It will be “safe” because you will not break the repository or your working copies or any future functionality.
But I like the idea of holding them simply because you could theoretically go and find out who actually performed a certain part of the work and what its real history was (theoretically, if you cared enough). If you delete merge information, you no longer have any clues where the linked code came from.
However, this is a weak argument for storing this information if it gets out of hand and causes problems, or even if it just annoys you.
Perhaps you could leave it for a couple of revisions, and then get rid of the "old" ones. Thus, the fault will include a merger revision, which will also allow you to find related information about the merger, as it was then.
Jim t source share