| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
| |
gone (fixes segfault)
|
|
|
|
|
|
| |
event framework now calls interface to dispatch events rather than
itself implementing multiple interfaces for fdevent_process() to
be able to dispatch events generically.
|
|
|
|
|
|
|
|
|
|
|
| |
regression in mod_evhost in lighttpd 1.4.53
regression in mod_flv_streaming in lighttpd 1.4.51 - lighttpd 1.4.53
(thx moisseev)
x-ref:
"[regression] %0 pattern does not match hostnames without the domain part"
https://redmine.lighttpd.net/issues/2932
|
| |
|
| |
|
|
|
|
| |
(return value was always 0, so remove checks for other values)
|
|
|
|
|
| |
flag to indicate if socket protocol is TLS
(future use: will be needed with connection upgrade to https; not impl)
|
|
|
|
|
|
|
|
|
| |
reduce code duplication in list resizing
realloc() of NULL ptr has behavior similar to malloc()
Note that if initial size == 0, then code used to adjust size
must be += x to ensure the size is non-zero for reallocation.
(Multiplying 0 * x, e.g. power-2 resizing, will result in 0.)
|
|
|
|
| |
fix gw_backend spelling of bin-path directive in error message
|
|
|
|
| |
send Connection: close if POLLRDHUP received and read queue is empty
|
|
|
|
|
|
|
| |
x-ref:
"PVS-Studio Analysis Results"
https://redmine.lighttpd.net/boards/3/topics/8459
http://www.fly-server.ru/pvs-studio/lighttpd/
|
|
|
|
|
|
|
| |
x-ref:
"PVS-Studio Analysis Results"
https://redmine.lighttpd.net/boards/3/topics/8459
http://www.fly-server.ru/pvs-studio/lighttpd/
|
|
|
|
|
|
|
|
|
| |
honor request for x-gzip, x-bzip2 if gzip, bzip2 not requested
x-ref:
"PVS-Studio Analysis Results"
https://redmine.lighttpd.net/boards/3/topics/8459
http://www.fly-server.ru/pvs-studio/lighttpd/
|
|
|
|
|
|
|
| |
x-ref:
"PVS-Studio Analysis Results"
https://redmine.lighttpd.net/boards/3/topics/8459
http://www.fly-server.ru/pvs-studio/lighttpd/
|
|
|
|
|
| |
cease http_request_parse_reqline() unconditionally copying request line,
as request line is currently used only by mod_accesslog 'r' format
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
set con->keep_alive = 0 to indicate backend request to close connection
|
| |
|
|
|
|
| |
prefer buffer_caseless_compare() to strcasecmp()
|
| |
|
|
|
|
| |
copy request header only if we may need to log it upon error
|
| |
|
| |
|
| |
|
|
|
|
| |
carve connection_state_machine() into separate functions per state
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
process fdevents in fdevent.c
|
| |
|
| |
|
|
|
|
|
| |
move server main loop into separate func and mark hot
separate funcs for signal handling
|
| |
|
| |
|
| |
|
|
|
|
|
| |
ssl.privkey can be used to specify path to file containing private key
in lieu of concatenating certificate and private key into single .pem
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When the output of include_shell calls include_shell itself,
that second invocation must not truncate the buffer used for the
outer include_shell.
This might sound like a pathological setup in itself, but with
e.g. debian's include-conf-enabled.pl, which outputs a list of
include statements for all files in /etc/lighttpd/conf-enabled,
if any of the *.conf files in that directory invokes include_shell,
the parsing of the rest of the files in that directory is effectively
aborted.
This fixes a regression since commit
a46bc4f5de49c742e4bdc6ac7be6a4e567c8d92a in such setups.
github: closes #95
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
ssl.acme-tls-1 = "/path/to/dir" containing .crt.pem and .key.pem
named with the SNI name ("<SNI>.crt.pem" and "<SNI>.key.pem")
x-ref:
"Transport Layer Security (TLS) Application-Layer Protocol Negotiation Extension"
https://tools.ietf.org/html/rfc7301
"ACME TLS ALPN Challenge Extension" (TLS-ALPN-01)
https://tools.ietf.org/html/draft-ietf-acme-tls-alpn-05
"Support for TLS-ALPN-01"
https://redmine.lighttpd.net/issues/2931
|
|
|
|
|
| |
con->uri.scheme is maintained lowercase "http" or "https"
so scheme string comparisons need not be case-insensitive
|
|
|
|
| |
compare COPY, MOVE Destination scheme with request URI scheme
|
|
|
|
| |
quiet clang --analyze warnings where trivial to do so
|
| |
|
|
|
|
|
|
|
| |
equivalent to prior code, but more direct (legibile in code) to look up
empty string than to walk array. Might be marginally faster to walk
array when the list is short, but the lookup is also fairly quick in
that case, too.
|