You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This could get most of the benefits of #4872, since if you have a file you won't need later in your workflow you can split out a sub-workflow to scope it and the file will go away when the sub-workflow ends. And it doesn't need any reference-counting or hooking into expression evaluation or weird un-designed synchronization primitives between jobs, just the ability to track all files created (and gather them at the end of the workflow), and the ability to dig through the output values, and the ability to delete what's in one set and not the other.
If we keep the current environment-only-grows system, the environment could be the all-files-ever-created collection, and we wouldn't need a separate one. I liked the simplification I was working on in https://github.com/adamnovak/toil/tree/issues/4872-drop-wdl-files, where a job will produce an environment with just what it actually defines itself, but maybe that's not needed.
┆Issue is synchronized with this Jira Story
┆Issue Number: TOIL-1702
The text was updated successfully, but these errors were encountered: