.shtests/init.bats
Comment on lines
+92
to
+116
| local _gtr_selection _gtr_key _gtr_line _gtr_branch | ||
| _gtr_selection="$(printf '%s\n' "$_gtr_porcelain" | fzf \ | ||
| --delimiter=$'\t' \ | ||
| --with-nth=2 \ | ||
| --ansi \ | ||
| --layout=reverse \ | ||
| --border \ | ||
| --prompt='Worktree> ' \ | ||
| --header='enter:cd │ ctrl-e:editor │ ctrl-a:ai │ ctrl-d:delete │ ctrl-y:copy │ ctrl-r:refresh' \ | ||
| --header='enter:cd │ ctrl-n:new │ ctrl-e:editor │ ctrl-a:ai │ ctrl-d:delete │ ctrl-y:copy │ ctrl-r:refresh' \ | ||
| --expect=ctrl-n,ctrl-a,ctrl-e \ | ||
| --preview='git -C {1} log --oneline --graph --color=always -15 2>/dev/null; echo "---"; git -C {1} status --short 2>/dev/null' \ | ||
| --preview-window=right:50% \ | ||
| --expect=ctrl-a,ctrl-e \ | ||
| --bind='ctrl-d:execute(git gtr rm {2} > /dev/tty 2>&1 < /dev/tty)+reload(git gtr list --porcelain)' \ | ||
| --bind='ctrl-y:execute(git gtr copy {2} > /dev/tty 2>&1 < /dev/tty)' \ | ||
| --bind='ctrl-r:reload(git gtr list --porcelain)')" || return 0 | ||
| [ -z "$_gtr_selection" ] && return 0 | ||
| _gtr_key="$(head -1 <<< "$_gtr_selection")" | ||
| _gtr_line="$(sed -n '2p' <<< "$_gtr_selection")" | ||
| if [ "$_gtr_key" = "ctrl-n" ]; then | ||
| printf "Branch name: " >&2 | ||
| read -r _gtr_branch | ||
| [ -z "$_gtr_branch" ] && return 0 | ||
| command git gtr new "$_gtr_branch" | ||
| return $? | ||
| fi |
There was a problem hiding this comment.
Guard read to handle EOF (Ctrl-D) gracefully.
The read -r _gtr_branch at line 112 is unguarded. If a user's interactive shell has set -e active and they press Ctrl-D (EOF) instead of entering a branch name, the read command will return non-zero and the function will exit before reaching the empty-string check on line 113. Although interactive shells typically don't have set -e active, defensive coding would guard this.
Note: A previous review flagged this same issue and it was marked as addressed, but the current code does not reflect the fix.
Proposed fix
if [ "$_gtr_key" = "ctrl-n" ]; then
printf "Branch name: " >&2
- read -r _gtr_branch
+ read -r _gtr_branch || true
[ -z "$_gtr_branch" ] && return 0
command git gtr new "$_gtr_branch"
return $?
fi📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| local _gtr_selection _gtr_key _gtr_line _gtr_branch | |
| _gtr_selection="$(printf '%s\n' "$_gtr_porcelain" | fzf \ | |
| --delimiter=$'\t' \ | |
| --with-nth=2 \ | |
| --ansi \ | |
| --layout=reverse \ | |
| --border \ | |
| --prompt='Worktree> ' \ | |
| --header='enter:cd │ ctrl-e:editor │ ctrl-a:ai │ ctrl-d:delete │ ctrl-y:copy │ ctrl-r:refresh' \ | |
| --header='enter:cd │ ctrl-n:new │ ctrl-e:editor │ ctrl-a:ai │ ctrl-d:delete │ ctrl-y:copy │ ctrl-r:refresh' \ | |
| --expect=ctrl-n,ctrl-a,ctrl-e \ | |
| --preview='git -C {1} log --oneline --graph --color=always -15 2>/dev/null; echo "---"; git -C {1} status --short 2>/dev/null' \ | |
| --preview-window=right:50% \ | |
| --expect=ctrl-a,ctrl-e \ | |
| --bind='ctrl-d:execute(git gtr rm {2} > /dev/tty 2>&1 < /dev/tty)+reload(git gtr list --porcelain)' \ | |
| --bind='ctrl-y:execute(git gtr copy {2} > /dev/tty 2>&1 < /dev/tty)' \ | |
| --bind='ctrl-r:reload(git gtr list --porcelain)')" || return 0 | |
| [ -z "$_gtr_selection" ] && return 0 | |
| _gtr_key="$(head -1 <<< "$_gtr_selection")" | |
| _gtr_line="$(sed -n '2p' <<< "$_gtr_selection")" | |
| if [ "$_gtr_key" = "ctrl-n" ]; then | |
| printf "Branch name: " >&2 | |
| read -r _gtr_branch | |
| [ -z "$_gtr_branch" ] && return 0 | |
| command git gtr new "$_gtr_branch" | |
| return $? | |
| fi | |
| local _gtr_selection _gtr_key _gtr_line _gtr_branch | |
| _gtr_selection="$(printf '%s\n' "$_gtr_porcelain" | fzf \ | |
| --delimiter=$'\t' \ | |
| --with-nth=2 \ | |
| --ansi \ | |
| --layout=reverse \ | |
| --border \ | |
| --prompt='Worktree> ' \ | |
| --header='enter:cd │ ctrl-n:new │ ctrl-e:editor │ ctrl-a:ai │ ctrl-d:delete │ ctrl-y:copy │ ctrl-r:refresh' \ | |
| --expect=ctrl-n,ctrl-a,ctrl-e \ | |
| --preview='git -C {1} log --oneline --graph --color=always -15 2>/dev/null; echo "---"; git -C {1} status --short 2>/dev/null' \ | |
| --preview-window=right:50% \ | |
| --bind='ctrl-d:execute(git gtr rm {2} > /dev/tty 2>&1 < /dev/tty)+reload(git gtr list --porcelain)' \ | |
| --bind='ctrl-y:execute(git gtr copy {2} > /dev/tty 2>&1 < /dev/tty)' \ | |
| --bind='ctrl-r:reload(git gtr list --porcelain)')" || return 0 | |
| [ -z "$_gtr_selection" ] && return 0 | |
| _gtr_key="$(head -1 <<< "$_gtr_selection")" | |
| _gtr_line="$(sed -n '2p' <<< "$_gtr_selection")" | |
| if [ "$_gtr_key" = "ctrl-n" ]; then | |
| printf "Branch name: " >&2 | |
| read -r _gtr_branch || true | |
| [ -z "$_gtr_branch" ] && return 0 | |
| command git gtr new "$_gtr_branch" | |
| return $? | |
| fi |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@lib/commands/init.sh` around lines 92 - 116, The branch-name prompt's
unguarded read (read -r _gtr_branch) can fail on EOF/ctrl-D and, under set -e,
abort before the subsequent empty-string check; update the ctrl-n branch
handling to check the return status of read -r _gtr_branch and return 0 if the
read fails (EOF) before proceeding to the [ -z "$_gtr_branch" ] check, so that
functions like git gtr new are only called when a valid branch name is obtained.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Pull Request
Description
Motivation
Fixes # (issue)
Type of Change
Testing
Manual Testing Checklist
Tested on:
Core functionality tested:
git gtr new <branch>- Create worktreegit gtr go <branch>- Navigate to worktreegit gtr editor <branch>- Open in editor (if applicable)git gtr ai <branch>- Start AI tool (if applicable)git gtr rm <branch>- Remove worktreegit gtr list- List worktreesgit gtr config- Configuration commands (if applicable)Test Steps
Expected behavior:
Actual behavior:
Breaking Changes
Checklist
Before submitting this PR, please check:
git gtr(production) and./bin/gtr(development)Additional Context
License Acknowledgment
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache License 2.0.
Summary by CodeRabbit
New Features
Tests