login
Header Space

 
 

Re: Stupid question on getting branch from yesterday

Score:
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]
To: Bill Lear <rael@...>
Cc: <git@...>
Date: Wednesday, February 14, 2007 - 5:13 pm

On Wed, 14 Feb 2007, Bill Lear wrote:

No, use that only if you want to throw away everything you did after 
that particular point.

If you just want to go and look at it more closely, with git-1.5 you do

	git checkout <SHA>

which will just detatch your HEAD from any current branch, and just move 
you there. But with an older git that doesn't like the detached state, 
you'd could instead do

	git checkout -b nonlame-branch <SHA>

which actually creates a new branch with the <SHA> as the beginning point, 
and switches to it. That will work with any version (including 1.5, of 
course).

You can then decide to re-do the branch without the lame commit. One 
particularly simple thing to do (if it's just one lame commit, and you 
don't want to *remove* it, but just fix it up a bit) is to literally check 
out the tip of tree juat AT the lame commit, and then you can use

	git commit --amend

to actually change it.

So depending on what your situation is, this sequence actually works:

	#
	# Create and check out a "fixes" branch that has the 
	# known-broken commit as its head commit
	#
	git checkout -b fixes <BROKEN-COMMIT>

	.. edit edit edit to fix the broken commit ..

	#
	# Then, just _replace_ the broken commit with the fixed state
	# by doing a "git commit --amend"
	#
	git commit -a --amend

	#
	# ok, now the "fixed-branch" is in good shape, but we
	# want to re-surrect our original 'master' branch WITH the 
	# fix, and based on the fixed branch, so we rebase the
	# master branch# _onto_ the fixed state in "fixes", with
	# the broken commit (that we do _not_ want to include) as
	# the base.
	#
	git rebase --onto fixes <BROKEN-COMMIT> master

Somebody else should verify that "git rebase" thing. I still find the 
syntax for that thing rather illogical for these kinds of things.  
Possibly because I never do it myself.

		Linus
-
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]

Messages in current thread:
Stupid question on getting branch from yesterday, Bill Lear, (Wed Feb 14, 3:38 pm)
Re: Stupid question on getting branch from yesterday, Linus Torvalds, (Wed Feb 14, 5:13 pm)
Re: Stupid question on getting branch from yesterday, Junio C Hamano, (Wed Feb 14, 5:25 pm)
Re: Stupid question on getting branch from yesterday, Junio C Hamano, (Wed Feb 14, 3:50 pm)
Re: Stupid question on getting branch from yesterday, Shawn O. Pearce, (Wed Feb 14, 3:49 pm)
Re: Stupid question on getting branch from yesterday, Junio C Hamano, (Wed Feb 14, 3:54 pm)
speck-geostationary