Some improvements for http-server sample:
- getopt
- persistent port via -p option
- IOCP for win32 via -I
- disable buffering
- enable debug logging via -v/EVENT_DEBUG_LOGGING_ALL
- cleanup (by signal and separate error path on errors)

* sample-http-server:
  s/http-server: graceful cleanup
  s/http-server: enable debug logging if EVENT_DEBUG_LOGGING_ALL env isset
  s/http-server: turn off buffering (otherwise do output on win32)
  s/http-server: add an option to use IOCP
  s/http-server: add options (for persistent port)

Refs: libevent#709
(cherry picked from commit 9a4b8ec)

…s_()

Refs: 7c8d015 ("Free event queues even
for recursive finalizers")

(cherry picked from commit f2da619)

…/stderr

Useful for win32

(cherry picked from commit b2d4fb4)

Done with coccinelle and manual line rewrap:
  $ cat > BEV_UPCAST.cocci
  @@
  expression field_;
  expression var;
  @@

  - EVUTIL_UPCAST(var, struct bufferevent_private, field_)
  + BEV_UPCAST(var)

  $ spatch --sp-file BEV_UPCASE.cocci --in-place bufferevent*.c > /dev/null

(cherry picked from commit 7bcf576)

Although this is not a problem, since bufferevent uses finalizers and
will free itself only from the loop (well this is not a problem if you
do not play games with various event_base in different threads) it
generates questions, so rewrite it in more reliable way.

Fixes: libevent#712
(cherry picked from commit 5dc88b3)

* iocp-fixes:
  regress: test for HTTP/HTTPS with IOCP enabled
  bev_async: trigger/run only deferred callbacks
  bev_async: do not initialize timeouts multiple times
  bev_async: set "ok" on setfd if fd>=0 (like we do during creation)
  bev_async: ignore ERROR_INVALID_PARAMETER on .setfd for iocp

Closes: libevent#709
Refs: nmathewson/Libevent#160
(cherry picked from commit 3d815cf)

(cherry picked from commit 56f3bde)

Refs: libevent#255
(cherry picked from commit 903c6ac)

(cherry picked from commit f8d510f)

Fixes: 811c63f ("regress: test for HTTP/HTTPS with IOCP enabled")
(cherry picked from commit 3036f15)

We have calls to the next functions but do not check return values,
though they can be invalid and it is better to show this somehow.

Also do bufferevent_setfd() first and only after it
bufferevent_enable()/bufferevent_disable() since:
a) it is more natural
b) it will avoid extra operations
c) it will not fail first bufferevent_enable() (this is the case for
   buffbufferevent_async at least)

In this case we could add more information for issues like libevent#709

(cherry picked from commit b98d32d)

…ailed

(cherry picked from commit 57765b2)

* TT_RETRIABLE:
  Mark a lot of flacky tests with TT_RETRIABLE (for linux/win32 only)
  regress: introduce TT_RETRIABLE

Fixes: libevent#704
(cherry picked from commit 4d2f013)

(cherry picked from commit d705e8c)

Long time ago in [1] cmake build was forced to compile both libraries
(SHARED and STATIC), since this is how our autotools build works.

  [1]: 7182c2f ("cmake: build SHARED and STATIC libraries (like autoconf does)")

And there is no way to configure this (and indeed you need to do this
for MSVC for example), so let's introduce option for this --
EVENT__LIBRARY_TYPE.

Plus now we have INTERFACE libraries, that we can use internally in
libevent's cmake rules to avoid strict to _shared/_static variant of the
libraries to link with samples/tests (we prefer SHARED over STATIC for
linking).

Also bump minimal cmake required version to 3.1 by the following
reasons:
- 3.1 is required for RPATH configuration under APPLE
- 3.0 is required for add_library(INTERFACE) (did not found it in 2.8.x
documentation)
- remove extra conditions
(anyway 3.1 was release 4 years ago, so I guess that most of the systems
will have it)

(cherry picked from commit c9a073e)

MSVC does not support SHARED and STATIC libraries with the same name,
so let's just build SHARED libraries by default instead (yes we can add
prefix but let's stick with this).

The reason for this is that in windows shared libraries requires .lib
file too, but this is not static library it is imported library for
shared (doh...), for more info [1] and [2].

  [1]: https://docs.microsoft.com/en-us/windows/desktop/dlls/dynamic-link-library-creation
  [2]: https://blogs.msdn.microsoft.com/oldnewthing/20091013-00/?p=16403

And when we build both static library can and will override shared
library imported part, let's take a look at event_extra.lib:

- before patch [3]:
  $ less libevent-fail/lib/Debug/event_extra.lib | head
  ==> use library:contained_file to view a file in the archive
  rw-rw-rw- 100666/100666  59568 Nov 21 23:55 2018 event_extra_static.dir/Debug/evrpc.obj
  rw-rw-rw- 100666/100666 252219 Nov 21 23:55 2018 event_extra_static.dir/Debug/evdns.obj
  rw-rw-rw- 100666/100666 203850 Nov 21 23:55 2018 event_extra_static.dir/Debug/http.obj
  rw-rw-rw- 100666/100666  25907 Nov 21 23:55 2018 event_extra_static.dir/Debug/event_tagging.obj

  [3]: https://ci.appveyor.com/project/azat/libevent/builds/20472024/job/t0o93v042jai0dj7

- "after patch" [4] (not after but the same effect):
  $ less libevent-ok/lib/Debug/event_extra.lib | head
  ==> use library:contained_file to view a file in the archive
  --------- 0/0    509 Nov 21 23:38 2018 event_extra.dll
  ...

  [4]: https://ci.appveyor.com/project/azat/libevent/builds/20478998/job/ca9k3c76amc4qr76

Refs: libevent#691
(cherry picked from commit 90d80ef)

(cherry picked from commit 1a79fd3)

Could fail from time to time in travis-ci:
  https://travis-ci.org/libevent/libevent/jobs/458554097#L1702

Follow-up-for: fe5b071 ("Mark a lot of flacky tests with TT_RETRIABLE (for linux/win32 only)")
(cherry picked from commit 1d2ef90)

(cherry picked from commit 196e3a7)

This will fix coveralls badge

See: https://travis-ci.org/libevent/libevent/jobs/458590276#L1425
(cherry picked from commit 3e6bec7)

Fixes: libevent#726
(cherry picked from commit e5ec52d)

Fixes: 6f988ee ("Merge branch 'check-O_NONBLOCK-in-debug'")
(cherry picked from commit 9d3a415)

…nection

We should not attemp to establishe the connection if there is retry
timer active, since otherwise there will be a bug.

Imagine next situation:
  con = evhttp_connection_base_new()
  evhttp_connection_set_retries(con, 2)
  req = evhttp_request_new()
  evhttp_make_request(con, req, ...)
  # failed during connecting, and timer for 2 second scheduler (retry_ev)

Then another request scheduled for this evcon:
  evhttp_make_request(con, req, ...)
  # got request from server,
  # and now it tries to read the response from the server
  # (req.kind == EVHTTP_RESPONSE)
  #
  # but at this point retry_ev scheduled,
  # and it schedules the connect again,
  # and after the connect will succeeed, it will pick request with
  # EVHTTP_RESPONSE for sending and this is completelly wrong and will
  # fail in evhttp_make_header_response() since there is no
  # "http_server" for this evcon

This was a long standing issue, that I came across few years ago
firstly, bad only now I had time to dig into it (but right now it was
pretty simple, by limiting amount of CPU for the process and using rr
for debug to go back and forth).

(cherry picked from commit f3f7aa5)

(cherry picked from commit f33c2ce)

Closes: libevent#736 (cherry-picked)
(cherry picked from commit 972da7c)

(cherry picked from commit 51945fd)

Fixes: libevent#737
(cherry picked from commit 246f440)

Closes: libevent#738 (cherry-picked)
(cherry picked from commit 28d7221)

In 755fbf1 ("Add void* arguments to request_new and reply_new
evrpc hooks") this new functions had been introduced, but newer used,
what for? So let's use them.

(cherry picked from commit 99b231b)

This allows the object methods to be properly inherited and called via
`super(..)`, addressing breakage with python 2.x, introduced in
cfcc093.

Fixes libevent#1042.

Signed-off-by: Enji Cooper <yaneurabeya@gmail.com>
(cherry picked from commit 46c9ead)

…ers)

tinytest uses another way of detecting test failures, it uses pipe
between child and parent, and if the test function in child returns OK
it writes OK flag into pipe, and reads it in parent.

However sanitizers uses atexit handlers to detect leaks, and this will
not detect failures in case of exit() will be called from the atexit
handlers, fix this by checking status after waitpid().

(cherry picked from commit 6754740)

…oo tricky)

And cannot be suppressed with suppressions due to setup routines.

(cherry picked from commit 3b13a64)

…backs)

(cherry picked from commit cdbb237)

- http/https_filter_basic
- http/https_filter_chunk_out

(cherry picked from commit ec94a6b)

(cherry picked from commit ff4ec5f)

(cherry picked from commit 0ac3cfc)

With some modifications:
- drop coverage
- drop doxygen
- do not deploy ABI check, only artifacts

(cherry picked from commit 9adc9f1)

(cherry picked from commit ce8be23)

(cherry picked from commit e2f938c)

…x gcc)

(cherry picked from commit 7b9ba3b)

(cherry picked from commit 9da1743)

UBSAN reports:
  test/regress_buffer.c:2360:2: runtime error: null pointer passed as argument 1, which is declared to never be null
  /usr/include/string.h:140:33: note: nonnull attribute specified here

(cherry picked from commit ba19b35)

Some requests may get response (evutil_addrinfo) from gaic_server_cb,
in case of cancel_event (10000ms) will not be fast enough.

(cherry picked from commit 90bcf2d)

Fixes main/fork under ASAN (LSAN if to be precise)

* fix-signal-leak:
  select: requires reinit (otherwise it leaks signal handlers)
  poll: requires reinit (otherwise it leaks signal handlers)

(cherry picked from commit f6bfa8b)

Fixes: 512c88c ("ci/linux: add dist check")
(cherry picked from commit e9a6b12)

PR: libevent#1045

* release-2.1.12-stable-pull:
  Update ChangeLog
  ci/linux: create dist artifact only if dist archive was built
  Merge branch 'fix-signal-leak'
  test: fix leak in dns/getaddrinfo_cancel_stress
  test: fix UB in evbuffer/empty_reference_prepend_buffer
  ci: set build type to debug with sanitizers
  test: really disable bufferevent_pair_release_lock under ASAN (and fix gcc)
  test-closed: fix leak
  test-export: adjust libevent version
  cmake: set rpath for libraries on linux
  test-export: compatible with all versions of visual studio
  Bump version to 2.1.12-stable (w/o ABI breakage)
  ci: run ABI for release-*-pull branches too
  Purge travis-ci config
  Purge appveyor config
  Bump ChangeLog for 2.1.12
  Update AUTHORS for 2.1.12
  Backport github actions to 2.1
  test: add getaddrinfo(AI_ADDRCONFIG) test (off by default)
  build: remove duplicate -Wredundant-decls
  test: fix memory leaks for https (add BEV_OPT_CLOSE_ON_FREE)
  test: "fix" (with a quirk) leak in ssl/bufferevent_wm (w/o defer callbacks)
  test: disable bufferevent/bufferevent_pair_release_lock under ASAN (too tricky)
  test: detect test failures if atexit handler calls _exit(!0) (sanitizers)
  Make all classes Entry, Struct, etc) new-style classes
  buffer: do not pass NULL to memcpy() from evbuffer_pullup()
  test: do not pass NULL to memcmp() in evbuffer_datacmp() helper
  http: fix undefined-shift in EVUTIL_IS*_ helpers
  Check error code of evhttp_add_header_internal() in evhttp_parse_query_impl()
  cmake: avoid problems from use of CMAKE_USE_PTHREADS_INIT
  test/regress_testutils: use inet_addr()
  remove FindGit.cmake, improve `git describe` command
  checkpatch.sh: fix clang-format-diff usage
  checkpatch.sh: fix usage
  Fix clang-format-diff usage
  variable redefinition in win32_dispatch
  test: http/autofree_connection cleanup
  http: fix EVHTTP_CON_AUTOFREE in case of timeout (and some else)
  test: cleanup http/autofree_connection
  test: fix http/autofree_connection
  evdns: Add additional validation for values of dns options
  test: Fix test_simpleclose for Windows platform
  abi-check: abi-monitor 1.10 does not support -make -j8 (1.12 supports though)
  Add API/ABI checker (using LVC)
  Update list of cmake files for autotools dist archive
  Pass --quiet to the event_rcpgen.py (autotools already does this)
  There is typo in GetAdaptersAddresses windows library. It should be iphlpapi.dll
  Support EV_CLOSED on linux for poll(2)
  Fix EV_CLOSED detection/reporting (epoll only)
  Merge branch 'EV_CLOSED-and-EV_ET-fixes'
  bufferevent: allow setting priority on socket and openssl type
  cmake: set a default value for LIBEVENT_STATIC_LINK
  evutil_time: improve evutil_gettimeofday on Windows
  bench: Allow backend method selection
  cmake: missing test-closed binary
  Merge branch 'event_rpcgen.py-cleanup'
  cmake: replace CheckFunctionExists with CheckSymbolExists
  LibeventConfig.cmake: restore CMAKE_FIND_LIBRARY_SUFFIXES and LIBEVENT_STATIC_LINK default
  cmake: fix getaddrinfo checking error
  cmake: remove CheckFunctionExistsEx
  autoconf: fix getaddrinfo checking errors on mingw
  test-time: do not use deprecated API
  test-time: enable debug mode if EVENT_DEBUG_LOGGING_ALL env set
  increase segment refcnt only if evbuffer_add_file_segment() succeeds
  evdns: fix a crash when evdns_base with waiting requests is freed
  event_base_once: fix potential null pointer threat
  test-ratelim: add missing free
  Do not use shared global structures on CYGWIN
  test: move thread into realtime class even on EVENT__DISABLE_THREAD_SUPPORT
  test: fix compilation without thread support (EVENT__DISABLE_THREAD_SUPPORT=ON)
  test: fix bufferevent/bufferevent_connect_fail_eventcb* under osx/freebsd
  test: fix dst thread in move_pthread_to_realtime_scheduling_class (osx)
  test: fix compilation under win32 (rearrange thread_setup() code)
  test: use THREAD_* wrappers over pthread* in del_notify
  Merge branch 'osx-clock'
  cmake: do not print used method (EVENT_SHOW_METHOD) while running tests
  cmake: run regress test quietly like autotools (makes CI logs cleaner)
  Merge branch 'http-connect'
  http: do not assume body for CONNECT
  Added uninstall target check to cmakelists
  Fix compilation without OPENSSL_API_COMPAT
  evutil_time: Implements usleep() using wait funtion on Windows
  Initialize variable to 0 replace use memset function in sample/hello-world.c
  cmake: set CMAKE_{RUNTIME,LIBRARY,ARCHIVE}_OUTPUT_DIRECTORY they are not defined
  cmake: use CMAKE_LIBRARY_OUTPUT_DIRECTORY for the final shared library symlink
  test-ratelim: calculate timers bias (for slow CPUs) to avoid false-positive
  mailmap: add name/email aliases for yuangongji (name and email)
  cmake: test for find_package()
  cmake: improve package config file
  Link with iphlpapi only on windows
  Parse IPv6 scope IDs.
  Relax bufferevent_connect_hostname_emfile
  autotools: fails build when need but can not find openssl
  cmake: eliminate duplicate installation of public headers
  append to CMAKE_MODULE_PATH
  Do not use sysctl.h on linux (it had been deprecated)
  sample/https-client: link crypt32 explicitly when build with mingw-w64
  Fix compat with NetBSD >= 10
  Avoid transforming base C_FLAGS set deliberately
  tinytest: support timeout on Windows
  Merge branch 'upstream/pr/899' (evbuffer_freeze testcase enhancements)
  evbuffer_add_file: fix freeing of segment in the error path
  evutil_time: detect and use _gmtime64_s()/_gmtime64()
  http: rename bind_socket_ai() to create_bind_socket_nonblock()
  https-client: load certificates from the system cert store on Windows
  Fix checking return value of the evdns_base_resolv_conf_parse()
  cmake: fix getrandom() detection
  arc4random: replace sysctl() with getrandom (on linux)
  Upgrade autoconf (after upgrading minimum required to 2.67)
  Revert "Warn if forked from the event loop during event_reinit()"
  eliminate some C4267 warnings in Windows
  autotools: attach doxygen target into all target
  cmake: attach doxygen target into all target
  Fix memory corruption in EV_CLOSURE_EVENT_FINALIZE with debug enabled
  test: prevent duplicate event_enable_debug_mode() for TT_ENABLE_DEBUG_MODE
  test: introduce TT_ENABLE_DEBUG_MODE flag
  sample/http-server: fix parameter parsing
  sample/signal-test: fix use of uninitialized variable
  Fix typos in comments (sample/test/event-internal.h)
  sample/signal-test: add NULL checks
  Change the minimum version of automake to 1.13 and autoconf to 2.67
  Add Uninstall.cmake.in into dist archive