Re: [PATCH] http: add support for specifying the SSL version

From: Remi Galan Alfonso <remi.galan-alfonso@ensimag.grenoble-inp.fr>
Date: Aug 12, 2015, 13:33 UTC
Message-ID: <950187984.739852.1439386399510.JavaMail.zimbra@ensimag.grenoble-inp.fr>
In-Reply-To: <1439384456-16335-1-git-send-email-gitter.spiros@gmail.com>

Hello, Elia

Elia Pinto <gitter.spiros@gmail.com> writes:

Show 20 quoted lines

> +        if (ssl_version != NULL && *ssl_version) {
> +                if (!strcmp(ssl_version,"tlsv1")) {
> +                        sslversion = CURL_SSLVERSION_TLSv1;
> +                } else if (!strcmp(ssl_version,"sslv2")) {
> +                        sslversion = CURL_SSLVERSION_SSLv2;
> +                } else if (!strcmp(ssl_version,"sslv3")) {
> +                        sslversion = CURL_SSLVERSION_SSLv3;
> +#if LIBCURL_VERSION_NUM >= 0x072200
> +                } else if (!strcmp(ssl_version,"tlsv1.0")) {
> +                        sslversion = CURL_SSLVERSION_TLSv1_0;
> +                } else if (!strcmp(ssl_version,"tlsv1.1")) {
> +                        sslversion = CURL_SSLVERSION_TLSv1_1;
> +                } else if (!strcmp(ssl_version,"tlsv1.2")) {
> +                        sslversion = CURL_SSLVERSION_TLSv1_2;
> +                } else {
> +                        warning("unsupported ssl version %s : using default",
> +                        ssl_version);
> +#endif
> +                }
> +        }

I'm curious about what would happen should 'sslVersion' be set to 'tlsv1.0' (or a value that doesn't belong to the possible values) and 'LIBCURL_VERSION_NUM' be inferior to '0x072200'. Since we wouldn't go through any 'if' and would also skip the 'else' case, wouldn't we have a silent 'unsupported ssl version: using default' ?

So I think that the 'else' case should be outside of the '#if [...] #endif'.

Thanks, Rémi