7 1) OpenSSL version tests not needed on LibreSSL and BoringSSL
9 2) Fix IPV6 build breakage
19 - There are only api additions, the api is compatible with v1.7.x. But
20 there is necessarily an soname bump to 8.
22 - If you are using lws client, you mainly need to be aware the option
23 LWS_SERVER_OPTION_DO_SSL_GLOBAL_INIT is needed at context-creation time
26 - If you are using lws for serving, the above is also true but there are
27 many new features to simplify your code (and life). There is a
30 https://libwebsockets.org/lws-2.0-new-features.html
32 but basically the keywords are vhosts, mounts and plugins. You can now
33 do the web serving part from lws without any user callback code at all.
34 See ./test-server/test-server-v2.0.c for an example, it has no user
35 code for ws either since it uses the protocol plugins... that one C file
36 is all that is needed to do the whole test server function.
38 You now have the option to use a small generic ws-capable webserver
39 "lwsws" and write your ws part as a plugin. That eliminates even
40 cut-and-pasting the test server code and offers more configurable
41 features like control over http cacheability in JSON.
47 These are already in 1.7.x series
49 1) MAJOR (Windows-only) fix assert firing
51 2) MAJOR http:/1.1 connections handled by lws_return_http_status() did not
52 get sent a content-length resulting in the link hanging until the peer closed
53 it. attack.sh updated to add a test for this.
55 3) MINOR An error about hdr struct in _lws_ws_related is corrected, it's not
56 known to affect anything until after it was fixed
58 4) MINOR During the close shutdown wait state introduced at v1.7, if something
59 requests callback on writeable for the socket it will busywait until the
62 5) MAJOR Although the test server has done it for a few versions already, it
63 is now required for the user code to explicitly call
65 if (lws_http_transaction_completed(wsi))
68 when it finishes replying to a transaction in http. Previously the library
69 did it for you, but that disallowed large, long transfers with multiple
70 trips around the event loop (and cgi...).
72 6) MAJOR connections on ah waiting list that closed did not get removed from
75 7) MAJOR since we added the ability to hold an ah across http keepalive
76 transactions where more headers had already arrived, we broke the ability
77 to tell if more headers had arrived. Result was if the browser didn't
78 close the keepalive, we retained ah for the lifetime of the keepalive,
81 8) MAJOR windows-only-POLLHUP was not coming
83 9) Client should not send ext hdr if no exts
88 1) MINOR test-server gained some new switches
90 -C <file> use external SSL cert file
91 -K <file> use external SSL key file
92 -A <file> use external SSL CA cert file
94 -u <uid> set effective uid
95 -g <gid> set effective gid
97 together you can use them like this to have the test-server work with the
98 usual purchased SSL certs from an official CA.
100 --ssl -C your.crt -K your.key -A your.cer -u 99 -g 99
102 2) MINOR the OpenSSL magic to setup ECDH cipher usage is implemented in the
103 library, and the ciphers restricted to use ECDH only.
104 Using this, the lws test server can score an A at SSLLABS test
106 3) MINOR STS (SSL always) header is added to the test server if you use --ssl. With
107 that, we score A+ at SSLLABS test
109 4) MINOR daemonize function (disabled at cmake by default) is updated to work
112 5) MINOR example systemd .service file now provided for test server
113 (not installed by default)
115 6) test server html is updated with tabs and a new live server monitoring
116 feature. Input sanitization added to the js.
118 7) client connections attempted when no ah is free no longer fail, they are
119 just deferred until an ah becomes available.
121 8) The test client pays attention to if you give it an http:/ or https://
122 protocol string to its argument in URL format. If so, it stays in http[s]
123 client mode and doesn't upgrade to ws[s], allowing you to do generic http client
124 operations. Receiving transfer-encoding: chunked is supported.
126 9) If you enable -DLWS_WITH_HTTP_PROXY=1 at cmake, the test server has a
127 new URI path http://localhost:7681/proxytest If you visit here, a client
128 connection to http://example.com:80 is spawned, and the results piped on
129 to your original connection.
131 10) Also with LWS_WITH_HTTP_PROXY enabled at cmake, lws wants to link to an
132 additional library, "libhubbub". This allows lws to do html rewriting on the
133 fly, adjusting proxied urls in a lightweight and fast way.
135 11) There's a new context creation flag LWS_SERVER_OPTION_DO_SSL_GLOBAL_INIT,
136 this is included automatically if you give any other SSL-related option flag.
137 If you give no SSL-related option flag, nor this one directly, then even
138 though SSL support may be compiled in, it is never initialized nor used for the
139 whole lifetime of the lws context.
141 Conversely in order to prepare the context to use SSL, even though, eg, you
142 are not listening on SSL but will use SSL client connections later, you must
143 give this flag explicitly to make sure SSL is initialized.
149 1) MINOR APIBREAK There's a new member in struct lws_context_creation_info, ecdh_curve,
150 which lets you set the name of the ECDH curve OpenSSL should use. By
151 default (if you leave ecdh_curve NULL) it will use "prime256v1"
153 2) MINOR NEWAPI It was already possible to adopt a foreign socket that had not
154 been read from using lws_adopt_socket() since v1.7. Now you can adopt a
155 partially-used socket if you don't need SSL, by passing it what you read
156 so it can drain that before reading from the socket.
158 LWS_VISIBLE LWS_EXTERN struct lws *
159 lws_adopt_socket_readbuf(struct lws_context *context, lws_sockfd_type accept_fd,
160 const char *readbuf, size_t len);
162 3) MINOR NEWAPI CGI type "network io" subprocess execution is now possible from
165 LWS_VISIBLE LWS_EXTERN int
166 lws_cgi(struct lws *wsi, char * const *exec_array, int script_uri_path_len,
169 LWS_VISIBLE LWS_EXTERN int
170 lws_cgi_kill(struct lws *wsi);
172 To use it, you must first set the cmake option
174 $ cmake .. -DLWS_WITH_CGI=1
176 See test-server-http.c and test server path
178 http://localhost:7681/cgitest
180 stdin gets http body, you can test it with wget
182 $ echo hello > hello.txt
183 $ wget http://localhost:7681/cgitest --post-file=hello.txt -O- --quiet
187 The test script returns text/html table showing /proc/meminfo. But the cgi
188 support is complete enough to run cgit cgi.
190 4) There is a helper api for forming logging timestamps
193 lwsl_timestamp(int level, char *p, int len)
195 this generates this kind of timestamp for use as logging preamble
197 lwsts[13116]: [2016/01/25 14:52:52:8386] NOTICE: Initial logging level 7
199 5) struct lws_client_connect_info has a new member
203 If it's NULL, then everything happens as before, lws_client_connect_via_info()
204 makes a ws or wss connection to the address given.
206 If you set method to a valid http method like "GET", though, then this method
207 is used and the connection remains in http[s], it's not upgraded to ws[s].
209 So with this, you can perform http[s] client operations as well as ws[s] ones.
211 There are 4 new related callbacks
213 LWS_CALLBACK_ESTABLISHED_CLIENT_HTTP = 44,
214 LWS_CALLBACK_CLOSED_CLIENT_HTTP = 45,
215 LWS_CALLBACK_RECEIVE_CLIENT_HTTP = 46,
216 LWS_CALLBACK_COMPLETED_CLIENT_HTTP = 47,
218 6) struct lws_client_connect_info has a new member
220 const char *parent_wsi
222 if non-NULL, the client wsi is set to be a child of parent_wsi. This ensures
223 if parent_wsi closes, then the client child is closed just before.
225 7) If you're using SSL, there's a new context creation-time option flag
226 LWS_SERVER_OPTION_REDIRECT_HTTP_TO_HTTPS. If you give this, non-ssl
227 connections to the server listen port are accepted and receive a 301
228 redirect to / on the same host and port using https://
230 8) User code may set per-connection extension options now, using a new api
231 "lws_set_extension_option()".
233 This should be called from the ESTABLISHED callback like this
235 lws_set_extension_option(wsi, "permessage-deflate",
236 "rx_buf_size", "12"); /* 1 << 12 */
238 If the extension is not active (missing or not negotiated for the
239 connection, or extensions are disabled on the library) the call is
240 just returns -1. Otherwise the connection's extension has its
241 named option changed.
243 The extension may decide to alter or disallow the change, in the
244 example above permessage-deflate restricts the size of his rx
245 output buffer also considering the protocol's rx_buf_size member.
248 New application lwsws
249 ---------------------
251 A libwebsockets-based general webserver is built by default now, lwsws.
253 It's configured by JSON, by default in
257 which contains global lws context settings like this
270 which contains zero or more files describing vhosts, like this
274 { "name": "warmcat.com",
276 "host-ssl-key": "/etc/pki/tls/private/warmcat.com.key",
277 "host-ssl-cert": "/etc/pki/tls/certs/warmcat.com.crt",
278 "host-ssl-ca": "/etc/pki/tls/certs/warmcat.com.cer",
281 { "home": "file:///var/www/warmcat.com" },
282 { "default": "index.html" }
298 1) There is now a "permessage-deflate" / RFC7692 implementation. It's very
299 similar to "deflate-frame" we have offered for a long while; deflate-frame is
300 now provided as an alias of permessage-deflate.
302 The main differences are that the new permessage-deflate implementation:
304 - properly performs streaming respecting input and output buffer limits. The
305 old deflate-frame implementation could only work on complete deflate input
306 and produce complete inflate output for each frame. The new implementation
307 only mallocs buffers at initialization.
309 - goes around the event loop after each input package is processed allowing
310 interleaved output processing. The RX flow control api can be used to
311 force compressed input processing to match the rate of compressed output
312 processing (test--echo shows an example of how to do this).
314 - when being "deflate-frame" for compatibility he uses the same default zlib
315 settings as the old "deflate-frame", but instead of exponentially increasing
316 malloc allocations until the whole output will fit, he observes the default
317 input and output chunking buffer sizes of "permessage-deflate", that's
318 1024 in and 1024 out at a time.
320 2) deflate-stream has been disabled for many versions (for over a year) and is
321 now removed. Browsers are now standardizing on "permessage-deflate" / RFC7692
323 3) struct lws_extension is simplified, and lws extensions now have a public
324 api (their callback) for use in user code to compose extensions and options
325 the user code wants. lws_get_internal_exts() is deprecated but kept around
326 as a NOP. The changes allow one extension implementation to go by different
327 names and allows the user client code to control option offers per-ext.
329 The test client and server are updated to use the new way. If you use
330 the old way it should still work, but extensions will be disabled until you
333 Extensions are now responsible for allocating and per-instance private struct
334 at instance construction time and freeing it when the instance is destroyed.
335 Not needing to know the size means the extension's struct can be opaque
342 1) The info struct gained three new members
344 - max_http_header_data: 0 for default (1024) or set the maximum amount of known
345 http header payload that lws can deal with. Payload in unknown http
346 headers is dropped silently. If for some reason you need to send huge
347 cookies or other HTTP-level headers, you can now increase this at context-
350 - max_http_header_pool: 0 for default (16) or set the maximum amount of http
351 headers that can be tracked by lws in this context. For the server, if
352 the header pool is completely in use then accepts on the listen socket
353 are disabled until one becomes free. For the client, if you simultaneously
354 have pending connects for more than this number of client connections,
355 additional connects will fail until some of the pending connections timeout
358 - timeout_secs: 0 for default (currently 20s), or set the library's
359 network activity timeout to the given number of seconds
361 HTTP header processing in lws only exists until just after the first main
362 callback after the HTTP handshake... for ws connections that is ESTABLISHED and
363 for HTTP connections the HTTP callback.
365 So these settings are not related to the maximum number of simultaneous
366 connections, but the number of HTTP handshakes that may be expected or ongoing,
367 or have just completed, at one time. The reason it's useful is it changes the
368 memory allocation for header processing to be one-time at context creation
369 instead of every time there is a new connection, and gives you control over
372 Setting max_http_header_pool to 1 is fine it will just queue incoming
373 connections before the accept as necessary, you can still have as many
374 simultaneous post-header connections as you like. Since the http header
375 processing is completed and the allocation released after ESTABLISHED or the
376 HTTP callback, even with a pool of 1 many connections can be handled rapidly.
378 2) There is a new callback that allows the user code to get acccess to the
379 optional close code + aux data that may have been sent by the peer.
381 LWS_CALLBACK_WS_PEER_INITIATED_CLOSE:
382 The peer has sent an unsolicited Close WS packet. @in and
383 @len are the optional close code (first 2 bytes, network
384 order) and the optional additional information which is not
385 defined in the standard, and may be a string or non-human-
387 If you return 0 lws will echo the close and then close the
388 connection. If you return nonzero lws will just close the
391 As usual not handling it does the right thing, if you're not interested in it
394 The test server has "open and close" testing buttons at the bottom, if you
395 open and close that connection, on close it will send a close code 3000 decimal
396 and the string "Bye!" as the aux data.
398 The test server dumb-increment callback handles this callback reason and prints
400 lwsts[15714]: LWS_CALLBACK_WS_PEER_INITIATED_CLOSE: len 6
401 lwsts[15714]: 0: 0x0B
402 lwsts[15714]: 1: 0xB8
403 lwsts[15714]: 2: 0x42
404 lwsts[15714]: 3: 0x79
405 lwsts[15714]: 4: 0x65
406 lwsts[15714]: 5: 0x21
408 3) There is a new API to allow the user code to control the content of the
409 close frame sent when about to return nonzero from the user callback to
410 indicate the connection should close.
413 * lws_close_reason - Set reason and aux data to send with Close packet
414 * If you are going to return nonzero from the callback
415 * requesting the connection to close, you can optionally
416 * call this to set the reason the peer will be told if
419 * @wsi: The websocket connection to set the close reason on
420 * @status: A valid close status from websocket standard
421 * @buf: NULL or buffer containing up to 124 bytes of auxiliary data
422 * @len: Length of data in @buf to send
424 LWS_VISIBLE LWS_EXTERN void
425 lws_close_reason(struct lws *wsi, enum lws_close_status status,
426 unsigned char *buf, size_t len);
428 An extra button is added to the "open and close" test server page that requests
429 that the test server close the connection from his end.
431 The test server code will do so by
433 lws_close_reason(wsi, LWS_CLOSE_STATUS_GOINGAWAY,
434 (unsigned char *)"seeya", 5);
437 The browser shows the close code and reason he received
439 websocket connection CLOSED, code: 1001, reason: seeya
441 4) There's a new context creation time option flag
443 LWS_SERVER_OPTION_VALIDATE_UTF8
445 if you set it in info->options, then TEXT and CLOSE frames will get checked to
446 confirm that they contain valid UTF-8. If they don't, the connection will get
449 5) ECDH Certs are now supported. Enable the CMake option
451 cmake .. -DLWS_SSL_SERVER_WITH_ECDH_CERT=1
453 **and** the info->options flag
455 LWS_SERVER_OPTION_SSL_ECDH
457 to build in support and select it at runtime.
459 6) There's a new api lws_parse_uri() that simplifies chopping up
460 https://xxx:yyy/zzz uris into parts nicely. The test client now uses this
461 to allow proper uris as well as the old address style.
463 7) SMP support is integrated into LWS without any internal threading. It's
464 very simple to use, libwebsockets-test-server-pthread shows how to do it,
465 use -j <n> argument there to control the number of service threads up to 32.
467 Two new members are added to the info struct
469 unsigned int count_threads;
470 unsigned int fd_limit_per_thread;
472 leave them at the default 0 to get the normal singlethreaded service loop.
474 Set count_threads to n to tell lws you will have n simultaneous service threads
475 operating on the context.
477 There is still a single listen socket on one port, no matter how many
480 When a connection is made, it is accepted by the service thread with the least
481 connections active to perform load balancing.
483 The user code is responsible for spawning n threads running the service loop
484 associated to a specific tsi (Thread Service Index, 0 .. n - 1). See
485 the libwebsockets-test-server-pthread for how to do.
487 If you leave fd_limit_per_thread at 0, then the process limit of fds is shared
488 between the service threads; if you process was allowed 1024 fds overall then
489 each thread is limited to 1024 / n.
491 You can set fd_limit_per_thread to a nonzero number to control this manually, eg
492 the overall supported fd limit is less than the process allowance.
494 You can control the context basic data allocation for multithreading from Cmake
495 using -DLWS_MAX_SMP=, if not given it's set to 32. The serv_buf allocation
496 for the threads (currently 4096) is made at runtime only for active threads.
498 Because lws will limit the requested number of actual threads supported
499 according to LWS_MAX_SMP, there is an api lws_get_count_threads(context) to
500 discover how many threads were actually allowed when the context was created.
502 It's required to implement locking in the user code in the same way that
503 libwebsockets-test-server-pthread does it, for the FD locking callbacks.
505 If LWS_MAX_SMP=1, then there is no code related to pthreads compiled in the
506 library. If more than 1, a small amount of pthread mutex code is built into
511 LWS_VISIBLE struct lws *
512 lws_adopt_socket(struct lws_context *context, lws_sockfd_type accept_fd)
514 allows foreign sockets accepted by non-lws code to be adopted by lws as if they
515 had just been accepted by lws' own listen socket.
517 9) X-Real-IP: header has been added as WSI_TOKEN_HTTP_X_REAL_IP
519 10) Libuv support is added, there are new related user apis
521 typedef void (lws_uv_signal_cb_t)(uv_loop_t *l, uv_signal_t *w, int revents);
523 LWS_VISIBLE LWS_EXTERN int
524 lws_uv_sigint_cfg(struct lws_context *context, int use_uv_sigint,
525 lws_uv_signal_cb_t *cb);
527 LWS_VISIBLE LWS_EXTERN int
528 lws_uv_initloop(struct lws_context *context, uv_loop_t *loop, int tsi);
531 lws_uv_sigint_cb(uv_loop_t *loop, uv_signal_t *watcher, int revents);
541 1) LWS_SEND_BUFFER_POST_PADDING is now 0 and deprecated. You can remove it; if
542 you still use it, obviously it does nothing. Old binary code with nonzero
543 LWS_SEND_BUFFER_POST_PADDING is perfectly compatible, the old code just
544 allocated a buffer bigger than the library is going to use.
546 The example apps no longer use LWS_SEND_BUFFER_POST_PADDING.
548 The only path who made use of it was sending with LWS_WRITE_CLOSE --->
550 2) Because of lws_close_reason() formalizing handling close frames,
551 LWS_WRITE_CLOSE is removed from libwebsockets.h. It was only of use to send
552 close frames...close frame content should be managed using lws_close_reason()
555 3) We check for invalid CLOSE codes and complain about protocol violation in
556 our close code. But it changes little since we were in the middle of closing
559 4) zero-length RX frames and zero length TX frames are now allowed.
561 5) Pings and close used to be limited to 124 bytes, the correct limit is 125
562 so that is now also allowed.
564 6) LWS_PRE is provided as a synonym for LWS_SEND_BUFFER_PRE_PADDING, either is
567 7) There's generic support for RFC7462 style extension options built into the
568 library now. As a consequence, a field "options" is added to lws_extension.
569 It can be NULL if there are no options on the extension. Extension internal
570 info is part of the public abi because extensions may be implemented outside
573 8) WSI_TOKEN_PROXY enum was accidentally defined to collide with another token
574 of value 73. That's now corrected and WSI_TOKEN_PROXY moved to his own place at
577 9) With the addition of libuv support, libev is not the only event loop
578 library in town and his api names must be elaborated with _ev_
580 Callback typedef: lws_signal_cb ---> lws_ev_signal_cb_t
581 lws_sigint_cfg --> lws_ev_sigint_cfg
582 lws_initloop --> lws_ev_initloop
583 lws_sigint_cb --> lws_ev_sigint_cb
585 10) Libev support is made compatible with multithreaded service,
586 lws_ev_initloop (was lws_initloop) gets an extra argument for the
587 thread service index (use 0 if you will just have 1 service thread).
589 LWS_VISIBLE LWS_EXTERN int
590 lws_ev_initloop(struct lws_context *context, ev_loop_t *loop, int tsi);
593 (for earlier changelogs, see the tagged releases)