Move the pragma ignoring outside the function to fix the build with gcc.
While here, add equivalent clang pragmas to mute the warning, as well.
Signed-off-by: Enji Cooper <yaneurabeya@gmail.com>
* upstream/pr/976:
github workflows: ignore previous cache
github workflows: test for mingw via cmake
cmake: replace CheckFunctionExists with CheckSymbolExists
Checking functions with `CheckFunctionExists` may
get wrong results, we should replace it with
`CheckSymbolExists`, which is recommended by the cmake
official documentation.
Before using `CheckSymbolExists`, we use
`CheckIncludeFiles` to check header files and save the
available header files in a variable that guarantees
`CheckSymbolExists` and `CheckTypeSize` to work correctly.
This approach is modeled after the cmake scripts of `curl`.
The following functions or files were not found before
modification, they can now be found:
- msys2 + mingw-8.1.0 on Windows10 or mingw-7.3.0 on Ubuntu-18.04
timerclear
timercmp
timerisset
- windows10
getaddrinfo
getnameinfo
getprotobynumber
getservbyname
putenv
strtoll
timerclear
timercmp
timerisset
- ubuntu-18.04
sys/sysctl.h
timeradd
timerclear
timercmp
timerisset
- MacOS 10.13
sys/random.h
timeradd
timerclear
timercmp
timerisset
The current cmake/LibeventConfig.cmake.in has a few problems and I'm not
sure how cleanly developed it is. It seems rater complex for the little
things I would assume it needs to do.
I found two problems that are fixed in this PR:
- If the downstream user does not explicitly set LIBEVENT_STATIC_LINK
before calling find_package(libevent) then they will not be able to
detect the static library, even if its the only one that exists. Since
this may be rather strict, I've changed the behavior so that
LIBEVENT_STATIC_LINK can be set to ON or OFF, but if unset, it defaults
to whatever configuration libevent was built as.
- The other problem is a bug. The package configuration needs to unset
CMAKE_FIND_LIBRARY_SUFFIXES after use, otherwise all packages that are
detected after libevent will be "infected" by this setting. This was a
significant problem for us, and is very hard to detect in downstream
project, because the order of dependencies will lead to different search
results.
Using `CheckFunctionExists` on Windows to check `getaddrinfo`
will get `not found`, but it actually exists. Using `CheckSymbolExists`
with headers will get correct results. Other functions such as
`getnameinfo`,`inet_ntop`,etc. have the same issue.
`CheckFunctionExistsEx` is copied from the cmake modules before
3.1.2, which is the minimum required version of libevent.
The internal module `CheckFunctionExists`of cmake can completely
replace it.
`AC_CHECK_FUNCS` can not properly check `getaddrinfo` because this
function requires some special headers on mingw.
Using `AC_CHECK_DECL` can effectively solve this issue.
Same for
- getnameinfo
- getprotobynumber
- getservbyname
- inet_ntop
- inet_pton
By default we are using CLOCK_MONOTONIC_COARSE, but if
EVENT_BASE_FLAG_PRECISE_TIMER isset, then CLOCK_MONOTONIC will be used,
however this will also enable timerfd, while this is not always what
someone wants, hence add a flag to control this (by default the old
behavior is preserved, set new flag to change it).
Fix undefined behaviour and application crash that might take
place in some rare cases after calling evdns_base_free when
there are requests in the waiting queue.
Current cleanup procedure in evdns_base_free_and_unlock
function includes 2 steps:
1. Finish all inflight requests.
2. Finish all waiting requests.
During the first step we iterate over each list in req_heads
structure and finish all requests in these lists. With current
logic finishing an inflight request (function request_finished)
removes it from the inflight requests container and forces
a wating connection to be sent (by calling
evdns_requests_pump_waiting_queue). When these new requests are
sent it is possible that they will be inserted to the list in
req_heads that we've already cleaned.
So in some cases container of the inflight requests is not empty
after this procedure and some requests are not finished and
deleted. When timeouts for these requests expire
evdns_request_timeout_callback is called but corresponding
evdns_base has been already deleted which causes undefined
behaviour and possible applicaton crash.
It is interesting to note that in old versions of libevent such
situation was not possible. This bug was introduced by the commit
14f84bbdc77d90b1d936076661443cdbf516c593. Before this commit
nameservers were deleted before finishing the requests. Therefore
it was not possible that requests from the waiting queue be sent
while we finish the inflight requests.
supposing if base is null, EVBASE_ACQUIRE_LOCK and EVBASE_RELEASE_LOCK
would get a coredump, so we add a guard for protection.
Signed-off-by: chenguolong <cgl.chenguolong@huawei.com>
* github-actions-v2:
travis: disable doxygen and coveralls, in favor of github actions
github actions: test and coverage
github actions: doxygen
Ignore truthy in yamllint (for github-actions)
test: mark common_timeout as retriable
cmake: set rpath for libraries on linux
test-export: compatible with all versions of visual studio
coverage: 'lcov --remove' need full path
Moves the thread into real-time scheduling class, as recommended in [1], it
fixes the separate test provided by @ygj6 [2] everywhere (github actions,
travis-ci, appveyor) under osx.
[1]: https://developer.apple.com/library/archive/technotes/tn2169/_index.html
[2]: dde1a6e4d0
Although even after this changes the following time-related tests failed
on travis-ci:
- no_events
- del_wait
But anyway I guess #940 can be closed, since this fixes the issue in common.
* osx-clock:
test: Use THREAD_* wrappers in del_notify/del_wait
test: move threads created with THREAD_START() to realtime scheduling class too
test: put thread into real time scheduling class on osx for better latencies
Closes: #940
* ci-improvements:
travis-ci: measure build/tests time
travis-ci: drop travis_wait (does not work with sub processes)
cmake: do not print used method (EVENT_SHOW_METHOD) while running tests
cmake: run regress test quietly like autotools (makes CI logs cleaner)
The existing error pages are very basic and don't allow for
multi-lingual support or for conformity with other pages in a web site.
The aim of the callback functionality is to allow custom error pages to
be supported for calls to evhttp_send_error() by both calling
applications and Libevent itself.
A backward-incompatible change has been made to the title of error pages
sent by evhttp_send_error(). The original version of the function used
the reason argument as part of the title. That might have unforeseen
side-effects if it contains HTML tags. Therefore the title has been
changed to always use the standard status text.
An example of the error callback can be found in this
[version](https://github.com/libevent/libevent/files/123607/http-server.zip)
of the 'http-server' sample. It will output error pages with very bright
backgrounds, the error code using a very large font size and the reason.
Closes: #323 (cherr-picked from PR)
After this patchset http-connect works with pproxy [1]:
$ pproxy -l http://:8000/ -vvv &
$ http-connect //127.1:8000 http://kernel.org:80/
[1]: https://pypi.org/project/pproxy/
* http-connect:
http-connect: do not check connection on GET cb
http-connect: set Host header (for CONNECT and GET) (like curl)
http-connect: cleanup and helpers
http: do not close connection for CONNECT
http: do not assume body for CONNECT
Fixes: #946
