Re: [PATCH 5/5] http: don't hardcode the value of CURL_SOCKOPT_OK
- From
Junio C Hamano <gitster@pobox.com>
- Date
- Sep 9, 2021, 23:22 UTC
- Message-ID
- <xmqqmtolwczv.fsf@gitster.g>
- In-Reply-To
- <xmqqr1dxwdbu.fsf@gitster.g>
Junio C Hamano <gitster@pobox.com> writes:
Show 27 quoted lines
> Ævar Arnfjörð Bjarmason <avarab@gmail.com> writes:
>
>> +/**
>> + * CURL_SOCKOPT_OK was added in 7.21.5, released in April 2011.
>> + */
>> +#ifndef CURL_SOCKOPT_OK
>> +#define CURL_SOCKOPT_OK 0
>> +#endif
>> +
>> /**
>> * CURLOPT_TCP_KEEPALIVE was added in 7.25.0, released in March 2012.
>> */
>> diff --git a/http.c b/http.c
>> index e38fcc34d64..c40439d39ce 100644
>> --- a/http.c
>> +++ b/http.c
>> @@ -533,7 +533,7 @@ static int sockopt_callback(void *client, curl_socket_t fd, curlsocktype type)
>> if (rc < 0)
>> warning_errno("unable to set SO_KEEPALIVE on socket");
>>
>> - return 0; /* CURL_SOCKOPT_OK only exists since curl 7.21.5 */
>> + return CURL_SOCKOPT_OK;
>> }
>
> This is much better than the one in the previous round where an
> extra CPP macro with GIT_CURL_SOCKOPT_OK_AVAILABLE or some other
> name was used to conditionally return 0 or CURL_SOCKOPT_OK.I hit <send> a bit too early.
Because the git-curl-compat.h header file is primarily to hold the GIT_CURL_HAVE_CURL_BLAH CPP macros, the fallback definition of CURL_SOCKOPT_OK added by this patch looks somewhat out of place. Doing so in http.c would be perfectly OK, though.