Write lock URI needs to have prepended.

Explicitly use addBinaryBody.

Attempt to avoid issues fetching external mapping (columns.csv) for upload.

Need to return to break out from retry loop.

Allow adding data from a remote HTTP source.

Typo, csvv -> csv.

Implement components pipeline call. Remove un-used methods.

Add test for write-lock (503) and implement retry/holdoff polling of write-lock path.

Implement codelist pipeline using HTTP Components direclty.

Test for jobDraft.replace() issue. Extend Exception rather than Throwable.

Test for and add cache buster config as part of publish.

Implement Drafter.publishDraftset()

Apache components HttpHost uses -1 as default port for scheme; see if that helps with pre-emptive basic auth.

Use husdon.FilePath() to get file InputStream. May need special handling for remote nodes, see https://stackoverflow.com/questions/41250567/using-filepath-to-access-workspace-on-slave-in-jenkins-pipeline

Use Apache http components directly in cube import pipeline, with attempt to use FileInputStream for streaming potentially large HTTP requests.

Try implementing with callable global to set things up and compiled class implementing drafter methods. Use test JSON files as responses.