In a Udacity lesson covering relative commit references, it says:
^ indicates the parent commit, ~ indicates the first parent commit
The m
X~n
means: The n
th ancestor of X
.
X^
means: The parent of X
. This is equivalent to X~1
.
If X
has more than one parent, one needs to distinguish between them when using the ^
notation. So X^1
would be the first parent, X^2
would be the second parent, and so on. X^
is equivalent to X^1
(and also equivalent to X~1
).
In your example, starting from commit 9ec05ca
, which is HEAD
:
db7e87a
is HEAD~1
(or alternatively HEAD^
).796ddb0
is HEAD~2
(or alternatively HEAD^^
).1a56a81
is HEAD~4
(or alternatively HEAD^^^^
, but nobody would use that).e014d91
, being the first parent of 1a56a81
, is HEAD~5
, or HEAD~4^
, or HEAD~4^1
.f69811c
, being the second parent of 1a56a81
, is HEAD~4^2
.https://git-scm.com/docs/gitrevisions
It's not totally wrong, but it's a bit too short: since the "~" operator is performing a depth search (nth ancestor, which actually corresponds to the nth last commit on a single branch), its argument can't specify a parent if they are many of them. It hence assumes it should always follow the first one of them. This correlates with your other question, where we were discussing about parent order.