[PATCH] t9200 - Allow cvs version 1.12

[PATCH] t9200 - Allow cvs version 1.12

[Mark Levedahl]

cvs v1.12 does not correctly handle "cvs co -d $DIR", which is shorthand
for "mkdir $DIR, cd $DIR, cvs co, cd -". So, use the latter form.

Also cvs v1.12 does not necessarily match cvs v1.11 in the format of
CVS/Entries, and this causes a false failure in subtest 14. Eliminate
checking CVS/Entries for this one test, but keep the test that the
created file exists and is checked out.

With these changes, all tests in t9200 pass on Cygwin using its default
cvs version 1.12.

Signed-off-by: Mark Levedahl <mlevedahl@gmail.com>
---
 t/t9200-git-cvsexportcommit.sh | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/t/t9200-git-cvsexportcommit.sh b/t/t9200-git-cvsexportcommit.sh
index 3fb3368..17cb554 100755
--- a/t/t9200-git-cvsexportcommit.sh
+++ b/t/t9200-git-cvsexportcommit.sh
@@ -28,7 +28,8 @@ rm -rf "$CVSROOT" "$CVSWORK"
 
 cvs init &&
 test -d "$CVSROOT" &&
-cvs -Q co -d "$CVSWORK" . &&
+mkdir -p "$CVSWORK" &&
+(cd "$CVSWORK" && cvs -Q co .) &&
 echo >empty &&
 git add empty &&
 git commit -q -a -m "Initial" 2>/dev/null ||
@@ -313,7 +314,6 @@ test_expect_success 'commit a file with leading spaces in the name' '
 	git commit -m "Add a file with a leading space" &&
 	id=$(git rev-parse HEAD) &&
 	git cvsexportcommit -w "$CVSWORK" -c $id &&
-	check_entries "$CVSWORK" " space/1.1/|DS/1.1/|attic_gremlin/1.3/|release-notes/1.2/" &&
 	test_cmp "$CVSWORK/ space" " space"
 
 '
-- 
1.8.3.2.0.63

Re: [PATCH] t9200 - Allow cvs version 1.12

[Junio C Hamano]

Mark Levedahl <mlevedahl@gmail.com> writes:

> cvs v1.12 does not correctly handle "cvs co -d $DIR", which is shorthand
> for "mkdir $DIR, cd $DIR, cvs co, cd -". So, use the latter form.

Hmph, I think I've been using 1.12.13 and without seeing such a
breakage.  Do you mean "exactly v1.12", not "v1.12.x series"?

> Also cvs v1.12 does not necessarily match cvs v1.11 in the format of
> CVS/Entries, and this causes a false failure in subtest 14. Eliminate
> checking CVS/Entries for this one test, but keep the test that the
> created file exists and is checked out.

Also I suspect this is not because we are expecting v1.11 format, as
v1.12.13 on my box seems to pass the test.  While your removal of
the "check_entries" step might be a valid workaround for whatever
version is shipped with Cygwin, the above problem description seems
somewhat inaccurate.

> With these changes, all tests in t9200 pass on Cygwin using its default
> cvs version 1.12.
>
> Signed-off-by: Mark Levedahl <mlevedahl@gmail.com>
> ---
>  t/t9200-git-cvsexportcommit.sh | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/t/t9200-git-cvsexportcommit.sh b/t/t9200-git-cvsexportcommit.sh
> index 3fb3368..17cb554 100755
> --- a/t/t9200-git-cvsexportcommit.sh
> +++ b/t/t9200-git-cvsexportcommit.sh
> @@ -28,7 +28,8 @@ rm -rf "$CVSROOT" "$CVSWORK"
>  
>  cvs init &&
>  test -d "$CVSROOT" &&
> -cvs -Q co -d "$CVSWORK" . &&
> +mkdir -p "$CVSWORK" &&
> +(cd "$CVSWORK" && cvs -Q co .) &&
>  echo >empty &&
>  git add empty &&
>  git commit -q -a -m "Initial" 2>/dev/null ||
> @@ -313,7 +314,6 @@ test_expect_success 'commit a file with leading spaces in the name' '
>  	git commit -m "Add a file with a leading space" &&
>  	id=$(git rev-parse HEAD) &&
>  	git cvsexportcommit -w "$CVSWORK" -c $id &&
> -	check_entries "$CVSWORK" " space/1.1/|DS/1.1/|attic_gremlin/1.3/|release-notes/1.2/" &&
>  	test_cmp "$CVSWORK/ space" " space"
>  
>  '

Re: [PATCH] t9200 - Allow cvs version 1.12

[Mark Levedahl]

On 07/15/2013 06:06 PM, Junio C Hamano wrote:
> Mark Levedahl <mlevedahl@gmail.com> writes:
>
>> cvs v1.12 does not correctly handle "cvs co -d $DIR", which is shorthand
>> for "mkdir $DIR, cd $DIR, cvs co, cd -". So, use the latter form.
> Hmph, I think I've been using 1.12.13 and without seeing such a
> breakage.  Do you mean "exactly v1.12", not "v1.12.x series"?
>
Hmm, good instincts. Cygwin includes 1.12.13 which is what I used. I 
downloaded the sources, rebuilt, everything works fine, so apparently 
the Cygwin provided cvs binary is corrupt. I apologize for the noise, 
will take this to the Cygwin list.

Mark