Background and constraints
P4StreamSwitcher targets Perforce development across multiple projects and branches. Workspace counts need to be controlled while branch directories may contain local UE caches. Repeatedly changing client configuration by hand makes workspace, stream, and directory associations harder to track.
This implementation in the source applies to Windows and the two-level stream layout //depot/project/branch.
Breaking down the problem
Switching to a target branch requires locating a workspace, deciding whether to create a client, selecting a directory, and choosing whether to update it from the server or preserve local content. Treating these actions as an unconditional sync misses the difference between existing workspaces and offline directories.
A full pool also requires reclaiming clients. Deleting by name alone does not account for usage order or opened files.
Implementation
The tool identifies workspaces as “tag_project_branch” and stores the directory, offline option, and pending synchronization flag per stream. An existing workspace opens directly in P4V when no sync is needed. A manually changed Root or an unfinished previous sync triggers the synchronization process again.
The pool chooses an eviction candidate using local timestamps, then checks its opened files and whether P4V is running before deletion. Tag conflict checks help prevent hosts from sharing a tag; ordinary manually created workspaces outside the naming convention are excluded from the pool.
Directories can be handled in two modes. Normal mode first updates the have list, then applies built-in cache protection and .p4ignore rules, removes extra files, and finally downloads files whose content differs or that are missing. Offline mode updates the have list before running reconcile, opening local additions, edits, and deletions in the default changelist for user review.
Results and boundaries
The implementation provides workspace selection, count management, and two directory handling modes, with logs and sync status in the interface. It does not provide automatic backups, and an empty opened-file list does not prove that no untracked local modifications exist on disk.
Normal cleanup can delete extra local files that are not ignored; offline mode changes pending records. Directory selection and mode confirmation are therefore required before use. Source capabilities described here are not production test results, and no guarantee of zero cache loss or fixed performance gains is made.