|
0144886e
|
2011-01-02T10:17:15
|
|
Check for POLLERR, POLLHUP and POLLNVAL for Solaris event ports
|
|
cb8059d2
|
2011-01-01T21:37:21
|
|
Fix compilation on Windows with NDEBUG
Dongsheng Song reports that when building on windows with NDEBUG, you
run into an attempt to do EVUTIL_ASSERT(x) where x is a bitfield,
which turns into _EVUTIL_NIL_CONDITION(x), which takes sizeof(x),
which is illegal. This patch fixes _EVUTIL_NIL_CONDITION to work on
bitfields too.
|
|
|
05f0195d
|
2010-12-16T12:57:23
|
|
Bump version to 2.0.10-stable-dev
|
|
|
5a12d835
|
2010-12-15T14:31:08
|
|
Add changelog for 2.0.10-stable
|
|
|
d34d21d9
|
2010-12-15T14:30:04
|
|
Add more acknowledgements to the readme for 2.0.10-stable
|
|
|
bffc411e
|
2010-12-15T14:29:50
|
|
Set the version to 2.0.10-stable
|
|
b63ab177
|
2010-12-06T14:17:44
|
|
EVUTIL_ASSERT: Use sizeof() to avoid "unused variable" warnings.
|
|
|
18adc3f0
|
2010-12-09T12:17:58
|
|
Add stuff to whatsnew based on reading include changes since August.
|
|
|
57689c44
|
2010-12-09T12:17:11
|
|
Document that the cpu_hint is only used on Windows with IOCP for now
|
|
|
bb0d2b4e
|
2010-12-09T11:47:54
|
|
Consistentize tabs
|
|
|
22f4af65
|
2010-12-09T11:43:12
|
|
Remove end-of-line whitespace
|
|
|
74f75758
|
2010-12-09T11:40:35
|
|
tweak readme, whatsnew
|
|
|
7011f9ec
|
2010-12-07T11:45:14
|
|
Fix a signed/unsigned comparison in the last commit
|
|
d23839fc
|
2010-12-07T11:43:52
|
|
Reject overlong http requests early when Expect:100-continue is set
|
|
|
4f332091
|
2010-12-02T14:09:52
|
|
tests: Use new event2 headers instead of old compatibility headers.
|
|
|
c0bf63ce
|
2010-12-02T14:13:33
|
|
tests: Use relative includes ("") instead of system includes (<>)
|
|
|
fbe64f21
|
2010-12-02T10:26:12
|
|
Use relative includes instead of system includes consistently.
|
|
|
ef5e65a4
|
2010-12-02T10:27:06
|
|
bufferevent-internal.h: Use the new event2/util.h header, not evutil.h
|
|
|
5beeec9d
|
2010-12-01T21:28:03
|
|
Correctly notify the main thread when activating an event from a subthread
|
|
2599b2d5
|
2010-11-30T11:34:26
|
|
Minor fix for IOCP shutdown handling fix
Call setsockopt() on the actual new socket, not on as->s, which was
set to INVALID_SOCKET.
|
|
|
2b0b06d7
|
2010-11-30T02:19:25
|
|
Increment version to 2.0.9-rc-dev
|
|
|
ce46db99
|
2010-11-30T01:21:00
|
|
Give the correct date for 2.0.9-rc in the changelog
|
|
|
9ce4cc4e
|
2010-11-23T11:17:24
|
|
Increment version to 2.0.9-rc
|
|
|
79957b89
|
2010-11-23T11:17:06
|
|
Bump VERSION_INFO to 5:0:0
|
|
|
0c54f176
|
2010-11-23T11:08:30
|
|
Changelog and acknowledgments for 2.0.9-rc
|
|
|
13fd242d
|
2010-11-30T00:53:16
|
|
Fix win32 build of http-server.c with unicode enabled
|
|
|
04861d5e
|
2010-11-30T00:05:54
|
|
Correctly detect timeouts during http connects
|
|
|
0faaa395
|
2010-11-29T22:44:18
|
|
Try to fix an assertion failure related to close detection
f700566c removed a line from evhttp_connection_stop_detectclose that
cleared the EVHTTP_CON_CLOSEDETECT flag. I think this was an
accident, and suspect that it may be the cause of bug 3069555.
|
|
|
29f7623c
|
2010-11-29T22:14:54
|
|
Possible crash fix when freeing an underlying bufferevent of an openssl bufferevent
|
|
|
0c4d5459
|
2010-11-29T21:56:23
|
|
Merge remote branch 'chrisd/http_continue'
|
|
fa9305f8
|
2010-11-29T18:25:04
|
|
Preliminary support for Continue expectation in evhttp.
|
|
|
18d03276
|
2010-11-29T20:39:09
|
|
make http-server example code build on win32
|
|
|
ad66dfd0
|
2010-11-29T20:13:19
|
|
Fix http-server.c compilation on freebsd
|
|
|
a12839b0
|
2010-11-29T14:25:33
|
|
Merge remote branch 'chrisd/http_fixes2'
|
|
4feedef9
|
2010-11-29T07:18:32
|
|
Add a few more comments.
|
|
|
026ac5d5
|
2010-11-26T13:29:56
|
|
add http-server to gitignore
|
|
|
4881778c
|
2010-11-26T13:28:48
|
|
Fix a misplaced ) in http_server.c
|
|
|
d40aa673
|
2010-11-26T12:59:02
|
|
Merge branch '20_empty_body'
|
|
|
4e794d5d
|
2010-10-18T13:44:05
|
|
Add a basic example of how to write a static HTTP server.
|
|
|
aab8c38b
|
2010-11-05T11:17:07
|
|
Add evhttp server alias interface, correct flagging of proxy requests.
evhttp needs to be mindful of all hostnames and addresses that clients
use to contact the main server and vhosts to know the difference between
proxy requests and non-proxy requests.
|
|
|
2e5a175b
|
2010-11-25T23:03:46
|
|
Merge remote branch 'github/20_once_fixes'
|
|
32390732
|
2010-11-23T19:40:23
|
|
Fix snprintf related failures on IRIX.
|
|
|
1cd45e56
|
2010-11-23T19:26:34
|
|
If not WIN32, include <sys/socket.h> in event2/util.h.
|
|
|
c2e5e22c
|
2010-11-23T14:09:03
|
|
xlC doesn't obey C99 comments here. autoconf isn't invoking full C99
mode.
|
|
|
c4dc3353
|
2010-11-23T00:22:01
|
|
Add some checks since lack of TAILQ_FOREACH doesn't imply lack of FIRST,
END, NEXT, or INSERT_BEFORE. Quiet some warnings in XL C.
|
|
|
2e2a3d7b
|
2010-11-23T19:09:08
|
|
Reworked AIX __ss_family workaround to use AC_STRUCT_MEMBER.
|
|
|
a3a9f6b2
|
2010-11-22T20:44:10
|
|
select comes from <sys/select.h> according to POSIX.1-2001, or from a
variety of other standard headers on older systems, but not <select.h>.
AIX build fix.
|
|
|
ec5c5aec
|
2010-11-23T20:31:28
|
|
Handle evhttp PUT/POST requests with an empty body
When we call evhttp_get_bodylen() [when transfer-encoding isn't set],
having req->ntoread == -1 means that we have no content-length. But a
request with no content-length has no body! We were treating the
absent content-length as meaning "read till closed", which only holds
for replies, not requests.
This patch also allows PATCH requests to have a body.
|
|
|
652024b6
|
2010-11-23T13:08:07
|
|
Remove _event_initialized(); make event_initialized() a function(); make it consistent on windows and non-windows
|
|
|
88be27dc
|
2010-11-23T12:26:46
|
|
Document event_get_assignment
|
|
|
e431bcd6
|
2010-11-23T12:26:34
|
|
Note that reentrant calls to libevent from logging cbs may fail badly
|
|
|
7bcace2d
|
2010-11-22T21:02:34
|
|
Fix some irix compilation warnings spotted by Kevin Bowling
|
|
|
150d7d0a
|
2010-11-22T20:24:03
|
|
Merge remote branch 'kev009/master'
|
|
|
e8749823
|
2010-11-22T16:40:31
|
|
Fix IRIX build. sa_family collides with a #define in sys/socket.h on IRIX.
|
|
|
26049c2f
|
2010-11-22T16:24:52
|
|
Merge remote branch 'github/20_getaddrinfo_cancel_v2'
|
|
|
da1bf528
|
2010-11-22T16:24:09
|
|
Add a stress test for getaddrinfo_cancel
|
|
|
568ac4fd
|
2010-11-22T15:52:34
|
|
Merge remote branches 'github/20_epoll_nochangelist_v4', 'github/20_openssl_closeonfree' and 'github/20_cloexec'
|
|
|
ece974fb
|
2010-11-22T14:13:24
|
|
Reorder backends in test.sh to match preference order in event.c
|
|
|
aed7e029
|
2010-11-22T14:10:01
|
|
Make unit tests for epoll-with-changelist pass
The only changes needed were to handle the fact that the methodname
"epoll (with changelist)" matches the environment variable
EVENT_NOEPOLL rather than the imaginary "EVENT_EPOLL (WITH CHANGELIST)".
|
|
|
9531763a
|
2010-11-14T17:52:16
|
|
Disable changelist for epoll by default because of Linux dup() bug; add an option and/or an envvar to reenable it for speed.
Rename option to control epoll changelist; make epoll changelist off by default
|
|
|
3ab578f8
|
2010-11-20T01:41:34
|
|
Make sure the CLOEXEC flag is set on fds we open for base notification
|
|
|
3a67d0bf
|
2010-11-19T17:09:30
|
|
Resolve an evport bug in the thread/forking test
|
|
|
d51b2fc6
|
2010-11-19T12:14:18
|
|
Make evdns_getaddrinfo_cancel threadsafe
|
|
|
c7cfbcf4
|
2010-11-19T12:01:05
|
|
Fix some more cancel-related bugs in getaddrinfo_async
Also imposed a new rule to make this much much simpler: no freeing
the getaddrinfo request until both dns callbacks have been invoked.
|
|
|
abf01ed1
|
2010-11-19T11:33:48
|
|
Avoid double-invocation of user callback with EVUTIL_EAI_CANCEL
|
|
|
49418612
|
2010-11-16T12:55:10
|
|
Use the US-English "canceled", not the UK "cancelled".
|
|
|
8faf223a
|
2010-11-17T00:09:10
|
|
Fix a memory leak in evhttp_uri_free.
|
|
|
4f228a1f
|
2010-11-14T19:52:18
|
|
Fix bug in bufferevent_connect on an openssl bufferevent that already had an fd
The problem was that we were using openssl's BIO code's shutdown flag
whenever BEV_OPT_CLOSE_ON_FREE was set. This made the BIO close the
socket when it was freed... but it would be freed whenever we did a
setfd on the bufferevent_openssl, even the no-op setfd in
bufferevent_connect.
So instead, we just set the shutdown flag to 0, and handle closing the
fd ourselves.
Spotted by Linus Nordberg
|
|
|
057a5146
|
2010-11-14T19:34:49
|
|
Clarify EVLOOP_* documentation to be more precise.
|
|
|
2d5e1bd0
|
2010-11-14T19:32:13
|
|
Do not let EVLOOP_ONCE exit the loop until all deferred callbacks have run
|
|
|
0617a818
|
2010-11-14T19:25:54
|
|
Make EVLOOP_ONCE ignore internal events
Merely getting an internal notification event from having an event
added or deleted from another thread should not cause
event_base_loop(base, EVLOOP_ONCE) to exit; previously, it did.
|
|
|
1ac5b230
|
2010-11-09T15:18:59
|
|
Only clear underlying callbacks when the user hasn't reset them.
|
|
|
fc7b1b00
|
2010-11-09T11:43:47
|
|
When closing a filtering bufferevent, clear callbacks on the underlying bufferevent
Previously, if BEV_OPT_CLOSE_ON_FREE wasn't set on a
bufferevent_filter or a filtering bufferevent_openssl, when we went
to free the filtering bufferevent, we'd leave the underlying
bufferevent unchanged. That's not so good, since the callbacks are
set to activate stuff in the filtering bufferevent that we're about
to free. Instead, set all the callbacks to NULL.
|
|
|
5c8a59e8
|
2010-11-09T10:19:05
|
|
Merge remote branches 'github/20_evdns_cancel_segfault_v2', 'github/20_http_close_detect', 'github/20_http_versions', 'github/20_more_http_methods', 'github/20_shutdown_iocp_listener' and 'github/20_win64_fixes'
|
|
|
a38140be
|
2010-11-09T10:14:32
|
|
Refactor http version parsing into a single function
Based on a suggestion by Chris Davis to make
evhttp_parse_response_line tolerate odd versions too.
|
|
|
3db6bc00
|
2010-11-09T10:03:00
|
|
Remove some debugging puts() calls from allow_methods test
|
|
|
52aa419b
|
2010-11-04T15:40:44
|
|
Set SO_UPDATE_ACCEPT_CONTEXT on sockets from AcceptEx so that shutdown() can work
Based on patch (and lots of debugging work) by Kelly Brock.
|
|
|
229714d1
|
2010-11-04T16:04:28
|
|
Fix a mistake in http documentation found by Julien Blache
|
|
|
05124879
|
2010-11-04T14:05:08
|
|
Never call evhttp_readcb while writing.
|
|
c76640b5
|
2010-06-01T04:45:55
|
|
Don't disable reading from the HTTP connection after sending the request to be notified of connection-close in time
|
|
|
75e3320e
|
2010-11-04T12:41:13
|
|
Units test for unexpected evhttp methods.
|
|
|
536311a4
|
2010-11-04T12:39:41
|
|
evhttp: Return 501 when we get an unrecognized method, not 400.
|
|
|
f5b391e2
|
2010-11-04T11:53:34
|
|
Tweak interface for allowed methods
|
|
|
75a73414
|
2010-11-04T11:25:35
|
|
Define enumerators for all HTTP methods, including PATCH from RFC5789
This patch defines enumerators for all HTTP methods that exist
(including PATCH introduced in RFC 5789).
It also makes them bit-masky (that's not a word, is it?), breaking
binary- but not source-code compatibility.
evhttp now stores a bitmask specifying for which methods requests to
dispatch and which ones to reject with "405 Method Not Allowed".
By default that's the ones we currently have (GET, POST, HEAD, PUT,
DELETE), thereby keeping functional compatibility (besides the minor
change that one of the other methods will now cause 405 instead of
400. But I believe that could even be considered a bug-fix).
evhttp is extended by evhttp_set_allowed_methods() with which the
user can change that bitmask.
no regressions here and my test-app still works. Haven't yet
actually tested any of the new methods.
What's obviously missing here is the special logic for the methods:
OPTIONS: We should be fine here - I believe our current dispatch
logic should work fine. Some convenience functions would be fine
though.
TRACE: I'm pretty certain we should never dispatch this to the
callbacks and simply implement the necessary functionality built-in.
CONNECT: Pretty straight-forward to implement (and considering the
framework in which we implement it very efficient too). Should
probably go built-in.
PATCH: Except for checking the RFC against our pre-dispatch logic
(there just might be some "MUST not have Some-Header" lurking
somewhere) there is nothing to be done here, this is completely up
to the user. Nothing to do.
|
|
|
9ed30de7
|
2010-11-03T12:37:37
|
|
Don't free evdns_request handles until after the callback is invoked
Previously, once the callback was scheduled, it was unsafe to cancel
a request, but there was no way to tell that. Now it is safe to
cancel a request until the callback is invoked, at which point it
isn't.
Found and diagnosed by Denis Bilenko.
|
|
|
34b84b97
|
2010-11-03T14:38:45
|
|
Fix more wn64 warnings.
|
|
|
78762383
|
2010-11-03T15:18:34
|
|
Merge branch 'http_nolegacy_v2'
|
|
|
985430ae
|
2010-11-03T15:17:57
|
|
Remove need for http_compat.h in http tests
|
|
|
0b137f45
|
2010-11-03T15:14:29
|
|
Stop accessing http request struct directly from in the unit tests.
|
|
|
22e0a9b2
|
2010-11-03T15:12:08
|
|
Add evhttp_response_code to remove one more reason to include http_struct.h
|
|
|
c91622d1
|
2010-11-03T15:04:44
|
|
Stop using event_compat.h in regress_http
|
|
|
9bb82393
|
2010-11-03T14:31:23
|
|
Convert the rest of the http tests to be non-legacy unit tests.
|
|
|
353402a8
|
2010-11-03T14:13:20
|
|
Rename the confusing "base" static variable in regress_http.c
|
|
|
8505a744
|
2010-11-03T13:55:20
|
|
Start porting http tests to not use legacy interfaces
|
|
|
647e094c
|
2010-11-02T15:19:12
|
|
Replace exact-version checks for HTTP/1.1 with >= or < checks
|
|
aa5f55fa
|
2010-11-02T13:50:57
|
|
reset "chunked" flag when sending non-chunked reply
|
|
|
ba014569
|
2010-11-02T12:42:35
|
|
Use the label_len local variable in evdns instead of recalculating it over and over
|
|
|
a3245afe
|
2010-11-01T14:23:33
|
|
Fix win32 build in response to fixes from win64 build.
|
|
|
74a91e5a
|
2010-11-01T14:16:39
|
|
fix signed/unsigned warnings in http.c
|
|
|
f8095d64
|
2010-11-01T14:15:34
|
|
Fix a typo in 7484df61c981fc33db2~
|