summaryrefslogtreecommitdiffhomepage
path: root/src/http (unfollow)
AgeCommit message (Collapse)AuthorFilesLines
2009-11-30support "*" in gzip_types, ssi_types, etcIgor Sysoev9-22/+45
2009-11-29log proxied HTTP/0.9 responses status as "009"Igor Sysoev2-4/+19
2009-11-29fix handling cached HTTP/0.9 responseIgor Sysoev2-1/+6
2009-11-29fix server_name regex named captures given by "?P<...>"Igor Sysoev1-4/+11
2009-11-25bump versionIgor Sysoev1-1/+1
2009-11-23fix building without PCRE, the bug had been introduced in r3326Igor Sysoev2-15/+23
2009-11-23bump versionIgor Sysoev1-1/+1
2009-11-17add comment from r2716 commit messageIgor Sysoev1-0/+6
2009-11-17use global perl variable in perl_destruct()/perl_free()Igor Sysoev1-16/+11
for non-mulitiplicity perl
2009-11-17bump versionIgor Sysoev1-1/+1
2009-11-16evaluate maximum captures size on configuration phaseIgor Sysoev2-1/+5
2009-11-16fix captures in "rewrite", the bug had been introduced in r3326Igor Sysoev2-4/+4
2009-11-16bump versionIgor Sysoev1-1/+1
2009-11-16backout r3325: postpone filter is a body only filterIgor Sysoev3-0/+6
2009-11-16ngx_http_degradation_moduleIgor Sysoev1-0/+223
2009-11-16regex named capturesIgor Sysoev12-237/+299
2009-11-16remove subrequest test in range header, chunked, and header filters: theyIgor Sysoev3-6/+0
are run after postpone filter which sends data only in main request context
2009-11-15allow "proxy_pass http://$backend" without URI partIgor Sysoev1-9/+14
2009-11-13ngx_regex_exec() calling optimiztion:Igor Sysoev1-1/+1
*) change NGX_REGEX_NO_MATCHED to PCRE_ERROR_NOMATCH *) declare ngx_regex_exec() as #define *) optimize SSI regex a little
2009-11-13style fix: lcf > clcfIgor Sysoev1-120/+120
2009-11-12do not log error if $r->variable was not foundIgor Sysoev1-3/+0
2009-11-12add MSIE padding for Chrome tooIgor Sysoev1-11/+11
2009-11-12r->chromeIgor Sysoev2-0/+4
2009-11-12test frequent r->header_only before three response statusIgor Sysoev1-1/+1
2009-11-12refactor gzip_vary handlingIgor Sysoev5-18/+35
2009-11-11nginx sent gzipped responses to clients those do not support gzip,Igor Sysoev1-1/+7
if "gzip_static on" and "gzip_vary off"; the bug had been introduced in r3136
2009-11-11bump versionIgor Sysoev1-1/+1
2009-11-11remove "Content-Encoding: gzip" in 304 responseIgor Sysoev1-0/+5
sent by ngx_http_gzip_static_module
2009-11-11fix "set_real_ip_from unix:" inheritanceIgor Sysoev1-4/+9
2009-11-11bump versionIgor Sysoev1-1/+1
2009-11-05export aio presence knowledge to prevent using "aio sendfile",Igor Sysoev2-30/+43
if aio does not present
2009-11-04allow to work single "set_real_ip_from unix:"Igor Sysoev1-1/+6
2009-11-04bump versionIgor Sysoev1-1/+1
2009-11-03fix segfault if there is single large_client_header_buffersIgor Sysoev1-1/+9
and a request line fills it completely
2009-11-03fix segfault if no rules are defined, introduced in r3279Igor Sysoev1-2/+2
2009-11-03fix segfault if $limit_rate was loggedIgor Sysoev1-1/+25
2009-11-03fix directive name in error message: "set_realip_from" to "set_real_ip_from"Igor Sysoev1-1/+1
2009-11-03fix directive name in error message: "realip_from" to "set_realip_from"Igor Sysoev1-1/+1
2009-11-03IPv6 support in ngx_http_access_moduleIgor Sysoev1-60/+195
2009-11-02optimize some cycles:Igor Sysoev1-3/+5
*) delete surplus variable; *) on i386/amd64 "while (n) / n--" is smaller than "while (n--)", because the platforms have no postfix operations
2009-11-02set_real_ip_from unix:Igor Sysoev1-51/+86
2009-11-02support IPv6 addresses in Real IP headersIgor Sysoev1-14/+20
2009-11-02change ngx_parse_addr() interfaceIgor Sysoev1-7/+17
2009-11-02proxy_bind, fastcgi_bind, and memcached_bindIgor Sysoev5-0/+52
2009-11-02rename ngx_peer_addr_t to ngx_addr_tIgor Sysoev1-1/+1
2009-11-02use sin6_addr.s6_addr instead of "(u_char *) & .sin6_addr"Igor Sysoev1-1/+1
2009-11-02replace inet_addr() with ngx_inet_addr()Igor Sysoev1-1/+1
2009-11-02style fixIgor Sysoev2-2/+0
2009-10-29fix segfaults if no listen directive was set in default server {} block:Igor Sysoev1-27/+27
ngx_http_add_listen() uses server's connection_pool_size and client_header_timeout values, therefore it must be called after the values have been merged, the bug had been introduced in r3218
2009-10-28style fixIgor Sysoev1-1/+1