Lots of derived files and images are in history

Question

Lots of derived files and images are in history

dme26 opened this issue 6 years ago · 4 comments

Running git clone gives a folder around 234M in size although the files present total 10M. This is due to images from checkPlots/ and various derived HTML files in the history. It should be straightforward to rewrite some of that git history, but that doesn't mean it's worth doing (I'm happy to look into it, but if so we should rewrite before publication since the commit IDs will change).

dataknut commented 6 years ago

fixed

Answer 1 · 2018-08-16T05:07:27.000Z

I know. I need to delete checkPlots/ from the history somehow

Answer 2 · 2018-08-16T05:18:11.000Z

https://rtyley.github.io/bfg-repo-cleaner/

hmm

Answer 3 · 2018-08-16T11:50:47.000Z

Looks good! ( and I like Scala, too… )The advantage here is that you have very specific directory paths to lop off the git trees (e.g., checkPlots/ ) which I think the cleaning tool would directly assist with. I was thinking of just using the git built-in commands, but I think your plan to use the cleaner would be better. I’d think it best to fork the repository before rewriting history, just so that the existing 250MB repository is still available should we need to go back to any part of it.