[PATCH] Replace misleading message during interactive rebasing

[PATCH] Replace misleading message during interactive rebasing

[Wincent Colaiuta]

git-rebase--interactive uses git-cherry-pick under the covers to reorder
commits, which in turn means that in the event of a conflict a message
will be shown advising the user to commit the results and use the -c
switch to retain authorship after fixing the conflict.

The message is misleading because what the user really needs to do is
run "git rebase --continue"; the committing is handled by git-rebase
and the authorship of the commit message is retained automatically.

We solve this problem by using an environment variable to communicate
to git-cherry-pick that rebasing is underway and replace the misleading
error message with a more helpful one.

Signed-off-by: Wincent Colaiuta <win@wincent.com>
---

This applies on top of "master".

The question of whether the environment variable should have a leading
underscore came up on the mailing list. I don't really care at all either
way, I'd just like to see the misleading message go away. I'll leave it
up to others to decide.

Another thing to decide is whether the help text should be more than
just "run 'git rebase --continue'", but should mention
"git rebase --abort" as well. Junio, please feel free to modify the
patch if you think it would be appropriate.

 builtin-revert.c           |    8 +++++---
 git-rebase--interactive.sh |    2 ++
 2 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/builtin-revert.c b/builtin-revert.c
index a0586f9..5a57574 100644
--- a/builtin-revert.c
+++ b/builtin-revert.c
@@ -229,7 +229,7 @@ static int revert_or_cherry_pick(int argc, const char **argv)
 	unsigned char head[20];
 	struct commit *base, *next, *parent;
 	int i;
-	char *oneline, *reencoded_message = NULL;
+	char *oneline, *reencoded_message = NULL, *help_message;
 	const char *message, *encoding;
 	const char *defmsg = xstrdup(git_path("MERGE_MSG"));
 
@@ -352,11 +352,13 @@ static int revert_or_cherry_pick(int argc, const char **argv)
 		}
 		if (close(msg_fd) || commit_lock_file(&msg_file) < 0)
 			die ("Error wrapping up %s", defmsg);
+		help_message = getenv("_GIT_CHERRY_PICK_HELP");
 		fprintf(stderr, "Automatic %s failed.  "
 			"After resolving the conflicts,\n"
 			"mark the corrected paths with 'git add <paths>' "
-			"and commit the result.\n", me);
-		if (action == CHERRY_PICK) {
+			"and %s.\n", me,
+			help_message ? help_message : "commit the result");
+		if (action == CHERRY_PICK && !help_message) {
 			fprintf(stderr, "When commiting, use the option "
 				"'-c %s' to retain authorship and message.\n",
 				find_unique_abbrev(commit->object.sha1,
diff --git a/git-rebase--interactive.sh b/git-rebase--interactive.sh
index bf44b6a..e5f9810 100755
--- a/git-rebase--interactive.sh
+++ b/git-rebase--interactive.sh
@@ -117,6 +117,7 @@ pick_one () {
 		sha1=$(git rev-parse --short $sha1)
 		output warn Fast forward to $sha1
 	else
+		export _GIT_CHERRY_PICK_HELP="run 'git rebase --continue'"
 		output git cherry-pick "$@"
 	fi
 }
@@ -187,6 +188,7 @@ pick_one_preserving_merges () {
 			fi
 			;;
 		*)
+			export _GIT_CHERRY_PICK_HELP="run 'git rebase --continue'"
 			output git cherry-pick "$@" ||
 				die_with_patch $sha1 "Could not pick $sha1"
 			;;
-- 
1.5.3.6.952.g84ef

Re: [PATCH] Replace misleading message during interactive rebasing

[Johannes Sixt]

Wincent Colaiuta schrieb:
> +		export _GIT_CHERRY_PICK_HELP="run 'git rebase --continue'"

Isn't this a bashism? Should be:

		_GIT_CHERRY_PICK_HELP="run 'git rebase --continue'"
		export _GIT_CHERRY_PICK_HELP

(and the second instance of this as well, of course)

-- Hannes

Re: [PATCH] Replace misleading message during interactive rebasing

[Wincent Colaiuta]

El 27/11/2007, a las 9:31, Johannes Sixt escribió:

> Wincent Colaiuta schrieb:
>> +		export _GIT_CHERRY_PICK_HELP="run 'git rebase --continue'"
>
> Isn't this a bashism? Should be:
>
> 		_GIT_CHERRY_PICK_HELP="run 'git rebase --continue'"
> 		export _GIT_CHERRY_PICK_HELP
>
> (and the second instance of this as well, of course)

I wondered that myself before submitting the patch, but then saw that  
the same pattern was used at other places as well:

git-clone.sh:

W=$(cd "$GIT_WORK_TREE" && pwd) && export GIT_WORK_TREE="$W"

git-filter-branch.sh:

export GIT_INDEX_FILE="$(pwd)/../index"
export GIT_COMMIT=$commit
export GIT_COMMIT="$sha1"

git-quiltimport.sh:

export GIT_AUTHOR_NAME=$(sed -ne 's/Author: //p' "$tmp_info")
export GIT_AUTHOR_EMAIL=$(sed -ne 's/Email: //p' "$tmp_info")
export GIT_AUTHOR_DATE=$(sed -ne 's/Date: //p' "$tmp_info")
export SUBJECT=$(sed -ne 's/Subject: //p' "$tmp_info")

So if this is a problem, those sites will need to be changed as well.

Cheers,
Wincent

Re: [PATCH] Replace misleading message during interactive rebasing

[Jeff King]

On Tue, Nov 27, 2007 at 09:31:06AM +0100, Johannes Sixt wrote:

> Isn't this a bashism? Should be:
>
> 		_GIT_CHERRY_PICK_HELP="run 'git rebase --continue'"
> 		export _GIT_CHERRY_PICK_HELP
>
> (and the second instance of this as well, of course)

No, it's in POSIX:

http://www.opengroup.org/onlinepubs/009695399/utilities/export.html

However, I'm sure you will be shocked to learn that /bin/sh on Solaris
doesn't understand it:

$ export foo=bar
foo=bar: is not an identifier

-Peff

Re: [PATCH] Replace misleading message during interactive rebasing

[Jakub Narebski]

Jeff King wrote:

> On Tue, Nov 27, 2007 at 09:31:06AM +0100, Johannes Sixt wrote:
> 
>> Isn't this a bashism? Should be:
>>
>>              _GIT_CHERRY_PICK_HELP="run 'git rebase --continue'"
>>              export _GIT_CHERRY_PICK_HELP
>>
>> (and the second instance of this as well, of course)
> 
> No, it's in POSIX:
> 
> http://www.opengroup.org/onlinepubs/009695399/utilities/export.html
> 
> However, I'm sure you will be shocked to learn that /bin/sh on Solaris
> doesn't understand it:
> 
> $ export foo=bar
> foo=bar: is not an identifier

If I remember correctly /bin/sh on Solaris cannot be used because
of other issues (like $(...) and such).

-- 
Jakub Narebski
Warsaw, Poland
ShadeHawk on #git

Re: [PATCH] Replace misleading message during interactive rebasing

[Jeff King]

On Tue, Nov 27, 2007 at 10:49:03AM +0100, Jakub Narebski wrote:

> > However, I'm sure you will be shocked to learn that /bin/sh on Solaris
> > doesn't understand it:
> > 
> > $ export foo=bar
> > foo=bar: is not an identifier
> 
> If I remember correctly /bin/sh on Solaris cannot be used because
> of other issues (like $(...) and such).

Yes, my response was not "we can't use this" (because clearly we have
been for some time) but rather "look how crappy the Solaris shell is.
Maybe you were thinking of it?"

IOW, the construct is fine to the best of my knowledge.

-Peff

Re: [PATCH] Replace misleading message during interactive rebasing

[Jeff King]

On Tue, Nov 27, 2007 at 10:49:03AM +0100, Jakub Narebski wrote:

> If I remember correctly /bin/sh on Solaris cannot be used because
> of other issues (like $(...) and such).

BTW, your new mail/news gateway solution seems to send two messages. I
got one that was addressed _only_ to me, and then this one came to the
list (and was addressed only to the list). That means that I had to
manually re-add the list when responding to the first one (or just
assume that you responded to me off-list). Personally, I think this is
worse than the previous "send only to the list" behavior because senders
are inevitably going to fail to re-add the list (but then, I was never
as bothered by others not to be cc'd in the first place).

-Peff

Re: [PATCH] Replace misleading message during interactive rebasing

[Junio C Hamano]

Johannes Sixt <j.sixt@viscovery.net> writes:

> Wincent Colaiuta schrieb:
>> +		export _GIT_CHERRY_PICK_HELP="run 'git rebase --continue'"
>
> Isn't this a bashism?

Being an old-fashioned shell programmer myself, "export VAR=VAL" does
raise my eyebrows.  It is in POSIX but are there shells that we do
support (dash, bash, ksh -- /bin/sh on Solaris does not count) that
want their exports old-fashioned way?

Re: [PATCH] Replace misleading message during interactive rebasing

[Wincent Colaiuta]

El 28/11/2007, a las 5:37, Junio C Hamano escribió:

> Johannes Sixt <j.sixt@viscovery.net> writes:
>
>> Wincent Colaiuta schrieb:
>>> +		export _GIT_CHERRY_PICK_HELP="run 'git rebase --continue'"
>>
>> Isn't this a bashism?
>
> Being an old-fashioned shell programmer myself, "export VAR=VAL" does
> raise my eyebrows.  It is in POSIX but are there shells that we do
> support (dash, bash, ksh -- /bin/sh on Solaris does not count) that
> want their exports old-fashioned way?


As I noted earlier in thread, the idiom is used elsewhere in Git  
already. Below is a list of the instances I found where shell scripts  
use the idiom. There are also other instances (documentation, test  
scripts, even a tcl file) where the idiom is used. As far as its  
suitability I only have directory experience with Bash, and it's  
always worked in all versions that I've tried in the 2.x and 3.x  
range. The other shells, I don't know.

In any case, below I'll paste in a revised patch that doesn't use the  
export FOO=... idiom. I can also submit another patch changing the  
other usages of the idiom elsewhere in the tree, but I'd need guidance  
on how far you'd want to go (only the Git commands? also the tests?  
don't worry about the docs? etc).

El 27/11/2007, a las 9:49, Wincent Colaiuta escribió:

> I wondered that myself before submitting the patch, but then saw  
> that the same pattern was used at other places as well:
>
> git-clone.sh:
>
> W=$(cd "$GIT_WORK_TREE" && pwd) && export GIT_WORK_TREE="$W"
>
> git-filter-branch.sh:
>
> export GIT_INDEX_FILE="$(pwd)/../index"
> export GIT_COMMIT=$commit
> export GIT_COMMIT="$sha1"
>
> git-quiltimport.sh:
>
> export GIT_AUTHOR_NAME=$(sed -ne 's/Author: //p' "$tmp_info")
> export GIT_AUTHOR_EMAIL=$(sed -ne 's/Email: //p' "$tmp_info")
> export GIT_AUTHOR_DATE=$(sed -ne 's/Date: //p' "$tmp_info")
> export SUBJECT=$(sed -ne 's/Subject: //p' "$tmp_info")
>
> So if this is a problem, those sites will need to be changed as well.
>
> Cheers,
> Wincent

Here's the revised patch:
--------8<---------
 From cd11e1355011796fe16d0eb7116fae4070f2a30d Mon Sep 17 00:00:00 2001
From: Wincent Colaiuta <win@wincent.com>
Date: Mon, 26 Nov 2007 13:35:46 +0100
Subject: [PATCH] Replace misleading message during interactive rebasing

git-rebase--interactive uses git-cherry-pick under the covers to reorder
commits, which in turn means that in the event of a conflict a message
will be shown advising the user to commit the results and use the -c
switch to retain authorship after fixing the conflict.

The message is misleading because what the user really needs to do is
run "git rebase --continue"; the committing is handled by git-rebase
and the authorship of the commit message is retained automatically.

We solve this problem by using an environment variable to communicate
to git-cherry-pick that rebasing is underway and replace the misleading
error message with a more helpful one.

Signed-off-by: Wincent Colaiuta <win@wincent.com>
---
  builtin-revert.c           |    8 +++++---
  git-rebase--interactive.sh |    4 ++++
  2 files changed, 9 insertions(+), 3 deletions(-)

diff --git a/builtin-revert.c b/builtin-revert.c
index a0586f9..5a57574 100644
--- a/builtin-revert.c
+++ b/builtin-revert.c
@@ -229,7 +229,7 @@ static int revert_or_cherry_pick(int argc, const  
char **argv)
  	unsigned char head[20];
  	struct commit *base, *next, *parent;
  	int i;
-	char *oneline, *reencoded_message = NULL;
+	char *oneline, *reencoded_message = NULL, *help_message;
  	const char *message, *encoding;
  	const char *defmsg = xstrdup(git_path("MERGE_MSG"));

@@ -352,11 +352,13 @@ static int revert_or_cherry_pick(int argc, const  
char **argv)
  		}
  		if (close(msg_fd) || commit_lock_file(&msg_file) < 0)
  			die ("Error wrapping up %s", defmsg);
+		help_message = getenv("_GIT_CHERRY_PICK_HELP");
  		fprintf(stderr, "Automatic %s failed.  "
  			"After resolving the conflicts,\n"
  			"mark the corrected paths with 'git add <paths>' "
-			"and commit the result.\n", me);
-		if (action == CHERRY_PICK) {
+			"and %s.\n", me,
+			help_message ? help_message : "commit the result");
+		if (action == CHERRY_PICK && !help_message) {
  			fprintf(stderr, "When commiting, use the option "
  				"'-c %s' to retain authorship and message.\n",
  				find_unique_abbrev(commit->object.sha1,
diff --git a/git-rebase--interactive.sh b/git-rebase--interactive.sh
index bf44b6a..3552c89 100755
--- a/git-rebase--interactive.sh
+++ b/git-rebase--interactive.sh
@@ -117,6 +117,8 @@ pick_one () {
  		sha1=$(git rev-parse --short $sha1)
  		output warn Fast forward to $sha1
  	else
+		_GIT_CHERRY_PICK_HELP="run 'git rebase --continue'"
+		export _GIT_CHERRY_PICK_HELP &&
  		output git cherry-pick "$@"
  	fi
  }
@@ -187,6 +189,8 @@ pick_one_preserving_merges () {
  			fi
  			;;
  		*)
+			_GIT_CHERRY_PICK_HELP="run 'git rebase --continue'"
+			export _GIT_CHERRY_PICK_HELP &&
  			output git cherry-pick "$@" ||
  				die_with_patch $sha1 "Could not pick $sha1"
  			;;
-- 
1.5.3.6.953.gdffc

Re: [PATCH] Replace misleading message during interactive rebasing

[Johannes Schindelin]

Hi,

On Wed, 28 Nov 2007, Wincent Colaiuta wrote:

> @@ -352,11 +352,13 @@ static int revert_or_cherry_pick(int argc, const char
> **argv)
> 		}
> 		if (close(msg_fd) || commit_lock_file(&msg_file) < 0)
> 			die ("Error wrapping up %s", defmsg);
> +		help_message = getenv("_GIT_CHERRY_PICK_HELP");
> 		fprintf(stderr, "Automatic %s failed.  "
> 			"After resolving the conflicts,\n"
> 			"mark the corrected paths with 'git add <paths>' "
> -			"and commit the result.\n", me);
> -		if (action == CHERRY_PICK) {
> +			"and %s.\n", me,
> +			help_message ? help_message : "commit the result");
> +		if (action == CHERRY_PICK && !help_message) {
> 			fprintf(stderr, "When commiting, use the option "
> 				"'-c %s' to retain authorship and message.\n",
> 				find_unique_abbrev(commit->object.sha1,

What about Junio's remark that _GIT_CHERRY_PICK_HELP should rather replace 
the _complete_ message?

You could still provide the "me" and unique_abbrev parameters, so that the 
first %s in _GIT_CHERRY_PICK_HELP would be replaced by the operation, and 
the second by the sha1.  Hmm?

Ciao,
Dscho

Re: [PATCH] Replace misleading message during interactive rebasing

[Johannes Sixt]

Johannes Schindelin schrieb:
> You could still provide the "me" and unique_abbrev parameters, so that the 
> first %s in _GIT_CHERRY_PICK_HELP would be replaced by the operation, and 
> the second by the sha1.  Hmm?

Gaah! First rule to safe programming: Don't use user input as format 
strings. getenv(_GIT_CHERRY_PICK_HELP) *is* user input. In this case, the 
caller knows exactly what the put into the help string, so it should do so.

-- Hannes