# [PATCH] stash: don't leak underlying error messages

3 messages from 2012-04-11 to 2012-04-13. Participants: Ross Lagerwall, Junio C Hamano.
Thread: https://gitlist.dev/t/30216

## Ross Lagerwall, 2012-04-11 18:36

Subject: [PATCH] stash: don't leak underlying error messages
Message-ID: <1334169379-20827-1-git-send-email-rosslagerwall@gmail.com>
URL: https://gitlist.dev/e/1334169379-20827-1-git-send-email-rosslagerwall%40gmail.com

```
When running git-stash on an empty repository, don't let the underlying
error messages leak through to the surface; instead, redirect them to
/dev/null.
---
 git-stash.sh |    6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/git-stash.sh b/git-stash.sh
index fe4ab28..5c72d1b 100755
--- a/git-stash.sh
+++ b/git-stash.sh
@@ -34,8 +34,8 @@ else
 fi
 
 no_changes () {
-	git diff-index --quiet --cached HEAD --ignore-submodules -- &&
-	git diff-files --quiet --ignore-submodules &&
+	git diff-index --quiet --cached HEAD --ignore-submodules -- 2>/dev/null &&
+	git diff-files --quiet --ignore-submodules 2>/dev/null &&
 	(test -z "$untracked" || test -z "$(untracked_files)")
 }
 
@@ -67,7 +67,7 @@ create_stash () {
 	fi
 
 	# state of the base commit
-	if b_commit=$(git rev-parse --verify HEAD)
+	if b_commit=$(git rev-parse --verify HEAD 2>/dev/null)
 	then
 		head=$(git rev-list --oneline -n 1 HEAD --)
 	else
-- 
1.7.10

```

## Junio C Hamano, 2012-04-12 19:30

Subject: Re: [PATCH] stash: don't leak underlying error messages
Message-ID: <7vaa2g694j.fsf@alter.siamese.dyndns.org>
URL: https://gitlist.dev/e/7vaa2g694j.fsf%40alter.siamese.dyndns.org
In-Reply-To: <1334169379-20827-1-git-send-email-rosslagerwall@gmail.com>

```
Ross Lagerwall <rosslagerwall@gmail.com> writes:

> When running git-stash on an empty repository, don't let the underlying
> error messages leak through to the surface; instead, redirect them to
> /dev/null.
> ---

Sign-off?

Is create_stash (hence save_stash) the only operation that do not make
sense when HEAD is not born yet?  I am wondering if it makes more sense to
either:

 (1) catch the case where HEAD is not born yet a lot earlier and do not
     let the control even reach these functions (i.e. die inside the
     case/esac statement at the end of the script); or

 (2) pretend as if HEAD is a commit that records an empty tree, and not
     error out to begin with.

If either one of the above turns out to make sense, then the issue this
patch addresses becomes irrelevant, so...

>  git-stash.sh |    6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/git-stash.sh b/git-stash.sh
> index fe4ab28..5c72d1b 100755
> --- a/git-stash.sh
> +++ b/git-stash.sh
> @@ -34,8 +34,8 @@ else
>  fi
>  
>  no_changes () {
> -	git diff-index --quiet --cached HEAD --ignore-submodules -- &&
> -	git diff-files --quiet --ignore-submodules &&
> +	git diff-index --quiet --cached HEAD --ignore-submodules -- 2>/dev/null &&
> +	git diff-files --quiet --ignore-submodules 2>/dev/null &&
>  	(test -z "$untracked" || test -z "$(untracked_files)")
>  }
>  
> @@ -67,7 +67,7 @@ create_stash () {
>  	fi
>  
>  	# state of the base commit
> -	if b_commit=$(git rev-parse --verify HEAD)
> +	if b_commit=$(git rev-parse --verify HEAD 2>/dev/null)
>  	then
>  		head=$(git rev-list --oneline -n 1 HEAD --)
>  	else

```

## Ross Lagerwall, 2012-04-13 03:59

Subject: Re: [PATCH] stash: don't leak underlying error messages
Message-ID: <4F87A4BA.50009@gmail.com>
URL: https://gitlist.dev/e/4F87A4BA.50009%40gmail.com
In-Reply-To: <7vaa2g694j.fsf@alter.siamese.dyndns.org>

```
On 04/12/2012 09:30 PM, Junio C Hamano wrote:
> Sign-off?

Oops, I'm not really used to the whole sign-off thing.

> Is create_stash (hence save_stash) the only operation that do not make
> sense when HEAD is not born yet?  I am wondering if it makes more sense to
> either:
> 
>  (1) catch the case where HEAD is not born yet a lot earlier and do not
>      let the control even reach these functions (i.e. die inside the
>      case/esac statement at the end of the script); or
> 
>  (2) pretend as if HEAD is a commit that records an empty tree, and not
>      error out to begin with.
> 
> If either one of the above turns out to make sense, then the issue this
> patch addresses becomes irrelevant, so...

I think it would be more consistent if stash worked without any commits
having taken place so I'll look at (2).
-- 
Ross Lagerwall


```