[
https://issues.apache.org/jira/browse/CAMEL-26413?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Claus Ibsen resolved CAMEL-26413.
---------------------------------
Resolution: Fixed
Fixed by https://github.com/apache/camel/pull/28723, merged to main for 4.24.0.
_Claude Code on behalf of davsclaus_
> camel-jbang - AI tools: an emptied body, an option of a to: that is not
> evaluated, a variable for a value of the route, and a line with no key
> ----------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: CAMEL-26413
> URL: https://issues.apache.org/jira/browse/CAMEL-26413
> Project: Camel
> Issue Type: Improvement
> Components: camel-jbang, camel-yaml-dsl
> Reporter: Claus Ibsen
> Assignee: Claus Ibsen
> Priority: Minor
> Fix For: 4.24.0
>
>
> Four more gaps from the r24 run of the local-model benchmark (qwen3.6:35b on
> Ollama, Oct 29-30 2026).
> * *An emptied body.* A setBody with $\{null\} before the split of
> $\{body[lines]\} left nothing to split, with no error (connect-http-client
> step 1, 2 runs). A setBody with constant "" before a GET made the call a
> POST, because the http producer sends a POST for any body that is not null
> (405 Method Not Allowed, 87 times). The validator now knows a body set to
> null or to empty text: a field read on it is reported, and so is empty text
> sent to an http call with no httpMethod, with the advice to use $\{null\} or
> httpMethod: GET. The steps of a split whose element is not known are now
> walked too, so what they set themselves is checked.
> * *An option of a to: that is not evaluated.* key: "$\{header.orderId\}" on a
> to: kafka gave every record that text as its key
> (connect-service-kafka-orders step 3, 2 runs). The options of a to: are fixed
> when the route starts. The validator now reports an expression in an option
> that has a header of its own for each message (CamelKafkaKey for the key of
> kafka), with the advice to set that header. Options that the catalog marks as
> taking an expression, or whose description says so, are left alone, as some
> components evaluate their options themselves.
> * *A variable for a value of the route.* Asked to set a variable, the model
> set an exchange property in 3 runs of 7 (run-order-generator step 1), and the
> header/property mix-up message then kept it there. That message now also
> offers setVariable and $\{variable.x\}.
> * *A line with no key.* A setHeader with the bare line directory (meant as
> name: directory) got the message that the key below it was in the wrong
> column, 15 times (ai-ticket-classifier step 2). The validator now says that
> the line has no colon, and what to write.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
