nginx 0.5.34 *) Change: now the full request line instead of URI only is written to error_log. *) Feature: Cygwin compatibility. Thanks to Vladimir Kutakov. *) Feature: the "merge_slashes" directive. *) Feature: the "gzip_vary" directive. *) Feature: the "server_tokens" directive. *) Feature: the "access_log" directive may be used inside the "limit_except" block. *) Bugfix: if the $server_protocol was used in FastCGI parameters and a request line length was near to the "client_header_buffer_size" directive value, then nginx issued an alert "fastcgi: the request record is too big". *) Bugfix: if a plain text HTTP/0.9 version request was made to HTTPS server, then nginx returned usual response. *) Bugfix: URL double escaping in a redirect of the "msie_refresh" directive; bug appeared in 0.5.28. *) Bugfix: a segmentation fault might occur in worker process if subrequests were used. *) Bugfix: the big responses may be transferred truncated if SSL and gzip were used. *) Bugfix: compatibility with mget. *) Bugfix: nginx did not unescape URI in the "include" SSI command. *) Bugfix: the segmentation fault was occurred on start or while reconfiguration if variable was used in the "charset" or "source_charset" directives. *) Bugfix: nginx returned the 400 response on requests like "GET http://www.domain.com HTTP/1.0". Thanks to James Oakley. *) Bugfix: a segmentation fault occurred in worker process if $date_local and $date_gmt were used outside the ngx_http_ssi_filter_module. *) Bugfix: a segmentation fault might occur in worker process if debug log was enabled. Thanks to Andrei Nigmatulin. *) Bugfix: ngx_http_memcached_module did not set $upstream_response_time. Thanks to Maxim Dounin. *) Bugfix: a worker process may got caught in an endless loop, if the memcached was used.