David Galeano [Wed, 9 Jan 2013 10:03:28 +0000 (18:03 +0800)]
Fixed deflate-stream extension.
When the output buffer was exhausted the input buffer was discarded
without checking if zlib had actually consumed all the input,
now we copy the remaining input data for the next call.
David Galeano [Wed, 9 Jan 2013 10:01:23 +0000 (18:01 +0800)]
Added private macro AWAITING_TIMEOUT instead of harcoded value 5.
David Galeano [Wed, 9 Jan 2013 08:35:18 +0000 (16:35 +0800)]
Fixed spacing.
David Galeano [Wed, 9 Jan 2013 08:25:54 +0000 (16:25 +0800)]
Added context creation parameter for CA certificates file.
David Galeano [Wed, 9 Jan 2013 08:17:04 +0000 (16:17 +0800)]
Return NULL if the handshake failed to complete,
libwebsocket_service_fd closes and frees the websocket in that case.
David Galeano [Wed, 9 Jan 2013 07:46:11 +0000 (15:46 +0800)]
Ignoring linux build files
David Galeano [Wed, 9 Jan 2013 07:39:50 +0000 (15:39 +0800)]
Use feature check rather than browser check.
David Galeano [Wed, 9 Jan 2013 07:31:46 +0000 (15:31 +0800)]
Changed client handshake to use "Origin" instead of "Sec-WebSocket-Origin" as defined by RFC 6455 when using version 13 of the protocol.
David Galeano [Wed, 9 Jan 2013 07:29:00 +0000 (15:29 +0800)]
Fixed compiler warnings on Windows.
David Galeano [Wed, 9 Jan 2013 07:26:08 +0000 (15:26 +0800)]
Added new status codes from RFC 6455.
David Galeano [Wed, 9 Jan 2013 07:25:05 +0000 (15:25 +0800)]
Fixed compiler warning on Windows.
Artem Baguinski [Wed, 9 Jan 2013 07:16:52 +0000 (15:16 +0800)]
required version of autoconf can be lower
David Galeano [Wed, 9 Jan 2013 07:14:31 +0000 (15:14 +0800)]
Static variable is now const.
Andy Green [Mon, 22 Oct 2012 11:29:57 +0000 (12:29 +0100)]
add context construction option to skip server hostname lookup
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Fri, 19 Oct 2012 10:36:28 +0000 (18:36 +0800)]
add missing docs for new context user pointer
Signed-off-by: Andy Green <andy.green@linaro.org>
Alon Levy [Fri, 19 Oct 2012 09:21:57 +0000 (11:21 +0200)]
libwebsocket_service_fd: EAGAIN is harmless, treat like EINTR
Only tested on linux.
Signed-off-by: Alon Levy <alevy@redhat.com>
Alon Levy [Fri, 19 Oct 2012 09:21:56 +0000 (11:21 +0200)]
libwebsocket_context: add userspace pointer for use before wsi creation
Signed-off-by: Alon Levy <alevy@redhat.com>
Alon Levy [Fri, 19 Oct 2012 09:21:55 +0000 (11:21 +0200)]
lib/Makefile.am: whitespace fix
Signed-off-by: Alon Levy <alevy@redhat.com>
Alon Levy [Fri, 19 Oct 2012 09:21:54 +0000 (11:21 +0200)]
add pkg-config file libwebsockets.pc
Signed-off-by: Alon Levy <alevy@redhat.com>
Anders Brander [Wed, 17 Oct 2012 12:23:28 +0000 (14:23 +0200)]
Added test.html favicon.ico to EXTRA_DIST.
Anders Brander [Wed, 17 Oct 2012 11:48:59 +0000 (13:48 +0200)]
Add missing .h files to sources.
Anders Brander [Wed, 17 Oct 2012 11:47:47 +0000 (13:47 +0200)]
Add kernel doc to extra_dist.
Andy Green [Wed, 17 Oct 2012 12:10:44 +0000 (20:10 +0800)]
always taking an interest in ppid wont hurt
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Sun, 7 Oct 2012 12:44:02 +0000 (20:44 +0800)]
remove depcomp
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Sun, 7 Oct 2012 12:40:35 +0000 (20:40 +0800)]
stop being so fragile on socket lifecycle
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Sat, 6 Oct 2012 07:17:01 +0000 (15:17 +0800)]
use autogen.sh
Signed-off-by: Andy Green <andy.green@linaro.org>
Rusty Lynch [Wed, 5 Sep 2012 23:15:16 +0000 (16:15 -0700)]
Updating changelog
Rusty Lynch [Thu, 23 Aug 2012 02:28:37 +0000 (19:28 -0700)]
Adding Tizen packaging files
Andy Green [Fri, 20 Jul 2012 05:00:12 +0000 (13:00 +0800)]
add not about autoreconf and libtoolize
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Fri, 20 Jul 2012 04:58:38 +0000 (12:58 +0800)]
restore accept error as closure signal
Signed-off-by: Andy Green <andy.green@linaro.org>
Artem Baguinski [Wed, 14 Dec 2011 21:14:03 +0000 (22:14 +0100)]
check for prctl, poll parent PID if not present
this allows forking code to be used on non-linux systems
Paulo Roberto Urio [Mon, 4 Jun 2012 02:52:19 +0000 (10:52 +0800)]
Fixing uninitialised memory
These were found with valgrind tool.
Signed-off-by: Paulo Roberto Urio <paulourio@gmail.com>
Paulo Roberto Urio [Mon, 4 Jun 2012 01:04:33 +0000 (09:04 +0800)]
some toolchains need stddef
Signed-off-by: Paulo Roberto Urio <paulourio@gmail.com>
Andy Green [Mon, 4 Jun 2012 00:53:26 +0000 (08:53 +0800)]
add missing sa_data init to canonical hostname code
Shay noticed we're no longer initializing the initial lookup of
server canonical hostname correctly
Reported-by: Shay Zuker <shay@boxee.tv>
Signed-off-by: Andy Green <andy.green@linaro.org>
Paulo Roberto Urio [Mon, 4 Jun 2012 00:40:28 +0000 (08:40 +0800)]
Fixed segfault in libwebsocket_context_destroy.
When creating a context with NULL extensions list,
a segmentation fault was yelled when trying to
destroy the context. This checks if the
extension list is NULL before go through the list.
Signed-off-by: Paulo Roberto Urio <paulourio@gmail.com>
Tobias Maier [Wed, 30 May 2012 04:46:42 +0000 (12:46 +0800)]
sockets were left open if connection could not be
established resulting in a mass of unusable open file
descriptors
Signed-off-by: Tobias Maier <tobias.maier@netplace.com>
yuval tal [Mon, 21 May 2012 02:16:06 +0000 (10:16 +0800)]
link lz explicitly for mips
fix link error since libz is not linked in
Signed-off-by: yuval tal <yuvalt@boxee.tv>
Andrew Chambers [Sun, 20 May 2012 00:17:09 +0000 (08:17 +0800)]
null exception with null extensions list
I was under the impression extensions could be null, so heres a patch to fix this error in libwebsockets. Cheers!
Signed-off-by: Andrew Chambers <andrewchamberss@gmail.com>
--
Andy Green [Sun, 20 May 2012 00:17:02 +0000 (08:17 +0800)]
refresh autotools stuff
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Thu, 3 May 2012 04:48:14 +0000 (12:48 +0800)]
update test html now mozilla uses normal ws apis
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Thu, 3 May 2012 04:32:38 +0000 (12:32 +0800)]
start migration to getnameinfo
Signed-off-by: Andy Green <andy.green@linaro.org>
Ken Atherton [Thu, 3 May 2012 03:45:04 +0000 (11:45 +0800)]
openssl allow connect multiple packets for SSL_connect
Ken found over the internet with real delays, SSL_connect can
fail to work. This patch adapts his workaround to stay in the
connect state until we either run out of time for the connect
or succeed.
Signed-off-by: Andy Green <andy.green@linaro.org>
Signed-off-by: Ken Atherton <katherton@echofirst.com>
Andy Green [Thu, 3 May 2012 03:17:08 +0000 (11:17 +0800)]
cleanup extra wrapping needed on braced define
Signed-off-by: Andy Green <andy.green@linaro.org>
Peter Hillier [Tue, 1 May 2012 22:09:45 +0000 (06:09 +0800)]
android toolchain SHA1 endian fix
Android toolchain needs an extra include if it's not to confuse
SHA-1 code probably with incorrect endian-ness from missing /bits/
Signed-off-by: Peter Hillier <peterhillier@yahoo.com>
Signed-off-by: Andy Green <andy@warmcat.com>
David Brooks [Tue, 1 May 2012 22:02:45 +0000 (06:02 +0800)]
ssize_t needs help on windows
ssize_t is needed, but absent in Windows.
This patch typedefs it to an int in that case as recommended by Tobias.
Signed-off-by: Andy Green <andy@warmcat.com>
Signed-off-by: David Brooks <dave@bcs.co.nz>
Signed-off-by: Tobias Maier <tobias.maier@netplace.com>
Reported-by: Rich Gossweiler <rich.gossweiler@gmail.com>
David Brooks [Fri, 20 Apr 2012 04:19:01 +0000 (12:19 +0800)]
introduce libwebsocket_client_connect_extended
Signed-off-by: David Brooks <dave@bcs.co.nz>
Signed-off-by: Andy Green <andy@warmcat.com>
--
David Brooks [Fri, 20 Apr 2012 04:18:47 +0000 (12:18 +0800)]
introduce LWS_CALLBACK_CLIENT_CONNECTION_ERROR
Signed-off-by: David Brooks <dave@bcs.co.nz>
Signed-off-by: Andy Green <andy@warmcat.com>
--
David Brooks [Fri, 20 Apr 2012 04:16:52 +0000 (12:16 +0800)]
set connection callback before connection completed to allow early messages
Signed-off-by: David Brooks <dave@bcs.co.nz>
Signed-off-by: Andy Green <andy@warmcat.com>
--
lib/client-handshake.c | 19 +++++++++++++++++++
1 files changed, 19 insertions(+), 0 deletions(-)
David Brooks [Fri, 20 Apr 2012 04:13:50 +0000 (12:13 +0800)]
improve callback prototypes to use single definition
Not sure all compilers can hack this, but give it a try
Signed-off-by: David Brooks <dave@bcs.co.nz>
Signed-off-by: Andy Green <andy@warmcat.com>
--
David Brooks [Fri, 20 Apr 2012 04:13:37 +0000 (12:13 +0800)]
reduce debug spew
Signed-off-by: David Brooks <dave@bcs.co.nz>
Signed-off-by: Andy Green <andy@warmcat.com>
--
Pasi Mankinen [Fri, 13 Apr 2012 23:28:19 +0000 (07:28 +0800)]
workaround osx compiler bug
Signed-off-by: Pasi Mankinen <pasi.t.mankinen@student.jyu.fi>
Andy Green [Thu, 12 Apr 2012 05:32:47 +0000 (13:32 +0800)]
document mingw build
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Thu, 12 Apr 2012 05:26:49 +0000 (13:26 +0800)]
mingw add win32helper to lib and fix extpoll
Based on work from Radu Sorici <soriciradu@gmail.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Thu, 12 Apr 2012 03:15:23 +0000 (11:15 +0800)]
doc add explanation missing from per context provate data
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Thu, 12 Apr 2012 03:11:10 +0000 (11:11 +0800)]
mingw add zlib paths
Based on work from Radu Sorici <soriciradu@gmail.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Thu, 12 Apr 2012 03:08:55 +0000 (11:08 +0800)]
remove duplicated netdb include from client handshake
Based on work from Radu Sorici <soriciradu@gmail.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Thu, 12 Apr 2012 03:06:05 +0000 (11:06 +0800)]
change DATADIR to INSTALL_DATADIR solve mingw conflict
Based on work from Radu Sorici <soriciradu@gmail.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Thu, 12 Apr 2012 03:02:06 +0000 (11:02 +0800)]
mingw gettimeofday avoid missing include
Based on work from Radu Sorici <soriciradu@gmail.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Thu, 12 Apr 2012 02:56:20 +0000 (10:56 +0800)]
mingw add in missing poll defs
Based on work from Radu Sorici <soriciradu@gmail.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Thu, 12 Apr 2012 02:54:22 +0000 (10:54 +0800)]
mingw hide some missing includes
Based on work from Radu Sorici <soriciradu@gmail.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Thu, 12 Apr 2012 02:11:29 +0000 (10:11 +0800)]
autoconf introduce enable mingw and adapt compiler options
Based on work from Radu Sorici <soriciradu@gmail.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Mon, 9 Apr 2012 07:09:01 +0000 (15:09 +0800)]
style cleaning
Signed-off-by: Andy Green <andy.green@linaro.org>
Andy Green [Mon, 9 Apr 2012 07:23:47 +0000 (15:23 +0800)]
fix onopen browser context patch
Signed-off-by: Andy Green <andy.green@linaro.org>
Yonathan Yusim [Sun, 8 Apr 2012 22:42:39 +0000 (06:42 +0800)]
allow message processing from onopen browser context
Found an issue in libwebsockets - if a server sends a message within its "onopen" context then the message is getting swallowed and not passed to calling client (means LWS_CALLBACK_CLIENT_RECEIVE is never getting called).
I've fixed the bug - pls find attached patch.
Signed-off-by: Yonathan Yusim <yonathan@boxee.tv>
Signed-off-by: Andy Green <andy@warmcat.com>
Tobias Maier [Thu, 5 Apr 2012 09:57:12 +0000 (11:57 +0200)]
use ssize_t instead of size_t for return values of read/recv/SSL_read
Signed-off-by: Tobias Maier <tobias.maier@netplace.com>
Nick Dowell [Thu, 5 Apr 2012 02:31:48 +0000 (10:31 +0800)]
make duplicate headers append instead of get ignored
The code to prevent memory leaks caused by reallocating a wsi->utf8_token
was also causing wsi->parser_state to not be updated, preventing the
handshake from completing.
This patch changes the logic to append to previous allocation which is
correct behaviour actually.
Signed-off-by: Nick Dowell <nick@nickdowell.com>
Signed-off-by: Andy Green <andy@warmcat.com>
Nick Dowell [Thu, 5 Apr 2012 02:29:39 +0000 (10:29 +0800)]
close connection if socket recv() fails.
This prevents a problem observed on Windows whereby a client that
disconnects before fully establishing the WebSocket would cause the
server to use 100% CPU since recv() would continually return -1 for that
socket.
Signed-off-by: Nick Dowell <nick@nickdowell.com>
Nick Dowell [Thu, 5 Apr 2012 02:29:29 +0000 (10:29 +0800)]
fix heap corruption if HTTP callback calls libwebsocket_close_and_free_session
Signed-off-by: Nick Dowell <nick@nickdowell.com>
Andy Green [Tue, 3 Apr 2012 15:02:20 +0000 (17:02 +0200)]
map wsorigin on to origin at an early point and stop dupe header mem leaks
Carlo wrote
''I'm interested to use the libwebsockets server (C based) with a Java
websocket client. I was able to implement a Java websocket client based
on JWebSocket library that works fine with the libwebsockets server, but
I was obliged to patch the libwebsockets server in order to reach my
goal. In particular I patched the handshake_00() function because
JWebSocket library uses hixie 76 version. The problem was that the
JWebSocket library sends only WSI_TOKEN_SWORIGIN token instead
WSI_TOKEN_ORIGIN token. For this reason the handshake_00 failed because
the token length of WSI_TOKEN_ORIGIN was 0. I have written a patch in
order to control if at least one of the two tokens (WSI_TOKEN_SWORIGIN
and WSI_TOKEN_ORIGIN) is defined. I have attached into the e-mail the
patch. Probably it can be written better, but in any case I think that
this problem should be fixed for client/server compatibility.''
I found a simpler approach which is convert WSORIGIN processing to
go into ORIGIN token at the earliest point, then everything else can
stay the same.
I also noticed we would leak on duplicated headers and fixed that.
Reported-by: Carlo PARATA <carlo.parata@st.com>
Signed-off-by: Andy Green <andy@warmcat.com>
Nick Dowell [Tue, 3 Apr 2012 16:22:19 +0000 (17:22 +0100)]
handle unsupported protocol correctly
Hi Andy,
First off, thanks for libwebsockets :-)
I've encountered a crash when a client connects to a libwebsockets server
but speicifies an unsupported protocol (Sec-WebSocket-Protocol).
handshake.c should probably be checking that wsi->protocol->name is not
null before doing a strcmp with it...
Attached is a patch for your consideration.
Cheers!
Nick
Signed-off-by: Nick Dowell <nick@nickdowell.com>
Andy Green [Thu, 29 Mar 2012 01:27:13 +0000 (09:27 +0800)]
fix issuing null extension header if none supported
Josh realized that with new Chrome, because we don't support the type of
compression extension yet we returned a null extension header.
This patch fixes that by deferring issuing the extension header until we
find we have something to say.
tested OK on google-chrome 19.0.1081.2-129295
Reported-by: Josh Roberson <josh@asteriasgi.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
Artem Egorkine [Wed, 23 Nov 2011 08:46:24 +0000 (10:46 +0200)]
Don't care for the human-readable part of "HTTP/1.1 101" reply
While the protocol specification gives example of the
human-readable message associated with a 101-code reply,
nowhere does it actually enforce it. Ws4py gevent middleware
is known to have a message other than "Switching Protocols".
Dan Zhang [Wed, 9 Nov 2011 07:55:16 +0000 (15:55 +0800)]
cleanup fixes tested on win32 compiler
Signed-off-by: Dan Zhang <emaildanzhang@gmail.com>
Andy Green [Tue, 8 Nov 2011 23:28:45 +0000 (07:28 +0800)]
win32 poll emulation clean and address PEEK issue
Dan Zhang needed to defeat the recv() with PEEK in order to get working
poll() emulation on 64-bit windows.
I cleaned up the style in that file and made a version of his workaround
(which was force recv result to 0). I can't test the result so it may
need more work.
Reported-by: Dan Zhang <emaildanzhang@gmail.com>
Signed-off-by: Andy Green <andy@warmcat.com>
Andy Green [Mon, 7 Nov 2011 11:53:23 +0000 (19:53 +0800)]
debug move most spew into debug macro
This quietens the spew so all typical debug info now is coming from
the user code (mirror protocol in the sample server / client case).
Signed-off-by: Andy Green <andy@warmcat.com>
Alex Bligh [Mon, 7 Nov 2011 09:19:25 +0000 (17:19 +0800)]
Add libwebsockets_ensure_user_space
This patch unifies the code for per-connection user allocation, and allows
it to be allocated earlier, duiring HTTP service time. It's also OK to
attempt allocation multiple times, it will just return any existing
allocation.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
David Galeano [Tue, 4 Oct 2011 12:59:24 +0000 (20:59 +0800)]
add missing win32 zlib pieces
Signed-off-by: David Galeano <davidgaleano@turbulenz.biz>
Oleg Golosovskiy [Tue, 4 Oct 2011 12:20:14 +0000 (20:20 +0800)]
http charsets have case insensitive tokens
Andy,
According http://www.ietf.org/rfc/rfc2616.txt HTTP character sets are identified by case-insensitive tokens.
Please replase
if (strcmp(lws_tokens[n].token, wsi->name_buffer))
continue;
to
if (stricmp(lws_tokens[n].token, wsi->name_buffer))
continue;
Oleg
Also introduce strcasecmp definition for win32
Signed-off-by: Oleg Golosovskiy <ogolosovskiy@unison.com>
David Galeano [Tue, 4 Oct 2011 11:55:18 +0000 (19:55 +0800)]
Added support for Windows XP.
David Galeano [Mon, 3 Oct 2011 11:30:27 +0000 (19:30 +0800)]
Added BYTE_ORDER definition for win32 builds.
Andy Green [Mon, 3 Oct 2011 11:30:22 +0000 (19:30 +0800)]
doc LWS_EXTERN
Signed-off-by: Andy Green <andy@warmcat.com>
David Galeno [Mon, 26 Sep 2011 16:33:45 +0000 (17:33 +0100)]
zlib update
Signed-off-by: David Galeno <davidgaleano@turbulenz.biz>
David Galeano [Mon, 26 Sep 2011 16:30:34 +0000 (17:30 +0100)]
Added "Release DLL" configuration.
David Galeano [Mon, 26 Sep 2011 11:17:20 +0000 (12:17 +0100)]
Added LWS_EXTERN to allow building a DLL.
Signed-off-by: David Galeano <davidgaleano@turbulenz.biz>
David Galeano [Mon, 26 Sep 2011 11:09:54 +0000 (12:09 +0100)]
Fixed compiler warnings.
Signed-off-by: David Galeano <davidgaleano@turbulenz.biz>
David Galeano [Mon, 26 Sep 2011 11:09:20 +0000 (12:09 +0100)]
Ignoring Release build and opensdf files.
Signed-off-by: David Galeano <davidgaleano@turbulenz.biz>
Andy Green [Sun, 25 Sep 2011 09:46:31 +0000 (10:46 +0100)]
fix failure to limit 7bit lengths
Reported-by: Tran Le Hung <tranhung44@gmail.com>
Reported-by: Tobias Maier <tobias.maier@netplace.com>
Signed-off-by: Andy Green <andy@warmcat.com>
M K [Sun, 25 Sep 2011 09:34:35 +0000 (10:34 +0100)]
Fix compile issue on Apple
Signed-off-by: M K <mkroehnert42@googlemail.com>
Andy Green [Sun, 25 Sep 2011 09:30:26 +0000 (10:30 +0100)]
correct cpp extern only being on win32
Signed-off-by: Andy Green <andy@warmcat.com>
Andy Green [Sun, 25 Sep 2011 09:01:02 +0000 (10:01 +0100)]
test html uplevel detect browser and MozWebSocket
Signed-off-by: Andy Green <andy@warmcat.com>
Andy Green [Sun, 25 Sep 2011 08:32:54 +0000 (09:32 +0100)]
v08v13 add support v13 default
Signed-off-by: Andy Green <andy@warmcat.com>
Andy Green [Sun, 25 Sep 2011 07:47:57 +0000 (08:47 +0100)]
v08 receiver ignore frame with unhandled ext opcode
Signed-off-by: Andy Green <andy@warmcat.com>
Andy Green [Sun, 25 Sep 2011 07:47:53 +0000 (08:47 +0100)]
v08 change ignore incoming set ext bits
Signed-off-by: Andy Green <andy@warmcat.com>
Andy Green [Sun, 26 Jun 2011 09:27:28 +0000 (10:27 +0100)]
my compiler version didn't find it, but there is an uninitialized usage of
len pointed out by Sonic on the blog
Reported-by: Sven <sven@killig.de>
Signed-off-by: Andy Green <andy@warmcat.com>
Andy Green [Sat, 28 May 2011 09:19:19 +0000 (10:19 +0100)]
extensions fix deflate stream vs mux and veto it on mux children
This patch gets deflate-stream working with x-google-mux.
It adds a clean veto system where are extension can veto the proposal
of any extension when opening a new connection. x-google-mux uses that
in its callback to defeat any use of deflate-stream on mux children.
However deflate stream is allowed on the parent connection and works
transparently now alongside x-google-mux.
Signed-off-by: Andy Green <andy@warmcat.com>
Andy Green [Thu, 26 May 2011 11:59:13 +0000 (12:59 +0100)]
x google mux handle illegal channel add indexes cleanup handle ch1 rx
Signed-off-by: Andy Green <andy@warmcat.com>
Andy Green [Wed, 25 May 2011 20:41:57 +0000 (21:41 +0100)]
extensions introduce pre close veto
This establishes a pre-close extension check to allow an extension to
veto a close. x-google-mux then uses this to stop ch1 going down
(subchannel 1 is the original socket connection that turns into a mux
parent) if it has active mux children; it just marks ch1 as closed in
its conn struct in that case and returns 1 from the callback to veto.
Code is also added to take care of the case ch1 is 'closed', and the
last child is subsequently closed, it actively calls close on the mux
parent then.
Signed-off-by: Andy Green <andy@warmcat.com>
Andy Green [Wed, 25 May 2011 20:07:20 +0000 (21:07 +0100)]
extensions change child count to highest child subchannel
This also changes the wsi_children array to be indexed by subchannel - 2,
non-existent channels are NULL in there and highest_child_subchannel
is a highwater mark for the highest subchannel ever used. That way we
can immediately get the wsi for a subchannel at the cost of some extra
sparse skipping during management.
This also takes care of scanning for empty slots on allocation and
NULLing on close of subchannel instead of deletion.
Signed-off-by: Andy Green <andy@warmcat.com>
Andy Green [Wed, 25 May 2011 19:46:18 +0000 (20:46 +0100)]
extensions ensure per conn allocation starts zeroed
Signed-off-by: Andy Green <andy@warmcat.com>
Andy Green [Wed, 25 May 2011 11:16:00 +0000 (12:16 +0100)]
x google mux complete introduction of MUX_REAL_CHILD_INDEX_OFFSET that got included accidentally
Accidentally included half-done MUX_REAL_CHILD_INDEX_OFFSET addition a couple
of patches back after late night hacking, this patch completes it.
This makes mux added channels start at channel index 2.
1 is reserved for original carrier channel, 0 for as-yet badly defined global mux commands
Signed-off-by: Andy Green <andy@warmcat.com>