3 ms·
Version control of large files is clearly a reasonable thing for a version control system to do. I think if you introspect you'll see you are defending a flaw
by IshKebab 6y ago
Version control of large files is clearly a reasonable thing for a version control system to do.
I think if you introspect you'll see you are defending a flaw in something you like with spurious technical objections because you don't want to admit it isn't perfect. That's understandable. Happens a lot.
- swiley 6y agoI didn't say versioning large files isn't a reasonable thing to do, I said it doesn't make sense in git. How is it supposed to work? The LFS way where you're storing a pointer to an external http resource? Just put a script in your repo to fetch that then. Or maybe stick the data in git's merkle tree and have a really slow repo? Why bother with LFS then?
- IshKebab 6y ago> How is it supposed to work? One obvious improvement would be for Git to use the hash of the object rather than the pointer file when calculating tree hashes. That makes the storage method for the actual files independent of the commit hashes. People have mentioned several other VCSs in this thread that do it already.