Re: [PATCH v3 3/3] credential: handle `credential.<partial-URL>.<key>` again

["Carlo Marcelo Arenas Belón" <carenas@gmail.com>]

On Fri, Apr 24, 2020 at 11:49:52AM +0000, Johannes Schindelin via GitGitGadget wrote:
> diff --git a/credential.c b/credential.c
> index 7dbbf26f174..c1a9ca4e485 100644
> --- a/credential.c
> +++ b/credential.c
> @@ -35,6 +35,10 @@ int credential_match(const struct credential *want,
>  #undef CHECK
>  }
>  
> +
> +static int credential_from_potentially_partial_url(struct credential *c,
> +						   const char *url);
> +
>  static int credential_config_callback(const char *var, const char *value,
>  				      void *data)
>  {

something like credential_from_url_partial might had been more grep friendly
but this would work as well.

> diff --git a/t/t0300-credentials.sh b/t/t0300-credentials.sh
> index efed3ea2955..f796bbfd48b 100755
> --- a/t/t0300-credentials.sh
> +++ b/t/t0300-credentials.sh
> @@ -448,4 +448,43 @@ test_expect_success 'credential system refuses to work with missing protocol' '
>  	test_i18ncmp expect stderr
>  '
>  
> +test_expect_success 'credential config with partial URLs' '
> +	echo "echo password=yep" | write_script git-credential-yep &&
> +	test_write_lines url=https://user@example.com/repo.git >stdin &&
> +	for partial in \
> +		example.com \
> +		user@example.com \

even if it works, configurations with a username might not be worth the
trouble to support IMHO

maybe better not to include them in the tests then either?

other than that, like the previous version (which is functionally equivalent)
should be IMHO good to go.

Carlo