Skip to content

omarchy plugin add --enable aborts with "omarchy-shell is not responding" when many plugins are installed #14429

Open

Description

@Nejcc

What happens

$ omarchy plugin add https://github.com/<owner>/<plugin>.git --enable
...
Added <id> into ~/.config/omarchy/plugins/<id>
omarchy-shell is not responding

The plugin is cloned but never enabled. Running omarchy plugin enable <id> a few seconds later works.

Why

In bin/omarchy-plugin-add, mv "$stage" "$target" lands the plugin under ~/.config/omarchy/plugins/, and the shell's watcher reloads all plugin code. The next line, omarchy-shell shell rescanPlugins, uses the default 2s IPC timeout in bin/omarchy-shell (OMARCHY_SHELL_IPC_TIMEOUT). With around 18 third-party plugins installed the reload takes longer than that, so the call times out, fail exits, and the enable step never runs. The same thing happens to any omarchy-shell call made right after a file change under the plugins directory.

Suggested fix

Before rescanPlugins (and before omarchy-plugin-enable), wait for the shell to answer: retry omarchy-shell shell ping for a bounded time (e.g. 30s), or call these with a longer OMARCHY_SHELL_IPC_TIMEOUT. A rescan that fails after the move could also print the omarchy plugin enable <id> hint instead of a bare error.

Seen on Omarchy 4.0.4, Hyprland 0.56.2.

Activity

  1. jandrusk commented on Oct 7, 2026

    @jandrusk

    Thanks for the clear write-up. This looks like the same failure as #9304 (plugin add --enable timing out while the shell rebuilds plugins). #11118 (an IPC budget sized for a reload) and #11117 (best-effort closing rescan) are open fixes for it, and #11115 is the same 2 s timeout hitting plugin update. Probably best to follow up on #9304 so it's all in one place.

  2. omarchybot commented on Oct 18, 2026

    @omarchybot
    Collaborator

    Checked against quattro: bin/omarchy-plugin-add still makes an unguarded omarchy-shell shell rescanPlugins call at line 150 under set -euo pipefail. bin/omarchy-shell still defaults that call to a 2s timeout, so a plugin reload that runs longer stops the script before it reaches omarchy-plugin-enable at line 173. That is the path you describe.

    This one has no label yet. A label goes on only when two models classify the item separately and agree. Claude Opus 5.5 classified it as a bug, but the second opinion (Codex Medium) did not run because today's review budget was spent. So the classification is unsettled and nothing has been verified on a worker.

    It reports the same symptom as #9304 (plugin installed but left disabled, and a later omarchy plugin enable works). #11117 makes this closing rescan best-effort, and #11118 gives the plugin commands a longer IPC budget. Both PRs are open against quattro. Whether this issue is a duplicate of #9304 will be decided once the classification is settled, and that is what it waits on now.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions