Skip to content

Latest commit

 

History

History
27 lines (18 loc) · 1.57 KB

DEVELOPER_NOTES.rst

File metadata and controls

27 lines (18 loc) · 1.57 KB
Raw

Items to Test/Check:

  • Using the .cookiecutterrc file regenerates the same output. (this doesn't work with fancy input selection)
  • Make sure output builds in cFS Bundle
  • Make sure output builds without the sample library, if that option is not selected

Future Work:

  • Template the file header generation so its not just the nasa Sample app output
  • Use Jinja loops to template out command, HK, and perf setup for dynamic quantities
  • Write python wizard that leverages the cookiecutter questions interface for more flexible template generation
  • -> Auto-generate the cfe_es.scr line, it needs variables for the app main function, and the .so name
  • -> Fix the table name generation so cFE can properly load the table with default output (maybe variableize it?)
  • -> Make separate underscore variable to allow shortening the display name for compiled text strings
  • -> Add app to HAVE list if using the #defines in sch, to, or ci

Open Questions:

  • Bootstrap the bootstrapper? Basically pick a set of configs that only need to be generated as one variation from the base env, make that an input step to build the setup.cfg that matrix uses for all the combination envs
  • Re-write the render.py script to call cookiecutter via the command line, because thats how the user does it, and its not always renderec quite the same way with the same effects. Found the click dependency this way, and would be a better way to get around the rendered field bug instead of using the workaround of re-writing the cookiecutter.json file.