"date" is a bit of a loose concept in git. A commit will have an author date that may be some time well in the past before someone actually pulls/commits the commit into their repository, also the commit may be rebased and updated to be on top of an apparently newer commit.
A commit also has an commit date which is updated if a commit is rebased or amended in any way. These commits are more likely to be in some sort of chronological order but you are still at the mercy of the committer having the correct time set on his computer and even so, an unmodified commit can sit on a feature branch on a remote repository indefinitely before being merged into the master branch of a central repository.
What is probably most useful for your purposes is the reflog date on the particular repository in question. If you have per-branch reflogs enabled (see git config core.logAllRefUpdates
) then you can use the ref@{date}
syntax to refer to where a branch was at a particular time.
E.g.
git log -p master@{2009-07-01}..master@{now}
You can also use 'fuzzy' descriptions like:
git log -p "master@{1 month ago}..master@{yesterday}"
These commands will show all commits that have 'appeared' in the given branch of the repository regardless of how 'old' they actually are according to their author and commit dates.
Note that the per-branch reflog is specific to a repository, so if you're running the log command on a clone, and you don't pull for (say) a month then pull all the changes for the last month at once, then all of the last month's changes will appear in a @{1 hour ago}..@{now}
range. If you are able to run the log command on the 'central' repostory that people push to, then it may do what you want.