WMK666666 commented on PR #29720: URL: https://github.com/apache/camel/pull/29720#issuecomment-6826688721
All three points are in `99c687f`: **1. Regeneration (the CI blocker).** The description now carries through all four copies, `+23/−10` over five files: | file | change | |---|---| | `DoclingConfiguration.java` | the annotation | | `components/.../docling.json` | the two entries that describe the option (`componentProperties` and `properties`) | | `catalog/camel-catalog/.../components/docling.json` | same file, it is a straight copy of the component's own JSON | | `DoclingComponentBuilderFactory.java` | one javadoc block | | `DoclingEndpointBuilderFactory.java` | the three javadoc blocks (consumer, producer, advanced) | No `deprecationNote` noise this time. On how the four generated copies were produced: I regenerated the component metadata and then applied the generator's own javadoc wrapping rule to the catalog copy and the DSL factories — `78 - indent` as the line width, breaking at the last space, `'.'` appended only when the description lacks one (`component-builder.vm:61`, `ComponentDslMojo.javadoc`, `EndpointDslMojo.createBaseDescription`). Before trusting it I checked that this rule reproduces every existing option block in both files byte for byte: 60 of 61 in `DoclingComponentBuilderFactory`, 358 of 361 in `DoclingEndpointBuilderFactory`. The only misses are the file header paragraphs, which carry deliberate newlines. So the four DSL paragraphs are what a full build produces, not hand-typed. **2. Wording.** Used yours verbatim: > Output format of the SUBMIT_ASYNC_CONVERSION operation; the other operations always use their own format. The html_split_page value is rejected, as docling-serve does not return the split pages; use html instead. I checked the premise first: `DoclingProducer.java:379-381` is the only place the option is read, and line `512` uses the format recorded on the task entry, so the description now matches the code. **3. `markdown` in `enums`.** Added — `enums = "markdown,md,json,html,html_split_page,text,doctags"`. One thing worth knowing: this is annotation-only, it changes no generated file, because the metadata generator does not emit `enums` for `String` options (`ocrLanguage` has a long `enums` list and no `enum` in the JSON either). So if the catalog validation you mentioned reads the JSON rather than the annotation, it still cannot see the list — tell me if you want that fixed properly in a separate issue. `formatter:validate`, `impsort:check` and a compile of the module pass. CI still needs the approval for a first-time contributor. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
