|
55d1e20e
|
2019-07-12T00:00:12
|
|
cmake: install shared library only if it was requested
$ cmake -DEVENT__LIBRARY_TYPE=static ..
...
CMake Error:
Error evaluating generator expression:
$<TARGET_FILE_DIR:event_core_shared>
No target "event_core_shared"
Fixes: #853
Fixes: 669a53f3 ("cmake: set library names to be the same as with autotools")
|
|
7806f3ec
|
2019-07-03T00:27:22
|
|
Missing <winerror.h> on win7/MinGW(MINGW32_NT-6.1)/MSYS
|
|
|
669a53f3
|
2019-07-10T00:30:12
|
|
cmake: set library names to be the same as with autotools
libtool has VERSION_INFO [1], cmake has SOVERSION/VERSION instead
(although it has different format). Also libtool has RELEASE [2] while
cmake do not have analog yet [3], hence manual symlinks should be
created.
[1]: https://www.gnu.org/software/libtool/manual/html_node/Libtool-versioning.html
[2]: https://www.gnu.org/software/libtool/manual/html_node/Release-numbers.html
[3]: https://gitlab.kitware.com/cmake/cmake/issues/17652
Plus osx has compatibility_version/current_version dylib properties and
cmake do not have separate properties for them [4], hence manual LINK_FLAGS.
And also there INSTALL_NAME_DIR property which should be adjusted too.
[4]: https://public.kitware.com/Bug/view.php?id=4383
So after all changes, here is an example before/after for osx and linux:
# osx
# autotools
.libs/libevent_pthreads-2.2.1.dylib
.libs/libevent_pthreads.dylib -> libevent_pthreads-2.2.1.dylib
/usr/local/lib/libevent_pthreads-2.2.1.dylib (compatibility version 2.0.0, current version 2.0.0)
# cmake
# before patch
lib/libevent_pthreads.2.2.0.dylib
lib/libevent_pthreads.dylib -> libevent_pthreads.2.2.0.dylib
@rpath/libevent_pthreads.2.2.0.dylib (compatibility version 2.2.0, current version 0.0.0)
# after patch
lib/libevent_pthreads-2.2.1.dylib
lib/libevent_pthreads.dylib -> libevent_pthreads-2.2.1.dylib
/vagrant/.cmake/inst/lib/libevent_pthreads-2.2.1.dylib (compatibility version 2.0.0, current version 2.0.0)
# linux
# autotools
.libs/libevent_pthreads-2.2.so.1 -> libevent_pthreads-2.2.so.1.0.0
.libs/libevent_pthreads-2.2.so.1.0.0
.libs/libevent_pthreads.so -> libevent_pthreads-2.2.so.1.0.0
# cmake
# before patch
lib/libevent_pthreads.so -> libevent_pthreads.so.2.2.0
lib/libevent_pthreads.so.2.2.0
# after patch
lib/libevent_pthreads-2.2.so -> libevent_pthreads-2.2.so.1
lib/libevent_pthreads-2.2.so.1 -> libevent_pthreads-2.2.so.1.0.0
lib/libevent_pthreads-2.2.so.1.0.0
lib/libevent_pthreads.so -> libevent_pthreads-2.2.so.1.0.0
Closes: #838 (cherry-picked)
Closes: #760
|
|
41c95abb
|
2019-07-08T14:06:55
|
|
Enable _GNU_SOURCE for Android
When targeting the Android NDK _GNU_SOURCE is not enabled by default:
```
/*
* With bionic, you always get all C and POSIX API.
*
* If you want BSD and/or GNU extensions, _BSD_SOURCE and/or _GNU_SOURCE are
* expected to be defined by callers before *any* standard header file is
* included.
*
* In our header files we test against __USE_BSD and __USE_GNU.
*/
#if defined(_GNU_SOURCE)
# define __USE_BSD 1
# define __USE_GNU 1
#endif
```
Because of this `pipe2` is not available:
```
#if defined(__USE_GNU)
int pipe2(int __fds[2], int __flags) __INTRODUCED_IN(9);
#endif
```
The function used to check if it does exist:
```
CHECK_FUNCTION_EXISTS_EX(pipe2 EVENT__HAVE_PIPE2)
```
Just check that the _linking_ succeeds, which it does, it's just not
visible in the import, leading to a warning (or error):
```
evutil.c:2637:6: error: implicit declaration of function 'pipe2' is invalid in C99 [-Werror,-Wimplicit-function-declaration]
if (pipe2(fd, O_NONBLOCK|O_CLOEXEC) == 0)
^
```
When targeting the NDK it should be safe to always opt into this. Clang
would pass the right flag for us automatically _if_ the source was C++
instead of C.
|
|
|
0d7d85c2
|
2019-07-02T10:47:47
|
|
Enable kqueue for APPLE targets
The CMAKE_CROSSCOMPILING variable is not set for Apple targets seemingly
because of cmake implementation details (more info
https://cmake.org/cmake/help/latest/variable/CMAKE_CROSSCOMPILING.html).
Since Apple targets have working kqueue implementations this check makes
sure we enable it always when those are the targets, without users
having to explicitly set EVENT__HAVE_WORKING_KQUEUE
|
|
|
497ef904
|
2019-07-01T23:54:02
|
|
Warn if forked from the event loop during event_reinit()
Calling fork() from the event loop is not a great idea, since at least
it shares some internal pipes (for handling signals) before
event_reinit() call
Closes: #833
|
|
|
5388a002
|
2019-06-26T09:36:49
|
|
autotools: do not install bufferevent_ssl.h under --disable-openssl
Refs: https://github.com/libevent/libevent/issues/760#issuecomment-502345788
|
|
|
69466cde
|
2019-06-26T01:27:20
|
|
cmake: link against shell32.lib/advapi32.lib
Fixes: #844
Refs: #760 (cmake-vs-autotools)
|
|
|
97488b2c
|
2019-06-26T01:20:53
|
|
Add README.md/Documentation into dist archive
Fixes: #841
|
|
|
bd2de483
|
2019-06-26T00:59:41
|
|
Require cmake >= 3.1.2 (for correct openssl 1.0.2 detection)
@ygj6 reported:
"My platform is MacOS 10.13.5, This problem only happens on Mac.
As written in the file CMakeLists.txt, the minimum required version of cmake is 3.1:
cmake_minimum_required(VERSION 3.1 FATAL_ERROR)
So I built this project with cmake-3.1.0, but I got the following errors:
CMake Error at /usr/local/cmake-3.1.0/share/cmake/Modules/FindOpenSSL.cmake:293 (list):
list GET given empty list
Call Stack (most recent call first):
CMakeLists.txt:824 (find_package)
CMake Error at /usr/local/cmake-3.1.0/share/cmake/Modules/FindOpenSSL.cmake:294 (list):
list GET given empty list
Call Stack (most recent call first):
CMakeLists.txt:824 (find_package)
CMake Error at /usr/local/cmake-3.1.0/share/cmake/Modules/FindOpenSSL.cmake:296 (list):
list GET given empty list
Call Stack (most recent call first):
CMakeLists.txt:824 (find_package)
CMake Error at /usr/local/cmake-3.1.0/share/cmake/Modules/FindOpenSSL.cmake:298 (list):
list GET given empty list
Call Stack (most recent call first):
CMakeLists.txt:824 (find_package)
I googled this error and got this answer:
https://bugs.archlinux.org/task/43688
It is a bug in FindOpenSSL.cmake on cmake-3.1.0 and fixed on cmake-3.1.2 .
Of course, It was successful when rebuilding with cmake-3.1.2 .
So I suggest setting the minimum version required for cmake to 3.1.2 or higher.
"
Closes: #845
Refs: https://github.com/Kitware/CMake/commit/de4ccee75a89519f95fcbcca75abc46577bfefea
|
|
|
538141eb
|
2019-06-15T23:18:05
|
|
evdns: add new options -- so-rcvbuf/so-sndbuf
This will allow to customize SO_RCVBUF/SO_SNDBUF for nameservers in this
evdns_base, you may want to adjust them if the kernel starts dropping
udp packages.
|
|
|
4f6fc092
|
2019-06-15T00:16:00
|
|
evutil: drop force_check from evutil_check_interfaces() (unused)
|
|
0de2b145
|
2019-06-14T10:19:51
|
|
evutil: set the have_checked_interfaces in evutil_check_interfaces()
Closes: #836 (cherry-picked)
Fixes: #834
|
|
|
7201062f
|
2019-06-09T20:15:37
|
|
cmake: add missing autotools targets (doxygen, uninstall, event_rpcgen.py)
Close: #832 (cherry-picked)
|
|
0374b559
|
2019-06-08T23:44:37
|
|
m4/libevent_openssl.m4: fix detection of openssl
Since commit 506df426dbeb0187bbd3654bd286b4100628fb16, OPENSSL_INCS is
added to CPPFLAGS without any white space, as a result openssl detection
fails on:
configure:10395: checking openssl/ssl.h usability
configure:10395: /home/fabrice/buildroot-test/buildroot/output/host/bin/riscv32-linux-gcc -c -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -Os -Wall -fno-strict-aliasing -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64-I/home/fabrice/buildroot-test/buildroot/output/host/bin/../riscv32-buildroot-linux-gnu/sysroot/usr/include conftest.c >&5
<command-line>:0:0: warning: "_FILE_OFFSET_BITS" redefined
<command-line>:0:0: note: this is the location of the previous definition
<command-line>:0:23: error: division by zero in #if
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
|
|
|
b002f04f
|
2019-06-03T23:22:35
|
|
doc: cmake command on Windows
Closes: #825 (cherry-pick)
|
|
|
2d4ac10e
|
2019-06-02T21:51:35
|
|
becat: remove extra SSL_CTX options usage [ci skip]
Because:
- this do not affects performance
- this breaks builds for with older openssl
|
|
|
7d569b0f
|
2019-06-02T20:20:02
|
|
Fix detection of the __has_attribute() for apple clang [ci skip]
Fixes build on the next osx env:
- Mac OS 10.9/clang 600
- Mac OS 10.8/clang 500
Refs: https://github.com/libevent/libevent/issues/447#issuecomment-497908488
|
|
891dd188
|
2019-05-30T17:38:49
|
|
https-client: correction error checking
When connecting to a non-existent HTTPS service, the "req" is not null
but the "evhttp_request_get_response_code(req)" is zero.
Closes: #822 (cherry-picked)
|
|
|
c03dabd7
|
2019-05-31T02:12:55
|
|
typo error in header file
|
|
236762a3
|
2019-05-28T12:45:59
|
|
cmake: limit MSVC to Clang-CL
|
|
|
1c573ab3
|
2019-05-25T23:41:38
|
|
http: do not name variable "sun" since this breaks solaris builds
-bash-3.2$ /opt/csw/bin/gcc -xc /dev/null -dM -E | grep '#define sun'
#define sun 1
|
|
|
51ac04ac
|
2019-05-25T17:29:25
|
|
test: mark bev_connect_hostname() as static (to avoid prototype requirement)
|
|
|
49a367e6
|
2019-05-25T17:28:53
|
|
cmake: add -Wmissing-prototypes (like autotools has)
|
|
|
4c774b6c
|
2019-05-21T10:59:05
|
|
time-test: disable buffering (mostly for windows)
|
|
|
3b1864b6
|
2019-05-16T10:25:50
|
|
Merge branch 'evbuffer-fixes-806-v2'
* evbuffer-fixes-806-v2:
evbuffer: fix last_with_datap after prepend with empty chain
test: regression for evbuffer_expand_fast_() with invalid last_with_datap
test: cover adjusting of last_with_datap in evbuffer_prepend()
Fixes: #806
|
|
|
401bd1c0
|
2019-05-11T00:17:03
|
|
evbuffer: fix last_with_datap after prepend with empty chain
last_with_datap should be adjusted only if it buf->first *was* not
empty, otherwise last_with_datap should point to the prepended chain.
|
|
|
244cacaf
|
2019-05-16T09:42:41
|
|
test: regression for evbuffer_expand_fast_() with invalid last_with_datap
Before the fix:
$ regress --no-fork evbuffer/reserve_invalid_last_with_datap
evbuffer/empty_chain_expand: [err] ../buffer.c:2138: Assertion chain == buf->first failed in evbuffer_expand_fast_
Aborted (core dumped)
This is the a shorter version of test from the #806 (with some
comments).
|
|
|
98ca3077
|
2019-05-16T09:50:43
|
|
test: cover adjusting of last_with_datap in evbuffer_prepend()
Before the fix:
$ regress evbuffer/empty_reference_prepend..
evbuffer/empty_reference_prepend: [forking]
FAIL ../test/regress_buffer.c:104: assert(chain == buf->first)
FAIL ../test/regress_buffer.c:2291: Buffer format invalid
[empty_reference_prepend FAILED]
evbuffer/empty_reference_prepend_buffer: [forking] OK
1/2 TESTS FAILED. (0 skipped)
|
|
|
c8b403a8
|
2019-05-15T21:29:03
|
|
Add getopt into dist archive
Fixes: #815
|
|
5e137f37
|
2014-12-08T17:32:07
|
|
Implement bufferevent_socket_connect_hostname_hints()
So that ai_flags (such as AI_ADDRCONFIG) can be specified.
Closes: #193 (cherry-picked with conflicts resolved)
|
|
|
76eded24
|
2019-05-12T19:06:43
|
|
Merge branch 'issue-807-accept4-getnameinfo-AF_UNIX'
* issue-807-accept4-getnameinfo-AF_UNIX:
http-server: add usage/help dialog
http: avoid use of uninitialized value for AF_UNIX/AF_LOCAL sockaddr
http-server: add ability to bind to unix-socket
build: struct sockaddr_un detection (sys/un.h, afunix.h)
Fixes: #807
|
|
|
c4de6024
|
2019-05-12T18:25:37
|
|
http-server: add usage/help dialog
|
|
|
ad51a3c1
|
2019-05-12T15:19:31
|
|
http: avoid use of uninitialized value for AF_UNIX/AF_LOCAL sockaddr
unixsock peer does not have sun_path initialized.
|
|
|
737d1beb
|
2019-05-11T18:49:28
|
|
http-server: add ability to bind to unix-socket
Usage example:
http-server -u -U /tmp/sock /tmp/no-such-dir
curl -v --unix-socket /tmp/sock 127.1:8080/foo
|
|
|
da112175
|
2019-05-12T16:12:06
|
|
build: struct sockaddr_un detection (sys/un.h, afunix.h)
- On UNIX: sys/un.h
- Since win10: afunix.h
And windows has AF_UNIX but do not have sockaddr_un (before windows
build 17061 [1]), hence the sockaddr_un detection.
[1]: https://devblogs.microsoft.com/commandline/af_unix-comes-to-windows/
|
|
2707a4ff
|
2019-05-10T23:54:14
|
|
kqueue: Avoid undefined behaviour.
As ploxiln pointed out in pull request 811 the check "newsize < 0"
is undefined behaviour (signed int overflow).
Follow the advice and check kqop->changes_size instead.
Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
Closes: #813 (cherry-picked)
|
|
|
cf8acae3
|
2019-05-07T20:53:17
|
|
Prevent integer overflow in kq_build_changes_list.
On amd64 systems with kqueue (e.g. *BSD systems) an integer overflow
could be triggered with an excessively huge amount of events.
Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
|
|
8701d0d3
|
2019-05-06T11:37:02
|
|
evdns: fix lock/unlock mismatch in evdns_close_server_port()
Closes: #809 (cherry-picked)
|
|
|
c6becb26
|
2019-04-30T08:43:27
|
|
Merge remote-tracking branch 'official/pr/804'
* official/pr/804:
Added test for evmap slot validations.
Prevent endless loop in evmap_make_space.
Enforce limit of NSIG signals.
|
|
|
9b2060c9
|
2019-04-24T22:55:24
|
|
Added test for evmap slot validations.
Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
|
|
|
bd817009
|
2019-04-24T22:54:04
|
|
Prevent endless loop in evmap_make_space.
If slot is larger than INT_MAX / 2, then the loop which increases
nentries until it is larger than slot would never return.
Also make sure that nentries * msize will never overflow INT_MAX.
Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
|
|
|
4a1088ba
|
2019-04-22T19:58:56
|
|
Enforce limit of NSIG signals.
It doesn't make sense to allow a signal number higher than NSIG.
The NSIG check already exists in signal.c for internally used
functions.
As this is a programming error of libevent consumers, this is a
purely defensive mechanism.
Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
|
|
|
176fd566
|
2019-04-16T21:12:21
|
|
Protect min_heap_push_ against integer overflow.
Converting unsigned to size_t for size of memory objects allows
proper handling of very large heaps on 64 bit systems.
Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
Closes: #799 (cherry-picked)
|
|
16d8564a
|
2019-04-18T17:10:33
|
|
le-proxy: initiate use of the Winsock DLL
Closes: #803 (cherry-picked)
|
|
1cd8830d
|
2019-04-17T15:44:59
|
|
evwatch: fix race condition
There was a race between event_base_loop and evwatch_new (adding a
prepare/check watcher while iterating over the watcher list). Only
release the mutex immediately before invoking each watcher callback,
and reacquire it immediately afterwards (same as is done for normal
event handlers).
|
|
|
04563d59
|
2019-04-13T13:22:55
|
|
Ignore build files for autotools build in subfolder too
|
|
|
c80f6be1
|
2019-04-11T22:52:10
|
|
Merge branch '21_http_extended_method'
* 21_http_extended_method:
Added http method extending
http: Update allowed_methods field from 16 to 32bits.
Fixes: #327
|
|
8dcb94a4
|
2016-01-08T13:36:20
|
|
Added http method extending
User can define his own response method by calling
evhttp_set_ext_method_cmp() on the struct http, or
evhttp_connection_set_ext_method_cmp() on the connection.
We expose a new stucture `evhttp_ext_method` which is passed to the
callback if it's set. So any field can be modified, with some exceptions
(in evhttp_method_):
If the cmp function is set, it has the ability to modify method, and
flags. Other fields will be ignored. Flags returned are OR'd with the
current flags.
Based on changes to the #282 from: Mark Ellzey <socket@gmail.com>
|
|
|
96e56beb
|
2019-04-02T15:43:35
|
|
http: Update allowed_methods field from 16 to 32bits.
|
|
|
799053db
|
2019-04-08T22:27:33
|
|
http: replace EVHTTP_REQ_UNKNOWN_ with 0
From the server perspective the evhttp_response_phrase_internal() should
not be called with 0 before this patch, it will be called with
EVHTTP_REQ_UNKNOWN_ hence this patch should not change behavior.
Fixes: 68eb526d7b ("http: add WebDAV methods support")
Fixes: #789
Fixes: #796
Reported-by: Thomas Bernard <miniupnp@free.fr>
|
|
|
55f9863b
|
2019-04-04T09:09:09
|
|
Remove experimental note for finalizers API
|
|
|
428f36e5
|
2019-04-03T23:22:54
|
|
https-client: do not try to free not initialized base
Otherwise:
$ https-client --help
Syntax:
https-client -url <https-url> [-data data-file.bin] [-ignore-cert] [-retries num] [-timeout sec] [-crt crt]
Example:
https-client -url https://ip.appspot.com/
[warn] event_base_free_: no base to free
|
|
|
35caeff5
|
2019-04-03T23:20:06
|
|
http: drop unused internal macros
|
|
|
2f184f8b
|
2019-03-26T13:33:57
|
|
evwatch: Add "prepare" and "check" watchers.
Adds two new callbacks: "prepare" watchers, which fire immediately
before we poll for I/O, and "check" watchers, which fire immediately
after we finish polling and before we process events. This allows other
event loops to be embedded into libevent's, and enables certain
performance monitoring.
Closes: #710
|
|
|
47d348a6
|
2019-04-03T07:26:21
|
|
Disable logging for tests that assume printing warnings
To avoid possible confusion
But there is still one test that has some messages on windows:
main/methods
Because this test needs >1 of avaiable methods, otherwise it will warn.
|
|
|
d4c75450
|
2019-03-25T11:13:03
|
|
Remove manually written nmake makefiles (cmake should be used instead)
This nmake stuff is out dated, and nobody wants to support it anyway.
|
|
|
93a92547
|
2019-04-02T23:25:08
|
|
appveyor: check intermediate powershell commands exit codes and terminate early
Otherwise build errors will be ignored, i.e. if build fails but regress
binary exists (copied from artifacts) it will be runned instead of newly
compiled.
|
|
|
da33f768
|
2019-04-01T02:42:55
|
|
Merge branch 'fix-uchex-warnings'
There is one more report that is false positive, see [1]:
"In bufferevent_openssl.c, pointer wm is dereferenced on line 871
before it is null checked on line 873."
[1]: https://github.com/libevent/libevent/issues/382#issuecomment-238081938
* fix-uchex-warnings:
evdns: do not check server_req twice
evrpc: do not check req twice
Fixes: #382
|
|
|
991f0ed3
|
2019-04-01T02:41:17
|
|
evdns: do not check server_req twice
Reported by µchex:
"In evdns.c, pointer server_req is null checked on line 1289 after it
is dereferenced above. Since server_req was already null checked above
on line 1243, there is no risk of crashing and the only bug is the
redundant null check (and indentation) on line 1289.
"
|
|
|
7a8cc114
|
2019-04-01T02:37:41
|
|
evrpc: do not check req twice
reported by µchex:
"In evrpc.c, pointer req is dereferenced on line 881 before it is null
checked on line 894."
|
|
|
e5b8f4c1
|
2019-04-01T01:47:00
|
|
evdns: add DNS_OPTION_NAMESERVERS_NO_DEFAULT/EVDNS_BASE_NAMESERVERS_NO_DEFAULT
- DNS_OPTION_NAMESERVERS_NO_DEFAULT
Do not "default" nameserver (i.e. "127.0.0.1:53") if there is no nameservers
in resolv.conf, (iff DNS_OPTION_NAMESERVERS is set)
- EVDNS_BASE_NAMESERVERS_NO_DEFAULT
If EVDNS_BASE_INITIALIZE_NAMESERVERS isset, do not add default
nameserver if there are no nameservers in resolv.conf (just set
DNS_OPTION_NAMESERVERS_NO_DEFAULT internally)
Fixes: #569
|
|
|
3e0dc1a6
|
2019-04-01T01:40:18
|
|
evdns: add descriptions for DNS_OPTION_*/DNS_OPTIONS_ALL
|
|
|
2ae875ed
|
2019-03-25T01:40:46
|
|
Link test/regress with event_core/event_extra over event
Due to regress linked with event and event_core (both of them includes
evthread.c) there will be two different evthread_id_fn_ variables under
mingw64:
evthread_id_fn_: &0x5294f20a8
evthread_id_fn_: &0x4ba0030a8
And because of this evthread_use_pthreads() can/will set one copy of
variables while evthread*() functions will access another, which will
break a lot of things (for example main/del_notify test).
Fixes: #792
|
|
|
b8e2f016
|
2019-03-25T01:41:09
|
|
signal: guard __cdecl definition with #ifdef
Under mingw64:
../signal.c:88:0: warning: "__cdecl" redefined
#define __cdecl
<built-in>: note: this is the location of the previous definition
https://ci.appveyor.com/project/azat/libevent/builds/23321613#L427
|
|
|
8a674243
|
2019-03-24T20:36:16
|
|
tinytest: fix parsing --timeout argument
Fixes: 15b2f41d ("tinytest: implement per-test timeout (via alarm() under !win32 only)")
|
|
|
15b2f41d
|
2019-03-24T16:44:56
|
|
tinytest: implement per-test timeout (via alarm() under !win32 only)
|
|
|
1f4f8769
|
2019-03-16T17:41:09
|
|
Merge branch 'buffer-read-size'
And after this patch set default evbuffer max read via bufferevent is
16K not 4K.
Here is some numbers for the single max read in evbuffer impact:
function client() { becat "$@" | pv > /dev/null; }
function server() { cat /dev/zero | becat -l "$@"; }
Plain bufferevent:
- 40K
$ server -R $((40<<10)) & client -R $((40<<10))
700MiB/s
- 16K *default now*
$ server & client
1.81GiB/s
- 4K
$ server -R $((4<<10)) & client -R $((4<<10))
1.05GiB/s
With OpenSSL (-S):
- 40K *default now*
$ server -S -R $((40<<10)) & client -S -R $((40<<10))
900MiB/s
- 16K *default now*
$ server -S & client -S
745MiB/s
- 4K
$ server -S -R $((4<<10)) & client -S -R $((4<<10))
593MiB/s
So as you can see without openssl 16K is faster then 40K/4K, while for
openssl 40K is still faster then 16K (I guess that this is due to with
openssl SSL_read() more at at time, while with plain we have some
allocations splits in evbuffer and maybe due to some buffer in openssl)
* buffer-read-size:
sample/becat: bufferevent cat, ncat/nc/telnet analog
Adjust evbuffer max read for bufferevents
Maximum evbuffer read configuration
Fix leaks in error path of the bufferevent_init_common_()
|
|
|
d5b24cc0
|
2019-03-05T21:34:31
|
|
sample/becat: bufferevent cat, ncat/nc/telnet analog
|
|
|
5357c3d6
|
2019-03-16T17:32:17
|
|
Adjust evbuffer max read for bufferevents
|
|
|
8c2001e9
|
2019-03-16T17:09:51
|
|
Maximum evbuffer read configuration
Before this patch evbuffer always reads 4K at a time, while this is fine
most of time you can find an example when this will decrease throughput.
So add an API to change default limit:
- evbuffer_set_max_read()
- evbuffer_get_max_read()
And a notice that most of time default is sane.
|
|
|
6995b9a8
|
2019-03-16T17:28:31
|
|
Fix leaks in error path of the bufferevent_init_common_()
|
|
|
efcc1844
|
2019-03-16T16:52:05
|
|
bench: suppress int conversion warnings
|
|
|
61e4a651
|
2019-03-16T16:37:38
|
|
Suppress int conversion warnings in getopt_long compatibility
|
|
68eb526d
|
2019-03-13T10:51:55
|
|
http: add WebDAV methods support
WebDAV introduced new HTTP methods (RFC4918):
PROPFIND, PROPPATCH, MKCOL, LOCK, UNLOCK, COPY, MOVE.
Add support of the methods.
|
|
|
bf19462a
|
2019-03-13T00:20:25
|
|
http: suppress -Wwrite-string in evhttp_parse_query_impl()
|
|
|
42d5a36b
|
2019-03-13T00:02:39
|
|
cmake: sync warnings with autotools v2
By some reason gcc reports next error:
../http.c:3330:11: warning: assignment discards ‘const’ qualifier from pointer target type [-Wdiscarded-qualifiers]
value = "";
Only under -Wwrite-strings, well this is logical, but this information
does not reflected in any documentation.
Follow-up: 8348b413 ("cmake: add various warning flags like autotools has")
f
|
|
42cc731b
|
2019-03-10T19:51:44
|
|
Update documentation [ci skip]
The documentation for building and installing has been moved into a new
Documentation section
Closes: #788 (cherry-picked)
|
|
|
56c9551e
|
2019-03-08T16:47:56
|
|
Move list of contributors into separate file
Closes: #786 (cherry-picked)
|
|
|
5ee507c8
|
2019-03-04T06:53:42
|
|
http: implement separate timeouts for read/write/connect phase
This patch allows to change timeout for next events read/write/connect
separatelly, using new API:
- client:
evhttp_connection_set_connect_timeout_tv() -- for connect
evhttp_connection_set_read_timeout_tv() -- for read
evhttp_connection_set_write_timeout_tv() -- for write
- server:
evhttp_set_read_timeout_tv() -- for read
evhttp_set_write_timeout_tv() -- for write
It also changes a logic a little, before there was next fallbacks which
does not handled in new API:
- HTTP_CONNECT_TIMEOUT
- HTTP_WRITE_TIMEOUT
- HTTP_READ_TIMEOUT
And introduce another internal flag (EVHTTP_CON_TIMEOUT_ADJUSTED) that
will be used in evrpc, which adjust evhttp_connection timeout only if it
is not default.
Fixes: #692
Fixes: #715
|
|
62df1301
|
2014-12-09T21:43:18
|
|
Add support for EV_TIMEOUT to event_base_active_by_fd
Closes: #194 (cherry-pick)
|
|
|
c4fbae3a
|
2019-03-03T19:03:29
|
|
buffer: make evbuffer_prepend() of zero-length array no-op
Refs: #774
|
|
|
b69524c0
|
2019-03-03T18:58:57
|
|
Merge branch 'evbuffer-empty-chain-handling'
* evbuffer-empty-chain-handling:
buffer: do not rely on ->off in advance_last_with_data()
buffer: fix evbuffer_remove_buffer() with empty chain in front
test: verify content of the buffer in evbuffer/remove_buffer_with_empty*
|
|
|
5b19c9f6
|
2019-03-03T16:29:52
|
|
buffer: do not rely on ->off in advance_last_with_data()
advance_last_with_data() adjusts evbuffer.last_with_datap, and if we
will have empty chain in the middle advance_last_with_data() will stop,
while it should not, since while empty chains is not regular thing they
can pops up in various places like, and while I did not look through all
of them the most tricky I would say is:
evbuffer_reverse_space()/evbuffer_commit_space()
evbuffer_add_reference()
Test case from:
https://github.com/envoyproxy/envoy/pull/6062
Fixes: #778
v2: keep last_with_datap really last with data, i.e. update only if
chain has data in it
|
|
|
fdfabbec
|
2019-03-02T22:50:00
|
|
buffer: fix evbuffer_remove_buffer() with empty chain in front
In case we have empty chain (chain that do not have any data, i.e. ->off
== 0) at the beginning of the buffer, and no more full chains to move to
the dst, we will skip moving of this empty chain, and hence
last_with_datap will not be adjusted, and things will be broken after.
Fix this by not relying on ->off, just count if we have something to
move that's it.
Test case from:
https://github.com/envoyproxy/envoy/pull/6062
Fixes: #774
|
|
|
91acedcc
|
2019-03-03T17:43:37
|
|
test: verify content of the buffer in evbuffer/remove_buffer_with_empty*
And replace spaces with tab in remove_buffer_with_empty
|
|
16a151df
|
2019-02-21T16:19:10
|
|
Remove needless check for arc4_seeded_ok
|
|
91dad75f
|
2019-02-28T17:47:51
|
|
Fix typos: warnigns -> warnings
Signed-off-by: Enji Cooper <yaneurabeya@gmail.com>
|
|
|
0b468bb0
|
2019-02-28T17:22:30
|
|
Fix typos in a comment describing the purpose of `VersionViaGit.cmake`
Signed-off-by: Enji Cooper <yaneurabeya@gmail.com>
|
|
|
5f87be42
|
2019-02-25T11:59:15
|
|
Define `_GNU_SOURCE` properly/consistently per autoconf
Although `_GNU_SOURCE` can be defined as an arbitrary #define per the
glibc docs [1], it's best to define it in a manner consistent with the way
that autoconf defines it, i.e., `1`.
While this shouldn't matter in most cases, it does when the headers from
other projects follow the poorly defined GNU convention implemented by
autoconf and are included after the libevent's util.h header. An example
failure with clang, similar to the failure I encountered, is as follows:
```
$ printf "#define _GNU_SOURCE\n#define _GNU_SOURCE 1" | clang -c -x c -
<stdin>:2:9: warning: '_GNU_SOURCE' macro redefined [-Wmacro-redefined]
^
<stdin>:1:9: note: previous definition is here
^
1 warning generated.
```
This happened when compiling python [2] with a stale homebrew util.h file from
libevent (which admittedly would not happen in a correct libevent install, as the
header should be installed under /usr/local/include/event2/util.h). However, if
both headers had been combined (which is more likely), it would have failed as
shown above.
Removing the ad hoc definition unbreaks compiling python's pyconfig.h.in header
when included after util.h from libevent.
1. http://www.gnu.org/software/libc/manual/html_node/Feature-Test-Macros.html
2. https://github.com/python/cpython/blob/master/configure.ac#L126
Closes: #773 (cherry-picked)
Signed-off-by: Enji Cooper <yaneurabeya@gmail.com>
|
|
a30d6d85
|
2019-02-26T12:33:49
|
|
Don't loose top error in SSL
Closes: #775 (cherry-picked)
|
|
b34192ba
|
2019-02-25T00:26:39
|
|
Ignore bitrise branches for travis-ci/appvyeor [ci skip]
|
|
|
2fccb967
|
2019-02-24T22:59:33
|
|
Update link to the appveyor project
Since the migration request has been hanged for a while, let's switch it
for now without beauty API URL.
Fixes: #555
|
|
|
5722e13b
|
2019-02-24T17:34:20
|
|
Match definition with prototype for bufferevent_socket_set_conn_address_fd_()
Fixes: #772
Fixes: b29207dc ("Eliminate fd conversion warnings and introduce EVUTIL_INVALID_SOCKET (windows)")
|
|
|
14eb903b
|
2019-02-24T17:25:31
|
|
Revert "test: avoid regress hanging in macOS"
After we started to use kill() over raise() everything should work just
fine.
This reverts commit a86f89d333d870e6714bd28c695ba1774df3d7f5.
Fixed-in: 728c5dc1 ("Use kill() over raise() for raising the signal (fixes osx 10.14 with kqueue)")
Fixes: #747
|
|
|
728c5dc1
|
2019-02-24T17:07:18
|
|
Use kill() over raise() for raising the signal (fixes osx 10.14 with kqueue)
On OSX 10.14+ the raise() uses pthread_kill() (verified with dtruss) and
by some reason signals that has been raised with pthread_kill() do not
received by kqueue EVFILT_SIGNAL.
While on OSX 10.11 the raise()/pthread_kill() uses plain kill() and
everything work just fine (linux also does the same, but instead of
kill() it uses tgkill())
Here is a simple reproducer that installs alarm to show that the signal
does not received by the kqueue backend:
https://gist.github.com/azat/73638b8e3b0fa563a20dadcca9e652a1
Refs: #747
Fixes: #765
|
|
8d5b14d4
|
2019-02-21T10:21:02
|
|
Merge pull request #769 from sungjungk/fix-return-handling
Add error-handling routine for arc4_seed()
|
|
|
493a1da9
|
2019-02-21T18:13:05
|
|
Add error-handling routine for arc4_seed()
|
|
|
fb134939
|
2019-02-07T11:43:14
|
|
Added DNS header mask definitions.
Currently, we do a lot of data munging with manual hex. This is ugly
and can lead to bugs. I defined the following:
_QR_MASK 0x8000U
_OP_MASK 0x7800U
_AA_MASK 0x0400U
_TC_MASK 0x0200U
_RD_MASK 0x0100U
_RA_MASK 0x0080U
_Z_MASK 0x0040U
_AD_MASK 0x0020U
_CD_MASK 0x0010U
_RCODE_MASK 0x000fU
So that we can more easily twiddle flags.
v2: make evdns flag masks unsigned literal
Closes: #756 (cherry-picked)
|
|
|
63c6669f
|
2019-02-06T00:12:35
|
|
Use the logo from the libevent.org
|