Worktree operations

Every operation is reached from both ends: ddev branchery <command> in a shell, or the interface, which starts the same command and reads its log. Both leave the same record, so an operation started in a terminal shows on the page while it runs and stands in the worktree's history afterwards. One operation runs per worktree at a time, and operations on different worktrees run side by side.

Commands#

Run ddev branchery for the commands the installed image has, and append --help to any of them for its arguments and options.

Command	Purpose
`worktree:list`	list worktrees and their current state
`worktree:fork <branch>`	create a branch from the project or another worktree
`worktree:add <branch>`	check out an existing local or remote branch
`worktree:pull <name>`	fast-forward a worktree branch from its upstream
`worktree:restore <name>`	return to the branch the worktree was made for
`worktree:discard <name>`	reset that branch to its upstream
`worktree:provision <name>`	rebuild dependencies and application state
`worktree:config [<name>]`	rewrite generated configuration for one or all worktrees
`worktree:php <name> [<version>]`	show or change the PHP runtime
`worktree:remove <name>`	remove checkout, branch, database and state
`database:sync <name>`	replace a worktree database from another source
`database:prune`	list or drop databases without a worktree
`git:fetch [<remote>]`	fetch and prune remote branches
`jobs:list`	list operations that are running
`jobs:show <id>`	inspect or wait for an operation
`config:example`	print or write a starting project configuration
`docs [<page>]`	list or print pages of this manual

What follows is what each operation does, in order. A step marked config is not Branchery's own work but what the project's .ddev/branchery.yaml says, directly or through the shipped profile it names. Without that file a worktree is a checkout at an address of its own, with an empty database. Configuration reference is the file, key by key.

The build of a worktree, and which half of it a project owns. add, fork and provision differ in where the checkout and the data come from; the phases below are the same.

Start with Troubleshooting when an operation stopped.