Re: [PATCH v2] git-rebase: Teach rebase "-" shorthand.

2014-03-19 Thread Junio C Hamano
John Keeping  writes:

>> I thought your suggestion was:
>> 
>> 'rebase @{-1}' says 'Fast-fowarded HEAD to @{-1}'. It should say
>> 'Fast-forwarded HEAD to 4f407407 (rebase: allow "-" short-hand
>> for the previous branch, 2014-03-19)' instead.
>> 
>> Or it could be:
>> 
>> 'rebase @{-1}' says 'Fast-fowarded HEAD to @{-1}'. It should say
>> 'Fast-forwarded HEAD to master' instead.
>> 
>> In either case, it does not look like such a change is about
>> teaching "-" as a synonym to "@{-1}".
>
> My suggestion was specifically:
>
> 'rebase -' says 'Fast-forwarded HEAD to -'.  It should say
> 'Fast-forwarded HEAD to master' instead.

OK, it was closer to the latter.

But why is it OK to leave @{-1}, which is just as "hmm, I do not
remember what the previous branch was myself" when the user says
"@{-1}" in the output while it not OK to leave "-" in the output?

I do not think of any sane reason, and that is why I think this
improvement is not part of "teaching rebase that '-' can be used in
place of @{-1}" topic.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: [PATCH v2] git-rebase: Teach rebase "-" shorthand.

2014-03-19 Thread John Keeping
On Wed, Mar 19, 2014 at 12:41:31PM -0700, Junio C Hamano wrote:
> John Keeping  writes:
> 
> > On Wed, Mar 19, 2014 at 12:02:01PM -0700, Junio C Hamano wrote:
> >> John Keeping  writes:
> >> 
> >> > On Wed, Mar 19, 2014 at 10:53:01AM -0700, Junio C Hamano wrote:
> >> >>"rebase -" with your change still says something like this:
> >> >> 
> >> >> First, rewinding head to replay your work on top of it...
> >> >> Fast-forwarded HEAD to @{-1}.
> >> >> 
> >> >>instead of "Fast-forwarded HEAD to -".  Somebody may later want
> >> >>to "fix" this, making these two eye-candy output to be different
> >> >>from each other, and what your test expects will no longer hold
> >> >>(not that I think it is better to say "-" instead of @{-1}
> >> >>there).
> >> >
> >> > I don't think either of these is correct.  When using "-" with the
> >> > commands that already support it, I have occasionally found that "-"
> >> > isn't what I thought it was.
> >> >
> >> > Can we use `git name-rev` to put the actual name here, so that people
> >> > who have not done what they intended can hopefully notice sooner?
> >> 
> >> That sounds like a right thing to do.  It however is totally
> >> orthogonal to the change we are discussing, and should be done as a
> >> separate patch.
> >
> > Is it not part of adding support for "-"?
> 
> I thought your suggestion was:
> 
> 'rebase @{-1}' says 'Fast-fowarded HEAD to @{-1}'. It should say
> 'Fast-forwarded HEAD to 4f407407 (rebase: allow "-" short-hand
> for the previous branch, 2014-03-19)' instead.
> 
> Or it could be:
> 
> 'rebase @{-1}' says 'Fast-fowarded HEAD to @{-1}'. It should say
> 'Fast-forwarded HEAD to master' instead.
> 
> In either case, it does not look like such a change is about
> teaching "-" as a synonym to "@{-1}".

My suggestion was specifically:

'rebase -' says 'Fast-forwarded HEAD to -'.  It should say
'Fast-forwarded HEAD to master' instead.

I'm not sure it's desirable to attempt to canonicalise whatever the user
writes on the command line, but since we're special-casing '-' I think
it is a good thing to print the branch name in that case.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: [PATCH v2] git-rebase: Teach rebase "-" shorthand.

2014-03-19 Thread Junio C Hamano
John Keeping  writes:

> On Wed, Mar 19, 2014 at 12:02:01PM -0700, Junio C Hamano wrote:
>> John Keeping  writes:
>> 
>> > On Wed, Mar 19, 2014 at 10:53:01AM -0700, Junio C Hamano wrote:
>> >>"rebase -" with your change still says something like this:
>> >> 
>> >> First, rewinding head to replay your work on top of it...
>> >> Fast-forwarded HEAD to @{-1}.
>> >> 
>> >>instead of "Fast-forwarded HEAD to -".  Somebody may later want
>> >>to "fix" this, making these two eye-candy output to be different
>> >>from each other, and what your test expects will no longer hold
>> >>(not that I think it is better to say "-" instead of @{-1}
>> >>there).
>> >
>> > I don't think either of these is correct.  When using "-" with the
>> > commands that already support it, I have occasionally found that "-"
>> > isn't what I thought it was.
>> >
>> > Can we use `git name-rev` to put the actual name here, so that people
>> > who have not done what they intended can hopefully notice sooner?
>> 
>> That sounds like a right thing to do.  It however is totally
>> orthogonal to the change we are discussing, and should be done as a
>> separate patch.
>
> Is it not part of adding support for "-"?

I thought your suggestion was:

'rebase @{-1}' says 'Fast-fowarded HEAD to @{-1}'. It should say
'Fast-forwarded HEAD to 4f407407 (rebase: allow "-" short-hand
for the previous branch, 2014-03-19)' instead.

Or it could be:

'rebase @{-1}' says 'Fast-fowarded HEAD to @{-1}'. It should say
'Fast-forwarded HEAD to master' instead.

In either case, it does not look like such a change is about
teaching "-" as a synonym to "@{-1}".
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: [PATCH v2] git-rebase: Teach rebase "-" shorthand.

2014-03-19 Thread John Keeping
On Wed, Mar 19, 2014 at 12:02:01PM -0700, Junio C Hamano wrote:
> John Keeping  writes:
> 
> > On Wed, Mar 19, 2014 at 10:53:01AM -0700, Junio C Hamano wrote:
> >>"rebase -" with your change still says something like this:
> >> 
> >> First, rewinding head to replay your work on top of it...
> >> Fast-forwarded HEAD to @{-1}.
> >> 
> >>instead of "Fast-forwarded HEAD to -".  Somebody may later want
> >>to "fix" this, making these two eye-candy output to be different
> >>from each other, and what your test expects will no longer hold
> >>(not that I think it is better to say "-" instead of @{-1}
> >>there).
> >
> > I don't think either of these is correct.  When using "-" with the
> > commands that already support it, I have occasionally found that "-"
> > isn't what I thought it was.
> >
> > Can we use `git name-rev` to put the actual name here, so that people
> > who have not done what they intended can hopefully notice sooner?
> 
> That sounds like a right thing to do.  It however is totally
> orthogonal to the change we are discussing, and should be done as a
> separate patch.

Is it not part of adding support for "-"?

I'm not arguing for a change to any existing functionality, just to the
behaviour introduced by this patch, which is basically a change from
"@{-1}" to "$(git name-rev --name-only @{-1})" in the patch.  (The error
handling of name-rev appears not to be very useful here when the
previous branch has been deleted, so I don't think it's quite that
simple, but that's the principle.)
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: [PATCH v2] git-rebase: Teach rebase "-" shorthand.

2014-03-19 Thread Junio C Hamano
John Keeping  writes:

> On Wed, Mar 19, 2014 at 10:53:01AM -0700, Junio C Hamano wrote:
>>"rebase -" with your change still says something like this:
>> 
>> First, rewinding head to replay your work on top of it...
>> Fast-forwarded HEAD to @{-1}.
>> 
>>instead of "Fast-forwarded HEAD to -".  Somebody may later want
>>to "fix" this, making these two eye-candy output to be different
>>from each other, and what your test expects will no longer hold
>>(not that I think it is better to say "-" instead of @{-1}
>>there).
>
> I don't think either of these is correct.  When using "-" with the
> commands that already support it, I have occasionally found that "-"
> isn't what I thought it was.
>
> Can we use `git name-rev` to put the actual name here, so that people
> who have not done what they intended can hopefully notice sooner?

That sounds like a right thing to do.  It however is totally
orthogonal to the change we are discussing, and should be done as a
separate patch.

Thanks.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: [PATCH v2] git-rebase: Teach rebase "-" shorthand.

2014-03-19 Thread Brian Gesiak
Thank you for the feedback and tweaks!

> Is the eye-candy output to the standard output what is the most
> interesting during the execution of a rebase?  Wouldn't we be
> more interested to make sure that we did transplant the history
> on the same commit between two cases?

I agree. I'll consult the other tests to see how to write such a test.

> Can we use `git name-rev` to put the actual name here, so that people
> who have not done what they intended can hopefully notice sooner?

This sounds like a great idea! Doing so would mirror how `git checkout`
behaves; checkout informs the user of which branch they have switched
to when using the "-" shorthand: "Switched to branch 'master'".

Should I submit a new patch, or reroll this one?

- Brian Gesiak
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: [PATCH v2] git-rebase: Teach rebase "-" shorthand.

2014-03-19 Thread John Keeping
On Wed, Mar 19, 2014 at 10:53:01AM -0700, Junio C Hamano wrote:
>"rebase -" with your change still says something like this:
> 
> First, rewinding head to replay your work on top of it...
> Fast-forwarded HEAD to @{-1}.
> 
>instead of "Fast-forwarded HEAD to -".  Somebody may later want
>to "fix" this, making these two eye-candy output to be different
>from each other, and what your test expects will no longer hold
>(not that I think it is better to say "-" instead of @{-1}
>there).

I don't think either of these is correct.  When using "-" with the
commands that already support it, I have occasionally found that "-"
isn't what I thought it was.

Can we use `git name-rev` to put the actual name here, so that people
who have not done what they intended can hopefully notice sooner?
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: [PATCH v2] git-rebase: Teach rebase "-" shorthand.

2014-03-19 Thread Junio C Hamano
Brian Gesiak  writes:

> diff --git a/t/t3400-rebase.sh b/t/t3400-rebase.sh
> index 6d94b1f..6176754 100755
> --- a/t/t3400-rebase.sh
> +++ b/t/t3400-rebase.sh
> @@ -88,6 +88,17 @@ test_expect_success 'rebase from ambiguous branch name' '
>   git rebase master
>  '
>  
> +test_expect_success 'rebase using shorthand' '
> + git checkout master &&
> + git checkout -b shorthand HEAD^ &&
> + git rebase - 1>shorthand.stdout &&
> + git checkout master &&
> + git branch -D shorthand &&
> + git checkout -b shorthand HEAD^ &&
> + git rebase @{-1} 1>without_shorthand.stdout &&
> + test_i18ncmp without_shorthand.stdout shorthand.stdout
> +'

A handful of issues here:

 * "1>target" looks unconventional and wastes readers' time, forcing
   them to wonder if there is anything special going on, only to
   realize there isn't anything noteworthy.  Saying ">target" like
   everybody else does avoids attracting unnecessary attention.

 * "rebase using shorthand" is somewhat a myopic title; it assumes
   that the only short-hand relevant to rebase will be that a "-"
   stands for "@{-1}" to specify the branch we rebase the current
   branch off of.

 * The usual filename for the output from the command being tested
   is 'actual', and the usual filename for the expected output is
   'expect'.  In this case, you are verifying that the output from
   "rebase -" is the same as the output from "rebase @{-1}", so it
   is more conventional to call the former 'actual' and the latter
   'expect'.

 * Is the eye-candy output to the standard output what is the most
   interesting during the execution of a rebase?  Wouldn't we be
   more interested to make sure that we did transplant the history
   on the same commit between two cases?

   "rebase -" with your change still says something like this:

First, rewinding head to replay your work on top of it...
Fast-forwarded HEAD to @{-1}.

   instead of "Fast-forwarded HEAD to -".  Somebody may later want
   to "fix" this, making these two eye-candy output to be different
   from each other, and what your test expects will no longer hold
   (not that I think it is better to say "-" instead of @{-1}
   there).


I'll tentatively queue it with a minor tweak (see below).

Thanks.

-- >8 --
From: Brian Gesiak 
Date: Wed, 19 Mar 2014 20:02:15 +0900
Subject: [PATCH] rebase: allow "-" short-hand for the previous branch

Teach rebase the same shorthand as checkout and merge to name the
branch to rebase the current branch on; that is, that "-" means "the
branch we were previously on".

Requested-by: Tim Chase 
Signed-off-by: Brian Gesiak 
Signed-off-by: Junio C Hamano 
---
 git-rebase.sh |  4 
 t/t3400-rebase.sh | 17 +
 2 files changed, 21 insertions(+)

diff --git a/git-rebase.sh b/git-rebase.sh
index 8a3efa2..658c003 100755
--- a/git-rebase.sh
+++ b/git-rebase.sh
@@ -449,6 +449,10 @@ then
test "$fork_point" = auto && fork_point=t
;;
*)  upstream_name="$1"
+   if test "$upstream_name" = "-"
+   then
+   upstream_name="@{-1}"
+   fi
shift
;;
esac
diff --git a/t/t3400-rebase.sh b/t/t3400-rebase.sh
index 6d94b1f..80e0a95 100755
--- a/t/t3400-rebase.sh
+++ b/t/t3400-rebase.sh
@@ -88,6 +88,23 @@ test_expect_success 'rebase from ambiguous branch name' '
git rebase master
 '
 
+test_expect_success 'rebase off of the previous branch using "-"' '
+   git checkout master &&
+   git checkout HEAD^ &&
+   git rebase @{-1} >expect.messages &&
+   git merge-base master HEAD >expect.forkpoint &&
+
+   git checkout master &&
+   git checkout HEAD^ &&
+   git rebase - >actual.messages &&
+   git merge-base master HEAD >actual.forkpoint &&
+
+   test_cmp expect.forkpoint actual.forkpoint &&
+   # the next one is dubious---we may want to say "-",
+   # instead of @{-1}, in the message
+   test_i18ncmp expect.messages actual.messages
+'
+
 test_expect_success 'rebase a single mode change' '
git checkout master &&
git branch -D topic &&
-- 
1.9.1-423-g4596e3a

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


[PATCH v2] git-rebase: Teach rebase "-" shorthand.

2014-03-19 Thread Brian Gesiak
Teach rebase the same shorthand as checkout and merge; that is, that "-"
means "the branch we were previously on".

Reported-by: Tim Chase 
Signed-off-by: Brian Gesiak 
---
 git-rebase.sh |  4 
 t/t3400-rebase.sh | 11 +++
 2 files changed, 15 insertions(+)

diff --git a/git-rebase.sh b/git-rebase.sh
index 5f6732b..2c75e9f 100755
--- a/git-rebase.sh
+++ b/git-rebase.sh
@@ -453,6 +453,10 @@ then
test "$fork_point" = auto && fork_point=t
;;
*)  upstream_name="$1"
+   if test "$upstream_name" = "-"
+   then
+   upstream_name="@{-1}"
+   fi
shift
;;
esac
diff --git a/t/t3400-rebase.sh b/t/t3400-rebase.sh
index 6d94b1f..6176754 100755
--- a/t/t3400-rebase.sh
+++ b/t/t3400-rebase.sh
@@ -88,6 +88,17 @@ test_expect_success 'rebase from ambiguous branch name' '
git rebase master
 '
 
+test_expect_success 'rebase using shorthand' '
+   git checkout master &&
+   git checkout -b shorthand HEAD^ &&
+   git rebase - 1>shorthand.stdout &&
+   git checkout master &&
+   git branch -D shorthand &&
+   git checkout -b shorthand HEAD^ &&
+   git rebase @{-1} 1>without_shorthand.stdout &&
+   test_i18ncmp without_shorthand.stdout shorthand.stdout
+'
+
 test_expect_success 'rebase a single mode change' '
git checkout master &&
git branch -D topic &&
-- 
1.8.5.2 (Apple Git-48)

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html