It said there was one conflict, but I looked in the file and there were no conflict markers.

Without more details about what Git said (the message about conflict), I can only guess what happened. And I guess that Git said that there was textual (context) conflict that it was able to automatically resolve, or the conflict was of the tree-level variety like for example CONFLICT (delete/modify) where one side deleted file and the other side modified it.

I added the file, committed the change, and did the merge command again. "Already up-to-date" is the message I get.

If one branch in ancestor of the other, you would get fast-forward or up-to-date situation. Git refuses to do pointless merges unless explicitely requested.

If you want to redo a merge, you must first undo it:

git reset --hard HEAD^ git merge B 

Just in case you decided that the resolution you made was good: there is reflog (git reflog) just in case you change your mind.

It said there was one conflict, but I looked in the file and there were no conflict markers.

Without more details about what Git said (the message about conflict), I can only guess what happened. And I guess that Git said that there was textual (context) conflict that it was able to automatically resolve, or the conflict was of the tree-level variety like for example CONFLICT (delete/modify) where one side deleted file and the other side modified it.

I added the file, committed the change, and did the merge command again. "Already up-to-date" is the message I get.

If one branch in ancestor of the other, you would get fast-forward or up-to-date situation. Git refuses to do pointless merges unless explicitely requested.

If you want to redo a comitted merge, you must first undo it:

git reset --hard HEAD^ git merge B 

Just in case you decided that the resolution you made was good: there is reflog (git reflog) just in case you change your mind.