On Wed, May 12, 2021 at 4:35 PM Ruediger Pluem <[email protected]> wrote:
>
> On 5/12/21 12:10 PM, [email protected] wrote:
> >
> > +static const char * proxyws_fallback_to_proxy_http(cmd_parms *cmd, void
> > *conf, int arg)
> > +{
> > + proxyws_dir_conf *dconf = conf;
> > + dconf->fallback_to_proxy_http = !!arg;
>
> Why !!? Shouldn't just arg be enough?
That's because fallback_to_proxy_http is a bit (i.e. unsigned int :1),
and !!arg is guaranteed to be 0 or 1 (I could have used arg != 0 too).
For instance dconf->fallback_to_proxy_http = 2 wouldn't work as
expected, maybe arg is always 0 or 1 but some compilers warn with
-Woverflow here anyway.
>
> > + dconf->fallback_to_proxy_http_set = 1;
> > + return NULL;
> > +}
> > +
> > static int proxy_wstunnel_post_config(apr_pool_t *pconf, apr_pool_t *plog,
> > apr_pool_t *ptemp, server_rec *s)
> > {
> > - fallback_to_mod_proxy_http =
> > + can_fallback_to_proxy_http =
> > (ap_find_linked_module("mod_proxy_http.c") != NULL);
> >
> > return OK;
>
> > @@ -467,6 +507,7 @@ static const command_rec ws_proxy_cmds[]
> > AP_INIT_TAKE1("ProxyWebsocketAsyncDelay", proxyws_set_asynch_delay,
> > NULL,
> > RSRC_CONF|ACCESS_CONF,
> > "amount of time to poll before going asynchronous"),
> > +
>
> Stray empty line?
Well there was one between the existing directives, so I thought I'd
separate the new and {NULL} ones with an empty line too.
>
> > {NULL}
> > };
Regards;
Yann.
