Automatic transformation of XML namespaces/Pipeline

From Wikiversity
Jump to navigation Jump to search

The develop branch of XML Boiler does contain running several automatic transformations joined in a pipeline.

It should also support running a particular transformation or a particular script as a step of the pipeline.

Both for transformations and scripts there are three variants. I will describe for scripts:

  1. Pass script RESOURCE SCRIPT with two URIs.
  2. Use all assets loaded till the point of processing to get script URIs.
  3. Use only these assets which were loaded with --preload option.

It is unclear how to add the resource to the algorithm of downloading assets in variant 1. In variant 3 we need also download assets which are specified (by --preload) before the asset in question.

In variant 2 we have underdetermined behavior that validity of the pipeline would depend on details of execution of possible previous chains in the pipeline.

In variant 3 there is the problem that non-preload assets may be downloaded before preload assets because (see here) the traversal is repeated (unless stuck in an infinite loop) for every outer list element.

TODO: Other issues?