Re: [PATCH] Makefile: default to -lcurl when no CURL_CONFIG or CURLDIR

[Jonathan Nieder <jrnieder@gmail.com>]

Dave Borowitz wrote:

> Instead, if CURL_CONFIG is empty or returns an empty result (e.g. due
> to curl-config being missing), use the old behavior of falling back to
> -lcurl.
> ---
>  Makefile | 36 +++++++++++++++++++++++++-----------
>  1 file changed, 25 insertions(+), 11 deletions(-)

Sign-off?

[...]
> +++ b/Makefile
> @@ -35,7 +35,9 @@ all::
>  # transports (neither smart nor dumb).
>  #
>  # Define CURL_CONFIG to the path to a curl-config binary other than the
> -# default 'curl-config'.
> +# default 'curl-config'. If CURL_CONFIG is unset or points to a binary that
> +# is not found, defaults to the CURLDIR behavior, or if CURLDIR is not set,
> +# uses -lcurl with no additional library detection.

I'm having a little trouble parsing this but don't have any better
suggestion.

[...]
> @@ -1127,9 +1129,27 @@ ifdef NO_CURL
>  	REMOTE_CURL_NAMES =
>  else
>  	ifdef CURLDIR
> -		# Try "-Wl,-rpath=$(CURLDIR)/$(lib)" in such a case.
> -		BASIC_CFLAGS += -I$(CURLDIR)/include
> -		CURL_LIBCURL = -L$(CURLDIR)/$(lib) $(CC_LD_DYNPATH)$(CURLDIR)/$(lib) -lcurl
> +		CURL_LIBCURL=

Tiny nit: elsewhere the makefile seems to prefer having a space before
the '='.

If I explicitly set CURL_LIBCURL to the empty string and CURLDIR was
set then my setting will still override the setting below and the
behavior is unchanged from before this patch --- good.

If I explicitly set CURL_LIBCURL to empty and CURLDIR was unset then
that used to produce an error so it was an invalid configuration and
couldn't regress.

So this should be safe --- good.

> +	else
> +		CURL_CONFIG ?= curl-config

Not about this patch, but the above '?=' should probably be plain '='
for consistency with the rest of the makefile's behavior wrt envvars.

[...]
> -				$(error libcurl not detected; try setting CURLDIR)
> +                                $(error libcurl not detected or not compiled with static support)

Whitespace damage.

Except for the whitespace issues,
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>

Thanks.