[
https://issues.apache.org/jira/browse/CAMEL-26936?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Work on CAMEL-26936 started by Claus Ibsen.
-------------------------------------------
> camel-jbang - camel run --background and camel_run report readiness up to 1.5
> s late
> ------------------------------------------------------------------------------------
>
> Key: CAMEL-26936
> URL: https://issues.apache.org/jira/browse/CAMEL-26936
> Project: Camel
> Issue Type: Sub-task
> Components: camel-jbang
> Reporter: Claus Ibsen
> Assignee: Claus Ibsen
> Priority: Major
> Fix For: 4.24.0
>
>
> Measured on 2026-11-13 with the unpacked 4.24.0-SNAPSHOT camel-launcher on
> macOS (Apple M4 Pro, JDK 25), warm CDS archives, medians of 5-9 runs on a
> machine with other load (treat as +/-30 %). Estimates are marked as such.
> Paths are relative to {{dsl/camel-jbang/}}.
> {{run --background}} returns at 2.75-2.88 s, while the routes are started at
> 1.05-1.40 s (10 runs, connector lag 0.83-0.89 s every time):
> * {{FileCliConnectorTransport}} writes the status file every 1 s ({{delay =
> 1000}}, line 75), and its first {{statusTask()}} runs before the context has
> started.
> * {{Run.runBackgroundProcess}} polls the status file every 500 ms (around
> line 1358).
> * The MCP and TUI {{IntegrationLauncher}} polls the same file every 250 ms
> ({{started()}} line 359, poll line 240), so {{camel_run}} has the same lag.
> Proposal:
> * Write the status file when the context has started (event driven), and
> after each action, so a {{camel get}} right after a {{camel cmd}} is not
> stale (today it shows the old totals in 3 of 4 tries).
> * Poll at 50-100 ms in {{runBackgroundProcess}} and {{IntegrationLauncher}}.
> Expected: -1.2 to -1.5 s per run (estimate from the measured lag plus the
> average poll wait).
> Related, measured separately: the parent configures the whole run before it
> forks ({{ProjectRunner.configure}} at Run.java:815, 5,005 classes,
> {{--background-wait=false}} still 0.66-0.74 s), and on macOS a JVM that exits
> while its child lives costs +0.3 s (a double fork takes 0.04 s). Forking
> first and a double fork would save a further 0.3-0.6 s (estimate).
> _Claude Code on behalf of Claus Ibsen_
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)