7 1) OpenSSL version tests not needed on LibreSSL and BoringSSL
9 2) Fix IPV6 build breakage
11 3) Some fixes for WinCE build
20 - There are only api additions, the api is compatible with v1.7.x. But
21 there is necessarily an soname bump to 8.
23 - If you are using lws client, you mainly need to be aware the option
24 LWS_SERVER_OPTION_DO_SSL_GLOBAL_INIT is needed at context-creation time
27 - If you are using lws for serving, the above is also true but there are
28 many new features to simplify your code (and life). There is a
31 https://libwebsockets.org/lws-2.0-new-features.html
33 but basically the keywords are vhosts, mounts and plugins. You can now
34 do the web serving part from lws without any user callback code at all.
35 See ./test-server/test-server-v2.0.c for an example, it has no user
36 code for ws either since it uses the protocol plugins... that one C file
37 is all that is needed to do the whole test server function.
39 You now have the option to use a small generic ws-capable webserver
40 "lwsws" and write your ws part as a plugin. That eliminates even
41 cut-and-pasting the test server code and offers more configurable
42 features like control over http cacheability in JSON.
48 These are already in 1.7.x series
50 1) MAJOR (Windows-only) fix assert firing
52 2) MAJOR http:/1.1 connections handled by lws_return_http_status() did not
53 get sent a content-length resulting in the link hanging until the peer closed
54 it. attack.sh updated to add a test for this.
56 3) MINOR An error about hdr struct in _lws_ws_related is corrected, it's not
57 known to affect anything until after it was fixed
59 4) MINOR During the close shutdown wait state introduced at v1.7, if something
60 requests callback on writeable for the socket it will busywait until the
63 5) MAJOR Although the test server has done it for a few versions already, it
64 is now required for the user code to explicitly call
66 if (lws_http_transaction_completed(wsi))
69 when it finishes replying to a transaction in http. Previously the library
70 did it for you, but that disallowed large, long transfers with multiple
71 trips around the event loop (and cgi...).
73 6) MAJOR connections on ah waiting list that closed did not get removed from
76 7) MAJOR since we added the ability to hold an ah across http keepalive
77 transactions where more headers had already arrived, we broke the ability
78 to tell if more headers had arrived. Result was if the browser didn't
79 close the keepalive, we retained ah for the lifetime of the keepalive,
82 8) MAJOR windows-only-POLLHUP was not coming
84 9) Client should not send ext hdr if no exts
89 1) MINOR test-server gained some new switches
91 -C <file> use external SSL cert file
92 -K <file> use external SSL key file
93 -A <file> use external SSL CA cert file
95 -u <uid> set effective uid
96 -g <gid> set effective gid
98 together you can use them like this to have the test-server work with the
99 usual purchased SSL certs from an official CA.
101 --ssl -C your.crt -K your.key -A your.cer -u 99 -g 99
103 2) MINOR the OpenSSL magic to setup ECDH cipher usage is implemented in the
104 library, and the ciphers restricted to use ECDH only.
105 Using this, the lws test server can score an A at SSLLABS test
107 3) MINOR STS (SSL always) header is added to the test server if you use --ssl. With
108 that, we score A+ at SSLLABS test
110 4) MINOR daemonize function (disabled at cmake by default) is updated to work
113 5) MINOR example systemd .service file now provided for test server
114 (not installed by default)
116 6) test server html is updated with tabs and a new live server monitoring
117 feature. Input sanitization added to the js.
119 7) client connections attempted when no ah is free no longer fail, they are
120 just deferred until an ah becomes available.
122 8) The test client pays attention to if you give it an http:/ or https://
123 protocol string to its argument in URL format. If so, it stays in http[s]
124 client mode and doesn't upgrade to ws[s], allowing you to do generic http client
125 operations. Receiving transfer-encoding: chunked is supported.
127 9) If you enable -DLWS_WITH_HTTP_PROXY=1 at cmake, the test server has a
128 new URI path http://localhost:7681/proxytest If you visit here, a client
129 connection to http://example.com:80 is spawned, and the results piped on
130 to your original connection.
132 10) Also with LWS_WITH_HTTP_PROXY enabled at cmake, lws wants to link to an
133 additional library, "libhubbub". This allows lws to do html rewriting on the
134 fly, adjusting proxied urls in a lightweight and fast way.
136 11) There's a new context creation flag LWS_SERVER_OPTION_DO_SSL_GLOBAL_INIT,
137 this is included automatically if you give any other SSL-related option flag.
138 If you give no SSL-related option flag, nor this one directly, then even
139 though SSL support may be compiled in, it is never initialized nor used for the
140 whole lifetime of the lws context.
142 Conversely in order to prepare the context to use SSL, even though, eg, you
143 are not listening on SSL but will use SSL client connections later, you must
144 give this flag explicitly to make sure SSL is initialized.
150 1) MINOR APIBREAK There's a new member in struct lws_context_creation_info, ecdh_curve,
151 which lets you set the name of the ECDH curve OpenSSL should use. By
152 default (if you leave ecdh_curve NULL) it will use "prime256v1"
154 2) MINOR NEWAPI It was already possible to adopt a foreign socket that had not
155 been read from using lws_adopt_socket() since v1.7. Now you can adopt a
156 partially-used socket if you don't need SSL, by passing it what you read
157 so it can drain that before reading from the socket.
159 LWS_VISIBLE LWS_EXTERN struct lws *
160 lws_adopt_socket_readbuf(struct lws_context *context, lws_sockfd_type accept_fd,
161 const char *readbuf, size_t len);
163 3) MINOR NEWAPI CGI type "network io" subprocess execution is now possible from
166 LWS_VISIBLE LWS_EXTERN int
167 lws_cgi(struct lws *wsi, char * const *exec_array, int script_uri_path_len,
170 LWS_VISIBLE LWS_EXTERN int
171 lws_cgi_kill(struct lws *wsi);
173 To use it, you must first set the cmake option
175 $ cmake .. -DLWS_WITH_CGI=1
177 See test-server-http.c and test server path
179 http://localhost:7681/cgitest
181 stdin gets http body, you can test it with wget
183 $ echo hello > hello.txt
184 $ wget http://localhost:7681/cgitest --post-file=hello.txt -O- --quiet
188 The test script returns text/html table showing /proc/meminfo. But the cgi
189 support is complete enough to run cgit cgi.
191 4) There is a helper api for forming logging timestamps
194 lwsl_timestamp(int level, char *p, int len)
196 this generates this kind of timestamp for use as logging preamble
198 lwsts[13116]: [2016/01/25 14:52:52:8386] NOTICE: Initial logging level 7
200 5) struct lws_client_connect_info has a new member
204 If it's NULL, then everything happens as before, lws_client_connect_via_info()
205 makes a ws or wss connection to the address given.
207 If you set method to a valid http method like "GET", though, then this method
208 is used and the connection remains in http[s], it's not upgraded to ws[s].
210 So with this, you can perform http[s] client operations as well as ws[s] ones.
212 There are 4 new related callbacks
214 LWS_CALLBACK_ESTABLISHED_CLIENT_HTTP = 44,
215 LWS_CALLBACK_CLOSED_CLIENT_HTTP = 45,
216 LWS_CALLBACK_RECEIVE_CLIENT_HTTP = 46,
217 LWS_CALLBACK_COMPLETED_CLIENT_HTTP = 47,
219 6) struct lws_client_connect_info has a new member
221 const char *parent_wsi
223 if non-NULL, the client wsi is set to be a child of parent_wsi. This ensures
224 if parent_wsi closes, then the client child is closed just before.
226 7) If you're using SSL, there's a new context creation-time option flag
227 LWS_SERVER_OPTION_REDIRECT_HTTP_TO_HTTPS. If you give this, non-ssl
228 connections to the server listen port are accepted and receive a 301
229 redirect to / on the same host and port using https://
231 8) User code may set per-connection extension options now, using a new api
232 "lws_set_extension_option()".
234 This should be called from the ESTABLISHED callback like this
236 lws_set_extension_option(wsi, "permessage-deflate",
237 "rx_buf_size", "12"); /* 1 << 12 */
239 If the extension is not active (missing or not negotiated for the
240 connection, or extensions are disabled on the library) the call is
241 just returns -1. Otherwise the connection's extension has its
242 named option changed.
244 The extension may decide to alter or disallow the change, in the
245 example above permessage-deflate restricts the size of his rx
246 output buffer also considering the protocol's rx_buf_size member.
249 New application lwsws
250 ---------------------
252 A libwebsockets-based general webserver is built by default now, lwsws.
254 It's configured by JSON, by default in
258 which contains global lws context settings like this
271 which contains zero or more files describing vhosts, like this
275 { "name": "warmcat.com",
277 "host-ssl-key": "/etc/pki/tls/private/warmcat.com.key",
278 "host-ssl-cert": "/etc/pki/tls/certs/warmcat.com.crt",
279 "host-ssl-ca": "/etc/pki/tls/certs/warmcat.com.cer",
282 { "home": "file:///var/www/warmcat.com" },
283 { "default": "index.html" }
299 1) There is now a "permessage-deflate" / RFC7692 implementation. It's very
300 similar to "deflate-frame" we have offered for a long while; deflate-frame is
301 now provided as an alias of permessage-deflate.
303 The main differences are that the new permessage-deflate implementation:
305 - properly performs streaming respecting input and output buffer limits. The
306 old deflate-frame implementation could only work on complete deflate input
307 and produce complete inflate output for each frame. The new implementation
308 only mallocs buffers at initialization.
310 - goes around the event loop after each input package is processed allowing
311 interleaved output processing. The RX flow control api can be used to
312 force compressed input processing to match the rate of compressed output
313 processing (test--echo shows an example of how to do this).
315 - when being "deflate-frame" for compatibility he uses the same default zlib
316 settings as the old "deflate-frame", but instead of exponentially increasing
317 malloc allocations until the whole output will fit, he observes the default
318 input and output chunking buffer sizes of "permessage-deflate", that's
319 1024 in and 1024 out at a time.
321 2) deflate-stream has been disabled for many versions (for over a year) and is
322 now removed. Browsers are now standardizing on "permessage-deflate" / RFC7692
324 3) struct lws_extension is simplified, and lws extensions now have a public
325 api (their callback) for use in user code to compose extensions and options
326 the user code wants. lws_get_internal_exts() is deprecated but kept around
327 as a NOP. The changes allow one extension implementation to go by different
328 names and allows the user client code to control option offers per-ext.
330 The test client and server are updated to use the new way. If you use
331 the old way it should still work, but extensions will be disabled until you
334 Extensions are now responsible for allocating and per-instance private struct
335 at instance construction time and freeing it when the instance is destroyed.
336 Not needing to know the size means the extension's struct can be opaque
343 1) The info struct gained three new members
345 - max_http_header_data: 0 for default (1024) or set the maximum amount of known
346 http header payload that lws can deal with. Payload in unknown http
347 headers is dropped silently. If for some reason you need to send huge
348 cookies or other HTTP-level headers, you can now increase this at context-
351 - max_http_header_pool: 0 for default (16) or set the maximum amount of http
352 headers that can be tracked by lws in this context. For the server, if
353 the header pool is completely in use then accepts on the listen socket
354 are disabled until one becomes free. For the client, if you simultaneously
355 have pending connects for more than this number of client connections,
356 additional connects will fail until some of the pending connections timeout
359 - timeout_secs: 0 for default (currently 20s), or set the library's
360 network activity timeout to the given number of seconds
362 HTTP header processing in lws only exists until just after the first main
363 callback after the HTTP handshake... for ws connections that is ESTABLISHED and
364 for HTTP connections the HTTP callback.
366 So these settings are not related to the maximum number of simultaneous
367 connections, but the number of HTTP handshakes that may be expected or ongoing,
368 or have just completed, at one time. The reason it's useful is it changes the
369 memory allocation for header processing to be one-time at context creation
370 instead of every time there is a new connection, and gives you control over
373 Setting max_http_header_pool to 1 is fine it will just queue incoming
374 connections before the accept as necessary, you can still have as many
375 simultaneous post-header connections as you like. Since the http header
376 processing is completed and the allocation released after ESTABLISHED or the
377 HTTP callback, even with a pool of 1 many connections can be handled rapidly.
379 2) There is a new callback that allows the user code to get acccess to the
380 optional close code + aux data that may have been sent by the peer.
382 LWS_CALLBACK_WS_PEER_INITIATED_CLOSE:
383 The peer has sent an unsolicited Close WS packet. @in and
384 @len are the optional close code (first 2 bytes, network
385 order) and the optional additional information which is not
386 defined in the standard, and may be a string or non-human-
388 If you return 0 lws will echo the close and then close the
389 connection. If you return nonzero lws will just close the
392 As usual not handling it does the right thing, if you're not interested in it
395 The test server has "open and close" testing buttons at the bottom, if you
396 open and close that connection, on close it will send a close code 3000 decimal
397 and the string "Bye!" as the aux data.
399 The test server dumb-increment callback handles this callback reason and prints
401 lwsts[15714]: LWS_CALLBACK_WS_PEER_INITIATED_CLOSE: len 6
402 lwsts[15714]: 0: 0x0B
403 lwsts[15714]: 1: 0xB8
404 lwsts[15714]: 2: 0x42
405 lwsts[15714]: 3: 0x79
406 lwsts[15714]: 4: 0x65
407 lwsts[15714]: 5: 0x21
409 3) There is a new API to allow the user code to control the content of the
410 close frame sent when about to return nonzero from the user callback to
411 indicate the connection should close.
414 * lws_close_reason - Set reason and aux data to send with Close packet
415 * If you are going to return nonzero from the callback
416 * requesting the connection to close, you can optionally
417 * call this to set the reason the peer will be told if
420 * @wsi: The websocket connection to set the close reason on
421 * @status: A valid close status from websocket standard
422 * @buf: NULL or buffer containing up to 124 bytes of auxiliary data
423 * @len: Length of data in @buf to send
425 LWS_VISIBLE LWS_EXTERN void
426 lws_close_reason(struct lws *wsi, enum lws_close_status status,
427 unsigned char *buf, size_t len);
429 An extra button is added to the "open and close" test server page that requests
430 that the test server close the connection from his end.
432 The test server code will do so by
434 lws_close_reason(wsi, LWS_CLOSE_STATUS_GOINGAWAY,
435 (unsigned char *)"seeya", 5);
438 The browser shows the close code and reason he received
440 websocket connection CLOSED, code: 1001, reason: seeya
442 4) There's a new context creation time option flag
444 LWS_SERVER_OPTION_VALIDATE_UTF8
446 if you set it in info->options, then TEXT and CLOSE frames will get checked to
447 confirm that they contain valid UTF-8. If they don't, the connection will get
450 5) ECDH Certs are now supported. Enable the CMake option
452 cmake .. -DLWS_SSL_SERVER_WITH_ECDH_CERT=1
454 **and** the info->options flag
456 LWS_SERVER_OPTION_SSL_ECDH
458 to build in support and select it at runtime.
460 6) There's a new api lws_parse_uri() that simplifies chopping up
461 https://xxx:yyy/zzz uris into parts nicely. The test client now uses this
462 to allow proper uris as well as the old address style.
464 7) SMP support is integrated into LWS without any internal threading. It's
465 very simple to use, libwebsockets-test-server-pthread shows how to do it,
466 use -j <n> argument there to control the number of service threads up to 32.
468 Two new members are added to the info struct
470 unsigned int count_threads;
471 unsigned int fd_limit_per_thread;
473 leave them at the default 0 to get the normal singlethreaded service loop.
475 Set count_threads to n to tell lws you will have n simultaneous service threads
476 operating on the context.
478 There is still a single listen socket on one port, no matter how many
481 When a connection is made, it is accepted by the service thread with the least
482 connections active to perform load balancing.
484 The user code is responsible for spawning n threads running the service loop
485 associated to a specific tsi (Thread Service Index, 0 .. n - 1). See
486 the libwebsockets-test-server-pthread for how to do.
488 If you leave fd_limit_per_thread at 0, then the process limit of fds is shared
489 between the service threads; if you process was allowed 1024 fds overall then
490 each thread is limited to 1024 / n.
492 You can set fd_limit_per_thread to a nonzero number to control this manually, eg
493 the overall supported fd limit is less than the process allowance.
495 You can control the context basic data allocation for multithreading from Cmake
496 using -DLWS_MAX_SMP=, if not given it's set to 32. The serv_buf allocation
497 for the threads (currently 4096) is made at runtime only for active threads.
499 Because lws will limit the requested number of actual threads supported
500 according to LWS_MAX_SMP, there is an api lws_get_count_threads(context) to
501 discover how many threads were actually allowed when the context was created.
503 It's required to implement locking in the user code in the same way that
504 libwebsockets-test-server-pthread does it, for the FD locking callbacks.
506 If LWS_MAX_SMP=1, then there is no code related to pthreads compiled in the
507 library. If more than 1, a small amount of pthread mutex code is built into
512 LWS_VISIBLE struct lws *
513 lws_adopt_socket(struct lws_context *context, lws_sockfd_type accept_fd)
515 allows foreign sockets accepted by non-lws code to be adopted by lws as if they
516 had just been accepted by lws' own listen socket.
518 9) X-Real-IP: header has been added as WSI_TOKEN_HTTP_X_REAL_IP
520 10) Libuv support is added, there are new related user apis
522 typedef void (lws_uv_signal_cb_t)(uv_loop_t *l, uv_signal_t *w, int revents);
524 LWS_VISIBLE LWS_EXTERN int
525 lws_uv_sigint_cfg(struct lws_context *context, int use_uv_sigint,
526 lws_uv_signal_cb_t *cb);
528 LWS_VISIBLE LWS_EXTERN int
529 lws_uv_initloop(struct lws_context *context, uv_loop_t *loop, int tsi);
532 lws_uv_sigint_cb(uv_loop_t *loop, uv_signal_t *watcher, int revents);
542 1) LWS_SEND_BUFFER_POST_PADDING is now 0 and deprecated. You can remove it; if
543 you still use it, obviously it does nothing. Old binary code with nonzero
544 LWS_SEND_BUFFER_POST_PADDING is perfectly compatible, the old code just
545 allocated a buffer bigger than the library is going to use.
547 The example apps no longer use LWS_SEND_BUFFER_POST_PADDING.
549 The only path who made use of it was sending with LWS_WRITE_CLOSE --->
551 2) Because of lws_close_reason() formalizing handling close frames,
552 LWS_WRITE_CLOSE is removed from libwebsockets.h. It was only of use to send
553 close frames...close frame content should be managed using lws_close_reason()
556 3) We check for invalid CLOSE codes and complain about protocol violation in
557 our close code. But it changes little since we were in the middle of closing
560 4) zero-length RX frames and zero length TX frames are now allowed.
562 5) Pings and close used to be limited to 124 bytes, the correct limit is 125
563 so that is now also allowed.
565 6) LWS_PRE is provided as a synonym for LWS_SEND_BUFFER_PRE_PADDING, either is
568 7) There's generic support for RFC7462 style extension options built into the
569 library now. As a consequence, a field "options" is added to lws_extension.
570 It can be NULL if there are no options on the extension. Extension internal
571 info is part of the public abi because extensions may be implemented outside
574 8) WSI_TOKEN_PROXY enum was accidentally defined to collide with another token
575 of value 73. That's now corrected and WSI_TOKEN_PROXY moved to his own place at
578 9) With the addition of libuv support, libev is not the only event loop
579 library in town and his api names must be elaborated with _ev_
581 Callback typedef: lws_signal_cb ---> lws_ev_signal_cb_t
582 lws_sigint_cfg --> lws_ev_sigint_cfg
583 lws_initloop --> lws_ev_initloop
584 lws_sigint_cb --> lws_ev_sigint_cb
586 10) Libev support is made compatible with multithreaded service,
587 lws_ev_initloop (was lws_initloop) gets an extra argument for the
588 thread service index (use 0 if you will just have 1 service thread).
590 LWS_VISIBLE LWS_EXTERN int
591 lws_ev_initloop(struct lws_context *context, ev_loop_t *loop, int tsi);
594 (for earlier changelogs, see the tagged releases)