search

aaronpeikert / reproducible-research

A Reproducible Data Analysis Workflow with R Markdown, Git, Make, and Docker
https://psyarxiv.com/8xzqy/
Creative Commons Attribution 4.0 International
122 stars 17 forks source link

Hashing #4

Closed aaronpeikert closed 4 years ago

aaronpeikert commented 4 years ago

@gabrielesh pointed me to an error in our approach on hashing, we should think it over. What are your thoughts @brandmaier ? maybe we should hash the file instead of the R Object?

aaronpeikert commented 4 years ago

https://github.com/aaronpeikert/workflow-showcase/issues/2

brandmaier commented 4 years ago

Yes, very good point!

aaronpeikert commented 4 years ago

So there is md5sum-function that comes with R to hash files, which would be a good fit for the workflow-example, but in the manuscript, we have no file. On the other hand, I have realised that we just have failed to rebuild the docker images after fixing the date, so that an update of the digest package lead to another hash. So if we had done everything as we prescribed ourself, we would not have this problem.

aaronpeikert commented 4 years ago

After thinking about it a bit I decided to close this. Our general approach on package management should cover this. Anyways @brandmaier feel free to reopen if you want to continue the discussion.