In following example tree:
A-B-C-D-E (master branch)
\
F-G-H (xxx branch)
I'm looking for F - the first commit in xxx branch. I think that it is possible with:
git log xxx --not master
and the last listed commit should be F. Is it correct solution or maybe there are some disadvantages of it?
I know that there were similar questions on stackoverflow, but nobody proposed such solution, and I'm not sure if I do it right.
git log master..branch --oneline | tail -1
Where "branch" is your specific branch name. The dot-dot gives you all of the commits that the branch has that master doesn't have. tail -1 returns the last line from the previous output.
master, and as long as your branch hasn't been deleted, git log <sha>..branch --oneline | tail -1 should still give you the results you're looking for.
--reverse to read the logs backwards with colors, etc.git cherry master -v | head -n 1
cherry - Shows the commits on your current branch which aren't present on a branch upstream (docs). Upstream (master in this case) is where you consider your current branch to have derived from.
-v - Shows commit names (instead of just SHA's) - verbose.
head - Unix command which prints n number of lines from a text
You should use the merge-base functionality which is designed to solve exactly this:
git merge-base remotes/origin/<branch> develop
git cherry master -v | tail -1
however this will only give you the first commit on branch xxx that is not in master, not the first commit on branch xxx ever. the latter would be difficult if branch xxx has been deleted and/or re-created one or more times. it that case you could try the following:
git reflog | grep checkout | grep xxx | tail -1
head -n 1 instead of tail -1? tail -1 returns the most recent commit instead of the first.
master to main which is the default branch on my current repo (master gives a fatal error from git). It gives the most recent commit instead of the oldest. The 2nd suggestion works fine.If your branch (old one) once again merged back to master doesn't give expected result. I have using python script to find initial branch commit Id.
git rev-list --first-parent changeset
--first-parent follow only the first parent commit upon seeing a merge commit.
Iterate changeset's from above command until parent branch found.
def status_check(exec_command, exec_dir=None, background=False):
if exec_dir:
os.chdir(exec_dir)
res = subprocess.Popen(exec_command, shell=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
if not background:
result = res.communicate()
return result
def findNewBranchCommits(changeset=None):
cmd = "git rev-list --first-parent "+ changeset
rev_list = status_check(cmd,self.module_dir)
rev_list = str(rev_list[0]).split('\n')
rev_list = list(filter(None, rev_list))
for x in rev_list: # Iterate until branch base point
rev_cmd = "git branch --contains " + x
rev_cmd = status_check(rev_cmd,self.module_dir)
rev_cmd = str(rev_cmd[0]).split('\n')
if(len(rev_cmd) > 2):
print "First Commit in xxx branch",x
break
findNewBranchCommits(changeset)
git rev-list --ancestry-path $(git merge-base master xxx)..xxx | tail -1
Its very simple: You can use this to directly get the hash of the first commit in the feature branch:
git log <source_branch>..<feature_branch> --pretty=format:%h
I tried this command and it worked:
git log <source_branch> <feature_branch> --oneline | tail -1
