We set GIT_OPTIONAL_LOCKS=0 for every git command we run. That env var
only affects `git status`: it tells git not to take the optional lock it
would otherwise use to write the index back after refreshing the cached
stat information. The intent was to avoid contending for index.lock with
git commands the user runs in a terminal.
The downside is that our `git status` never persists the refreshed
stat-cache. So whenever the working tree's cached stat info goes stale
(e.g. editing files and discarding the changes, or a checkout), every
subsequent status re-hashes the affected files to confirm they're clean,
and stays slow until something else writes the index (such as the user
running `git status` in a terminal).
Fix this by only suppressing optional locks for refreshes that run
unattended in the background; foreground refreshes triggered by a user
action now run a plain `git status` that writes the refreshed index back,
just like the command line does. Background refreshes keep passing
--no-optional-locks so they still can't cause lock contention.
RefreshOptions gains a Background flag that the background routines set,
threaded down to the status command.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add a 2-second background poll that calls Status.RefsSnapshot and
compares against the snapshot stored at the end of the last refs-
touching refresh. On a diff, trigger a full refresh — same scope as the
focus-in handler, because once we know something changed externally
we can't be sure what (an agent might have created a worktree or
stashed something alongside the commit we detected).
Refresh runs in SYNC mode because goEvery already serializes iterations
via <-done: a slow refresh delays the next tick naturally instead of
letting work stack. The post-refresh hook from the previous commit
updates the snapshot, so in-app commands don't cause the next poll to
spuriously re-fire.
Disabled in the integration test config, like autoRefresh and autoFetch,
because demo replays make repo changes throughout the run; at 2-second
cadence the resulting full refreshes compete with the demo's own
choreography and push some demos past their 40-second timeout.
Also list the two new config keys in checkForChangedConfigsThatDontAutoReload
so a config edit warns the user that lazygit needs a restart.
Add the storage and snapshot-update half of the external-change-detection
mechanism. RefreshHelper now keeps a mutex-protected snapshot string and exposes
accessors for it; Refresh captures a fresh snapshot at the start of any refresh
whose scope set includes COMMITS or BRANCHES.
We capture before reading the git state, not after. Capturing after would let an
external change that lands between the git state read and the snapshot (say, the
next step of a rebase running in another terminal) leave the stored snapshot
newer than what we actually rendered; the poller would then see no difference
and never refresh again, stranding the UI on the intermediate state. Capturing
first keeps the snapshot from running ahead of the render, so if disk moves
during the refresh the next poll catches it.
No reader of the snapshot exists yet — the polling goroutine that consumes it
comes in a later commit. Keeping the snapshot hook in its own commit isolates
the invariant that the snapshot stays in sync with what the UI has observed,
which is what makes the poller's change-detection predicate work across in-app
commands and focus-in refreshes.
Several downstream conditions in Refresh() relied on multi-scope predicates to
express "if X is in scope, Y also needs refreshing". This makes it hard to add
new code that needs to ask "does this refresh re-read refs?", because the answer
involves mirroring one of those predicates and keeping them in sync forever.
Expand the co-refreshing relationships once, up front, right after the scope set
is built. The downstream conditions then collapse to single-scope checks against
the (now-expanded) set. Behavior is preserved.
Two of the scattered multi-scope conditions are intentionally left as-is because
they express subsumption rather than co-refresh (one branch already does the
work of another internally — expanding would cause double-refresh), and one
expresses mid-function coupling on a flag set inside the COMMITS/BRANCHES block.
Several commands (rewording or amending an earlier commit, custom patch
operations, etc.) are implemented by starting an interactive rebase that stops
at a commit, amending it, and continuing. When no conflict occurs, the user
isn't meant to notice a rebase happened at all.
But a background file refresh can fire while the rebase is mid-flight and render
a dirty working copy of whatever the behind-the-scenes rebase is doing (e.g.
applying a custom patch).
To fix this, we pause the background routines for the duration of any
waiting-status operation — exactly the window in which lazygit is driving the
git operation itself and will refresh once at the end. The boundary is also
right for the conflict case: when a rebase stops on a conflict the operation
returns, the pause releases, and background refreshes resume for the interactive
resolution that follows.
Replace the pauseBackgroundRefreshes bool with a count. The single existing
caller (subprocess suspend/resume) is unaffected, but we're about to add a
second, independent reason to pause — lazygit driving a git operation that the
background routines would otherwise catch mid-flight — and the two scopes can
overlap. A bool can't represent "two things both want refreshes paused"; a count
can.
With more than a couple of pagers, having to cycle forward through all
of them to reach the previous one (or to back out of an accidental press
of `|`) is tedious. Add a second binding that cycles backward.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A reverse-cycle handler is about to need the same re-render-and-toast
logic. Pull it out first so the behavior change that follows only has to
swap the cycle direction.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
When cycling pagers, "Selected pager 2 of 3" gives no clue which pager
you landed on; with several configured you have to remember the order.
Include the pager's name in the toast instead.
The name is normally derived from the first word of the pager command,
but that isn't always enough: two entries can share a command but differ
in options (e.g. "delta" and "delta --side-by-side"), and an entry may
have no command at all (the default entry, or when using
useExternalDiffGitConfig). So add an optional `name` field that
overrides the derived name.
The message was also hardcoded in English; localize it while we're here.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A submodule that only has dirty or untracked content (no new commit) can't
be staged from the parent repo, but it still shows up as having unstaged
changes. Pressing stage on it therefore briefly flashed as staged and then
reverted, without explaining why nothing was staged.
Detect this case (via `git submodule status`, where a '+' prefix marks a
stageable commit change) in the shared stage/unstage decision: if the only
thing that looks stageable is such a submodule, don't try to stage it.
Instead unstage if there's anything staged to unstage, so the toggle stays
symmetric; otherwise show an error explaining that there's nothing to stage.
Because the decision is shared, this covers both the stage (space) and
stage-all (a) keybindings.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This map only feeds the optimistic rendering that makes staging feel
instant; it doesn't affect the eventual status, which git reports after
the refresh. The "MM" entry can never be reached for a regular file: a
file at "MM" has stageable unstaged changes, so pressing space stages it
rather than unstaging, and the unstage path is where this map is used. The
only thing that reaches the unstage path at "MM" is a submodule whose
commit is staged on top of dirty content, so this entry exists purely to
update that submodule instantly instead of waiting for the next git
status.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The stage/unstage toggle decides what to do based on whether a node has
unstaged changes: if it does, it stages; otherwise it unstages. For a
submodule this breaks down, because dirty or untracked content inside the
submodule always reports as an unstaged change in the parent repo but can
never be staged from there. Once such a submodule's commit pointer is
staged it sits at "MM", and every subsequent press keeps trying to stage
the unstageable dirty content, so it can never be unstaged.
Treat a submodule's unstaged change as stageable only when its commit
isn't already staged, so that a staged submodule unstages on the next
press regardless of leftover dirty content. Because the decision is now
shared by press and stage-all, this fixes both at once.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
pressWithLock (acting on the selection) and toggleStagedAllWithLock (acting
on the whole tree) each independently decided whether to stage or unstage,
ran the optimistic update, and logged the action. That duplicated decision
has already drifted: the tracked-files filter was added to press months
before it was applied to stage-all, and fixes to one have repeatedly had to
be chased into the other.
Extract that shared decision into toggleStaged, leaving each caller to
supply only the git commands it runs (per-path for the selection, bulk
add -A / reset for the whole tree — the latter is required because the tree
root node has an empty path, so a per-path stage wouldn't work). This is a
pure refactor: the two callers' decisions were already equivalent, so
behavior is unchanged. It exists so the next change to the staging logic
only has to be made once.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
When a user switches into a repo whose .envrc hasn't been approved with
`direnv allow`, the previous behavior was to drop a "blocked" error
popup and leave the user to fix it externally. That meant opening a
terminal, running `direnv allow`, and then either restarting lazygit or
switching repos and back to refresh the env — easy to get wrong, easy
to forget.
When `direnv export json` exits non-zero, follow up with `direnv status
--json` to ask direnv whether the current directory has a not-yet-
allowed .envrc, and if so, get its path. Then show a confirmation popup
with the .envrc contents inline so the user can read what they're
approving. Confirming runs `direnv allow <path>` and re-runs the load
so the new env reaches subprocesses immediately; cancelling leaves the
env unloaded (the same state as before this commit when direnv refused
to load the .envrc).
Using `direnv status --json` instead of parsing the "is blocked"
stderr line means we rely on direnv's structured output rather than
its human-readable error format, which is more stable across versions
and avoids assumptions about output formatting.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
When a user opens a repo from the recent-repos menu or jumps between
worktrees inside lazygit, only the env vars present at process startup
reach subprocesses. That breaks pre-commit hooks and other tools whose
dependencies are pulled in by a per-repo .envrc — users were left with
read-only operations because the env their shell would normally load via
direnv never made it into lazygit's git invocations.
Shell out to `direnv export json` after each chdir and apply the JSON
delta via os.Setenv/Unsetenv. direnv tracks the previous load in its own
DIRENV_DIFF env var, so the delta also unloads vars from the old repo
when entering one without a matching .envrc. If direnv isn't on PATH the
call is a no-op, so users who don't use direnv pay nothing and users who
do need no config to opt in. Any stderr direnv emits (loading messages,
"blocked .envrc" errors, etc.) goes to the command log.
The integration test puts a fake direnv on PATH and asserts that a value
it exports reaches a custom command after switching repos. Wiring this
up needed runner.go to support `{{actualPath}}` placeholders in
ExtraEnvVars, mirroring the existing support for ExtraCmdArgs, so the
test can prepend a fixture-relative directory to PATH.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
If we return the error here, we don't switch repos, but the chdir
happened already, so this would be an inconsistent state (a lot of
lazygit's code assumes that the current directory is always the worktree
root). Only log the error; failing to record the current directory is
not the end of the world.
Also, it is very unlikely to happen; RecordCurrentDirectory only writes
to a small file, and if this fails, then either there is filesystem
corruption of the disk is full, and in both cases the user likely has
much bigger problems.
Commit 4f0393f97b caused a regression: for operations that use
WithWaitingStatusSync (examples are squashing fixups, moving commits up or down,
cherry-picking, creating fixup commits, and more), the waiting status wouldn't
show during the operation; however, it would show after the operation was done,
and then linger forever.
The cause: since 4f0393f97b, layout sizes the bottom line from the actual
content of the AppStatus view rather than from the status manager. The async
render path keeps the view in sync (it sets the buffer on the first tick and
clears it to "" when the status ends), but the sync path used by
WithWaitingStatusSync did not:
- It called ForceLayoutAndRedraw before writing anything to the view, so layout
saw an empty buffer and left no room; the status never appeared during the
operation.
- When the operation finished it just broke out of the loop, leaving the last
spinner frame in the buffer. Every subsequent layout kept reserving room for
that stale content, so the status stuck around forever.
Fix this by writing the status into the view before the initial layout, and
clearing it again when stopping.
This doesn't make a difference for the behavior, it just looks strange to
include the empty bindings first and then filter them out in the next statement.
If a keybinding that we want to display in the options bar was set to <disabled>
by the user, in pre-0.62 versions we would still display the command, but with
no keybinding. This was arguably not very useful before, but now it actually
crashes because we would now try to display the first key of the slice of
configured keys (crash introduced in 3d18ee8f91). Fix the crash by not showing
those commands at all.
The previous logic only re-rendered the main view when the side panel itself was
focused. When the user pressed `0` to focus the main view, the
Normal/NormalSecondary context becomes Current and the equality check failed, so
cycling pagers had no visible effect. Mirror the pattern from postRefreshUpdate:
when the main view is focused, call HandleRenderToMain on the side panel below
it on the stack (which CurrentSide already returns).
Convert the remaining *Alt/*Alt[12] sibling fields (PrevItem/NextItem,
GotoTop/GotoBottom, PrevBlock/NextBlock, ScrollUpMain/ScrollDownMain,
OptionMenu, ConfirmInEditor, DiffingMenu) so the merge mechanism folds
their values into the corresponding main multi-key binding at config
load. The redundant alt-only Binding registrations across the various
controllers and the global keybindings file are gone: the merged main
field already carries every key, so the for-loop in SetKeybinding
registers them all.
Previously the patch_explorer and merge_conflicts controllers reused
Universal.PrevBlock/NextBlock for moving between hunks (or conflicts) in the
main view, sharing keys with the global side-window cycle. The two operations
are conceptually distinct: cycling side windows is a global navigation gesture,
while next/prev hunk acts on the diff in the main view. Tying them together also
blocks adding <tab>/<backtab> as side- window-cycle keys, because <tab> already
means "toggle panel" in the staging view.
Add Main.PrevHunk/NextHunk to the existing KeybindingMainConfig (which already
groups bindings for the main view across staging, patch building, and merge
conflicts) and switch both controllers to it. The defaults match the active key
set those controllers had before (<left>/<right>/h/l), so the user-visible
behavior is unchanged.
Now that quit accepts multiple keys, the historical quit-alt1 field is
redundant: existing configs that set it should keep working without the user
having to migrate, but the lazygit code shouldn't have to register the alt
binding separately.
Add a merge step that runs after the user config is loaded (and from
NewDummyAppConfig, which the cheatsheet generator and integration tests go
through) folding the alt value into the main key list. Mark QuitAlt1 deprecated
so it disappears from the generated Config.md example, while staying in the JSON
schema with a description so editors can still steer users toward the new form.
Note that instead of marking the alt config as deprecated, we could have added a
migrator that changes users' config files and gets rid of the alt config for
good. I decided not to do that, because this would render the config file
invalid for older versions of lazygit, which would then refuse to start; and
that's annoying when bisecting bugs. We'll keep the deprecated configs in the
code for a year or so, and then add the migrator.
The next commit will fold the remaining ~15 -alt-style fields the same way; the
helper is shaped to keep that mechanical.
CustomCommand.Key and CustomCommandMenuOption.Key are user-configured
keybindings just like the built-in ones. Converting them to the Keybinding type
lets a user assign multiple keys to the same custom command, e.g. `key: [a, b]`,
the same way they would for any other keybinding.
The validator iterates over the elements rather than checking a single string,
the binding registration goes through GetValidatedKeyBindingKeys to register
every alternate, and the existing error messages use .String() so a multi-key
binding renders sensibly.
CustomCommandPrompt.Key (a form field name, not a keybinding) stays a plain
string.
JumpToBlock is special: each of its 5 elements is the binding for one side
window (status / files / branches / commits / stash), not an alternate for a
single command. Change the field from []string to []Keybinding so each window
slot can have alternates of its own.
The schema becomes "an array of 5 keybindings, each itself a string or array of
strings", which falls out cleanly from how the Keybinding type inlines into the
generated schema. Existing configs (a flat array of 5 strings) keep validating
because each element is unmarshalled through Keybinding's scalar-or-sequence
decoder.
Until now every keybinding config field was a plain string. That meant a user
couldn't ask for two keys to invoke a command — the config silently accepted
only one form.
Convert every string-typed field across all 13 KeybindingXxxConfig structs to
Keybinding so the union type extends to every command. Defaults wrap their
single-key value in Keybinding{...} so the generated Config.md still renders one
scalar key per binding.
The alt fields keep their separate Binding registrations for now: this commit
does not yet introduce the merge mechanism that folds them into the main field —
that comes in a follow-up. Consumers previously calling opts.GetKeys on a string
field now call opts.GetKeys on the Keybinding, or take .String() / Keys[0] where
a single value is needed.
Adds a Keybinding.String helper for rendering, schema-generator work that
inlines the Keybinding union into each consuming property, and a unit test
covering the user-facing scalar/sequence YAML forms for quit.
This is a pure refactor in preparation for letting users configure multiple
alternate bindings for a single command. Every Binding still has exactly one
key, so nothing changes visibly: the cheatsheet, the on-screen options bar,
and the keybindings menu all render identically.
When a Binding ends up with multiple keys, the on-screen options bar will
show only the first (to avoid clutter); the cheatsheet will show all of them (in
a later commit). For now both paths take Key[0].
MenuItem.Key is changed in the same way, it also has a slice of keys now.
In this commit we keep the name `Key` in Binding, KeybindingOpts and MenuItem,
instead of renaming them to `Keys` right away, in order to keep the diff a bit
more readable. We'll do the rename separately in the next commit.
Constructing a menu item key from a literal character requires
gocui.NewKeyRune('r'), which is a bit noisy. Add a private menuKey helper in
both the controllers and helpers packages so the common case in either reads as
menuKey('r'). Duplicating the one-liner is cheaper than a cross-package import
dependency and avoids forcing every controller file to qualify the call.
The reason for doing this now is that we are going to change MenuItem.Key to a
slice of keys later in the branch, which means we'd have to add `[]gocui.Key{`
at each call site, making them even more noisy. With the menuKey helper we can
just change its signature and leave all clients unchanged.
For legacy reasons, OptionMenu was set to `<disabled>`, and OptionMenuAlt1 to
`?`. This doesn't make a lot of sense any more; get rid of OptionMenuAlt1 and
bind OptionMenu to `?` by default. This is a breaking change for users who
rebound OptionMenuAlt1 in their config, but it doesn't strike me as very likely,
and it's easy enough to fix.
AutoForwardBranches relies on the worktree model to skip any branch
that's currently checked out in another worktree (so we don't update
its ref behind the worktree's back). The post-fetch refresh wasn't
including the worktrees scope, so any external change to the worktree
list between lazygit's startup and the fetch — a `git worktree add`,
a `git checkout` in a linked worktree, a branch rename — left the
in-memory model stale and the skip check returned false negatives.
Add WORKTREES to the post-fetch refresh scope when auto-forwarding is
enabled. We gate on the config so users with auto-forward disabled
don't pay for an extra `git worktree list` plus per-worktree rev-parse
on every fetch tick.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The post-fetch logic was duplicated in `backgroundFetch` and the manual
fetch handler: refresh a fixed set of views, then auto-forward branches
if the fetch succeeded. The two had already drifted on the refresh
scope; folding them into a single helper makes the duplication go
away and prevents it from drifting again.
Pass the fetch error through so we preserve the previous behaviour of
refreshing unconditionally but only auto-forwarding on success.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The background fetch path already includes PULL_REQUESTS in its
post-fetch refresh scope, but the manual fetch from the files view
doesn't. As far as I can tell that's an oversight from when
PULL_REQUESTS was added — there's no reason the two paths should
differ. Align them so both refresh PRs after fetching.
This also sets up the next commit to extract a shared helper for
the post-fetch refresh.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
If `git worktree list` fails, we want the Worktrees model to fall back
to an empty slice so callers iterating over it stay correct. The error
branch was setting it to `[]`, but the line below unconditionally
overwrote it with the nil `worktrees` value from the failed call.
Use an else branch so the empty-slice fallback actually sticks.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The branches-panel PR icons only worked for github.com remotes. There was no
fundamental reason — the auth library we already vendor (cli/go-gh) supports
enterprise tokens out of the box (GH_ENTERPRISE_TOKEN, gh auth's keyring),
and the user-facing 'services' config has long been the documented way to
tell lazygit "this domain is a github service" for the View-PR-URL feature.
The fetcher just hardcoded github.com in three places:
- a substring check on the remote URL to decide we're "in a github repo",
- the GraphQL endpoint (always api.github.com/graphql), and
- the auth lookup (always against the default host).
Plumb the resolved web domain through instead. Detection now goes through
the hosting_service ("is this remote's provider 'github'?"), which means a
user with services: { 'git.acme.com': 'github:git.acme.com' } configured
gets PR icons on their GHE remotes too.
Replacing the substring check with a provider check also tightens a latent
bug in getGithubRemotes: it previously accepted any remote whose URL parsed
with the default regex, including gitlab and bitbucket — masked today only
by the InGithubRepo gate, but exposed once the gate goes away.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
After a successful paste DidPaste is true, hiding the "X commits copied"
indicator but leaving the buffer populated. From the user's perspective
this looks like a clean slate, so a new shift+C should start fresh. It is
important to reset DidPaste first, before populating the buffer with the new
commits, because otherwise each loop iteration would overwrite the previous one
since Add() rebuilds the set via SelectedHashSet() which returns empty while
DidPaste is set.
Now that HandleRender no longer does an implicit Render(), we can use it inside
OnUIThreadContentOnly to save performance, on the assumption that rerendering a
view that contains an inline spinner never changes the layout.
self.c.Render() at the end of HandleRender was there to schedule a gocui Update
tick so the view content modified above would actually get drawn. For UI-thread
callers (the great majority -- keybinding handlers, the layout function itself,
popup resize, etc.) this was unnecessary work, since gocui already runs a
layout/redraw cycle after every event. SimpleContext.HandleRender doesn't call
Render() either, so this aligns the two implementations.
The few callers that drove HandleRender from a worker goroutine and relied on
Render() for the flush were wrapped in OnUIThread in the preceding commits, so
the implicit Render is no longer needed.
Also, Render() being called *before* setFooter() looks like it might have been a
theoretical race; this is no longer an issue now.
The redraw of the selection color (using <space>) would be tied to the
spinner drawing. To reproduce, having a high spinner refresh rate and
toggling a file would see a delay equivalent to the time spinner refresh
rate.
SetSuggestions has two callers: prepareConfirmationPanel calls it directly on
the UI thread, while editors.promptEditor and
SuggestionsContext.RefreshSuggestions call it via AsyncHandler, which runs the
result closure on a worker goroutine. The worker path currently relies on
HandleRender's self.c.Render() to flush the view update. Wrap the body in
OnUIThread so the worker path stays correct when Render() is removed; for the
UI-thread caller the extra bounce is harmless.
refreshBranches runs on a worker goroutine and re-renders the commits view
directly to refresh the branch-head visualization. As with refreshView, this
currently only flushes because HandleRender ends with self.c.Render(). Wrap the
explicit HandleRender call (along with the LocalCommitsMutex pair around it) in
OnUIThread so it keeps working once Render() is removed.
An async refresh dispatches refreshXyz on a worker goroutine, which then calls
refreshView -> PostRefreshUpdate -> HandleRender. Today the final
self.c.Render() inside ListContextTrait.HandleRender is what triggers a UI flush
from the worker. We're going to remove that Render() call, so prepare by
wrapping refreshView's body in OnUIThread.
This moves the entire rendering of the view (and the ReApplyFilter/ReApplySearch
stuff) to the UI thread, not just the layout. I don't expect this to make a
difference in practice, and it is already one step towards my long-term goal of
moving all view rendering to the UI thread (see
https://github.com/jesseduffield/lazygit/issues/2974#issuecomment-1729154768).
Unrelated to this branch, just because we're touching this code: there's little
reason to set the color on the background thread but the text on the UI thread.
Set them both together on the UI thread. Avoids a data race (unlikely to be a
problem in practice, we're talking about a 64-bit int, but still).
Stefan HallerandClaude Opus 4.8
Antoine Gaudreau Simard