bertfrees
commented
5 years ago - Make more use of utility steps (px:normalize-uri, px:assert, px:set-base-uri, px:fileset-load, px:fileset-update, px:validation-status).
- Move as much code as possible to utility steps (px:epub-load, px:opf-manifest-to-fileset, px:daisy-load, px:html-to-fileset; see also https://github.com/daisy/pipeline-scripts-utils/pull/98 and https://github.com/daisy/pipeline-modules-common/pull/114).
- Put as much code as possible in main steps and keep script wrappers as minimal as possible.
- Use the name pattern "px:x-to-y" for the main steps and "px:x-to-y.script" for the script wrappers.
- Add fileset and in-memory ports when possible.
- Main steps are responsible for loading filesets into memory.
- More XProcSpec tests
- ...