[PATCH] t5705: Use the correct file:// URL

[PATCH] t5705: Use the correct file:// URL

[Torsten Bögershausen]

A URL like file;//. is (no longer) supported by Git:
Typically there is no host, and RFC1738 says that file:///<path>
should be used.

Update t5705 to use a working URL.

Reported-by: Michael Blume <blume.mike@gmail.com>
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
---
 t/t5705-clone-2gb.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/t/t5705-clone-2gb.sh b/t/t5705-clone-2gb.sh
index e9783c3..191d6d3 100755
--- a/t/t5705-clone-2gb.sh
+++ b/t/t5705-clone-2gb.sh
@@ -46,7 +46,7 @@ test_expect_success CLONE_2GB 'clone - bare' '
 
 test_expect_success CLONE_2GB 'clone - with worktree, file:// protocol' '
 
-	git clone file://. clone-wt
+	git clone "file://$(pwd)" clone-wt
 
 '
 
-- 
1.9.1.dirty

Re: [PATCH] t5705: Use the correct file:// URL

[Jeff King]

On Thu, Nov 13, 2014 at 08:36:07AM +0100, Torsten Bögershausen wrote:

> A URL like file;//. is (no longer) supported by Git:
> Typically there is no host, and RFC1738 says that file:///<path>
> should be used.
> 
> Update t5705 to use a working URL.

Interesting. This looks like it was unintentionally lost in c59ab2e
(connect.c: refactor url parsing, 2013-11-28). Given RFC1738, and that
this is the first notice of it (and that it is not even a real use case,
but something questionable in the test script), it's probably OK to
declare the syntax dead and not treat it like a regression.

-Peff

Re: [PATCH] t5705: Use the correct file:// URL

[Junio C Hamano]

Jeff King <peff@peff.net> writes:

> On Thu, Nov 13, 2014 at 08:36:07AM +0100, Torsten Bögershausen wrote:
>
>> A URL like file;//. is (no longer) supported by Git:
>> Typically there is no host, and RFC1738 says that file:///<path>
>> should be used.
>> 
>> Update t5705 to use a working URL.
>
> Interesting. This looks like it was unintentionally lost in c59ab2e
> (connect.c: refactor url parsing, 2013-11-28). Given RFC1738, and that
> this is the first notice of it (and that it is not even a real use case,
> but something questionable in the test script), it's probably OK to
> declare the syntax dead and not treat it like a regression.

Yeah, I tend to agree.  I do not think there is any <scheme>:// that
lets you say things relative to something that is not "root" of the
namespace, and it was a bug the old test depended on that we did not
notice it as an error when "file://." tried to refer to "the current
directory".