Hi @luccahuguet,
Would you be open to upstreaming—or supporting someone else upstreaming—the three commits on yazelix-tab-activity-pipe, so Yazelix could eventually use official dj95/zjstatus instead of maintaining a fork?
I tested replaying the commits onto upstream v0.24.0. The rebase succeeded with two small conflict resolutions while preserving upstream’s focused-CWD support, newer command-lock implementation, and public invalidate_command_result API.
The resulting custom delta is limited to:
Cargo.toml
Cargo.lock
src/pipe.rs
src/widgets/tabs.rs
The version remains 0.24.0, zellij-tile remains 0.44.3, and the tests and warning-failing Clippy checks pass.
The functionality also seems potentially useful outside Yazelix: the pipe name and markers are configurable, any producer could send the schema-v1 activity messages, and the multiline/embedded-:: parsing fixes are generally applicable. Yazelix could continue owning its orchestrator and defaults while upstream owns the generic consumer.
I couldn’t find an exact upstream issue or PR. PR #228 explored a related per-pane status protocol but was closed as targeting the wrong repository, while PR #259 added native bell indicators without covering external busy/alert activity.
Is there a design or historical reason these changes need to remain Yazelix-specific? If not, would proposing them upstream—perhaps as a pipe-parsing PR followed by the tab-activity feature—now make sense?
Hi @luccahuguet,
Would you be open to upstreaming—or supporting someone else upstreaming—the three commits on
yazelix-tab-activity-pipe, so Yazelix could eventually use officialdj95/zjstatusinstead of maintaining a fork?I tested replaying the commits onto upstream
v0.24.0. The rebase succeeded with two small conflict resolutions while preserving upstream’s focused-CWD support, newer command-lock implementation, and publicinvalidate_command_resultAPI.The resulting custom delta is limited to:
Cargo.tomlCargo.locksrc/pipe.rssrc/widgets/tabs.rsThe version remains
0.24.0,zellij-tileremains0.44.3, and the tests and warning-failing Clippy checks pass.The functionality also seems potentially useful outside Yazelix: the pipe name and markers are configurable, any producer could send the schema-v1 activity messages, and the multiline/embedded-
::parsing fixes are generally applicable. Yazelix could continue owning its orchestrator and defaults while upstream owns the generic consumer.I couldn’t find an exact upstream issue or PR. PR #228 explored a related per-pane status protocol but was closed as targeting the wrong repository, while PR #259 added native bell indicators without covering external busy/alert activity.
Is there a design or historical reason these changes need to remain Yazelix-specific? If not, would proposing them upstream—perhaps as a pipe-parsing PR followed by the tab-activity feature—now make sense?