I've been practicing git checkout -- somefile.txt to checkout git files, in case I mess up. How does Git know which file gets checked out from which commit?
i.e.: I write git checkout -- index.html. How does Git know which commit to take index.html from, especially as your commit list gets larger and larger?
Using git checkout with a path but without specifying a commit will check out the file from the index, i.e. the staging area. So for example if you have staged some changes to a file, and then modified it further and only want to undo those additional changes, then you can use git checkout -- some/path to undo those changes.
If your index is clean, i.e. nothing is reported as being staged when using git status, then the index will be equivalent to HEAD, the currently checked out branch (or commit). So this is a way to undo changes too.
Of course specifying a commit with git checkout will always use the file version from that exact revision, regardless of your index.
git checkout -- some/path checks out the version from the index, not from the HEAD commit. In a lot of cases these are the same underlying blob anyway so that it makes no difference, though.
git checkout -- index.html, with the space, orgit checkout --index.html, without the space?