On 05/04/2009 10:29 PM, Roy T. Fielding wrote:
> On May 4, 2009, at 12:22 PM, Ruediger Pluem wrote:
>> During my work on the SNI backport I noticed that during my
>> preparation of
>> the backport patch for this proposal one import line fell through the
>> cracks
>> that was contained in r760866.
>>
>> The following patch would fix this missing initialization:
>>
>> Index: modules/ssl/ssl_engine_io.c
>> ===================================================================
>> --- modules/ssl/ssl_engine_io.c (revision 771213)
>> +++ modules/ssl/ssl_engine_io.c (working copy)
>> @@ -1060,6 +1060,7 @@
>> if (sslconn->is_proxy) {
>> const char *hostname_note;
>>
>> + sc = mySrvConfig(sslconn->server);
>> if ((n = SSL_connect(filter_ctx->pssl)) <= 0) {
>> ap_log_cerror(APLOG_MARK, APLOG_INFO, 0, c,
>> "SSL Proxy connect failed");
>>
>>
>>
>> Can I just correct this error and fix it in 2.2.x?
>
> Yes, just fix it.
Thanks. Sorry for the confusion I create now, but I now
find out that this line *must* be part of the SNI backport
proposal that I am currently working on and not part of this
backport. What confused me and created my messy question was
that r760866 was a backport of code that was committed to trunk
*after* some parts of the SNI code and this line in trunk is
actually the right and needed thing to do with the SNI patches
applied, but not without.
Hope that I will be less confusing next time :-).
Regards
Rüdiger
