platform/upstream/libwebsockets.git
11 years agotrac 29 convert more snprintf stowaways
Andy Green [Thu, 2 May 2013 00:15:53 +0000 (08:15 +0800)]
trac 29 convert more snprintf stowaways

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agotrac 29 echo remoe mention of syslog constants for win32
Andy Green [Wed, 1 May 2013 23:04:54 +0000 (07:04 +0800)]
trac 29 echo remoe mention of syslog constants for win32

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agotrac 29 avoid syslog in echo test if win32
Andy Green [Mon, 29 Apr 2013 22:53:56 +0000 (06:53 +0800)]
trac 29 avoid syslog in echo test if win32

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agotrac 29 eliminate snprintf again
Andy Green [Mon, 29 Apr 2013 22:51:10 +0000 (06:51 +0800)]
trac 29 eliminate snprintf again

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agotrac 29 win32 EWOULDBLOCK
Andy Green [Sat, 27 Apr 2013 00:33:26 +0000 (08:33 +0800)]
trac 29 win32 EWOULDBLOCK

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoTrac 29: win32 fix LWS_VISIBLE issue
Andy Green [Thu, 25 Apr 2013 23:37:16 +0000 (07:37 +0800)]
Trac 29: win32 fix LWS_VISIBLE issue

Reported-by: davetaflin
Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agocygwin compatibility
Bob Roberts [Thu, 25 Apr 2013 01:16:30 +0000 (09:16 +0800)]
cygwin compatibility

Signed-off-by: Bob Roberts <bobroberts177@gmail.com>
11 years agotrac 27: use fullpath in OSX dynamic libs
thinkski [Tue, 16 Apr 2013 11:48:05 +0000 (19:48 +0800)]
trac 27: use fullpath in OSX dynamic libs

Signed-off-by: thinkski <chiszp@gmail.com>
11 years agopkg config avoid hardcoded prefix
thinkski [Fri, 12 Apr 2013 02:31:46 +0000 (10:31 +0800)]
pkg config avoid hardcoded prefix

Signed-off-by: thinkski <chiszp@gmail.com>
11 years agocmake use existing spec file for RPM generation
Tomas Cejka [Thu, 11 Apr 2013 00:13:13 +0000 (08:13 +0800)]
cmake use existing spec file for RPM generation

Signed-off-by: Tomas Cejka <cejkat@cesnet.cz>
11 years agocmake added RPM generation
Tomas Cejka [Thu, 11 Apr 2013 00:38:02 +0000 (08:38 +0800)]
cmake added RPM generation

Signed-off-by: Tomas Cejka <cejkat@cesnet.cz>
11 years agocmake improve cross documentation
Andy Green [Sat, 30 Mar 2013 02:30:03 +0000 (10:30 +0800)]
cmake improve cross documentation

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoHide private library symbols
Peter Pentchev [Sat, 30 Mar 2013 01:52:21 +0000 (09:52 +0800)]
Hide private library symbols

Signed-off-by: Peter Pentchev <roam@ringlet.net>
11 years agoSOL_SNDBUF set to protocol rx size
Andy Green [Sat, 23 Mar 2013 01:53:17 +0000 (09:53 +0800)]
SOL_SNDBUF set to protocol rx size

This tells the OS to reserve a TX buffer at least the size of the biggest RX frame
expected, for both server and client connections.

In Linux, the OS reserves 2 x the requested amount.

This is aimed at reducing the partial large atomic frame send problem to the point
it's only coming at large atomic frames the OS balks at reserving the size for.

If you have a lot of data to send, it is better to split it into multiple writes,
and use the FIN / CONTINUATION websocket stuff to manage it.

See the "fraggle" test app for example code of how to do that.

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agotest server improve return checking
Andy Green [Sat, 23 Mar 2013 01:46:18 +0000 (09:46 +0800)]
test server improve return checking

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agocosmetic debug improvements
Andy Green [Sat, 23 Mar 2013 01:44:47 +0000 (09:44 +0800)]
cosmetic debug improvements

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoBUGFIX: Properly add libcyassl to the lib list.
Joakim Soderberg [Sat, 23 Mar 2013 00:55:28 +0000 (08:55 +0800)]
BUGFIX: Properly add libcyassl to the lib list.

This would cause failure to build the test apps with CyaSSL.

11 years agoCheck for GCC/G++ explicitly instead of simply Unix.
Joakim Soderberg [Sat, 23 Mar 2013 00:55:02 +0000 (08:55 +0800)]
Check for GCC/G++ explicitly instead of simply Unix.

Since we're setting gcc specific build flags.

11 years agoAllow setting the resource path via command line.
Joakim Soderberg [Mon, 18 Mar 2013 17:08:25 +0000 (17:08 +0000)]
Allow setting the resource path via command line.

It is awfully limiting when you have to set the resource path at compile time.

Instead also allow setting it via the --resource_path command line option.

11 years agointroduce api for unthrottle all connections of protocol
Andy Green [Sat, 16 Mar 2013 04:32:27 +0000 (12:32 +0800)]
introduce api for unthrottle all connections of protocol

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agotest client have u option also reject deflate stream
Andy Green [Sat, 16 Mar 2013 04:34:36 +0000 (12:34 +0800)]
test client have u option also reject deflate stream

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agocmake add gcc compile options
Andy Green [Sat, 16 Mar 2013 04:32:59 +0000 (12:32 +0800)]
cmake add gcc compile options

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agorxflow remove recursion and simplify
Andy Green [Sat, 16 Mar 2013 03:24:23 +0000 (11:24 +0800)]
rxflow remove recursion and simplify

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoreplaced last remaining close on wsi->sock with compatible_close
Tobias Maier [Thu, 14 Mar 2013 12:02:23 +0000 (13:02 +0100)]
replaced last remaining close on wsi->sock with compatible_close

Signed-off-by: Tobias Maier <tobias.maier@netplace.com>
11 years agoFixed read of the jpg in Windows
Michel Archambault [Thu, 14 Mar 2013 15:45:49 +0000 (11:45 -0400)]
Fixed read of the jpg in Windows

Windows needs to open things with _O_BINARY

Signed-off-by: Michel Archambault <marchamb@matrox.com>
11 years agoupdate service_fd comment and api docs
Andy Green [Mon, 11 Mar 2013 12:37:03 +0000 (20:37 +0800)]
update service_fd comment and api docs

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agocmake update specfile
Andy Green [Sat, 9 Mar 2013 23:16:04 +0000 (07:16 +0800)]
cmake update specfile

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agocmake add missing echo test app
Andy Green [Sun, 10 Mar 2013 00:09:25 +0000 (08:09 +0800)]
cmake add missing echo test app

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agocmake packageconfig installs to lib dir
Andy Green [Sun, 10 Mar 2013 00:10:19 +0000 (08:10 +0800)]
cmake packageconfig installs to lib dir

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agocmake ignore non source things for make dist
Andy Green [Sun, 10 Mar 2013 00:11:48 +0000 (08:11 +0800)]
cmake ignore non source things for make dist

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoandroid include gregorys example makefile
Gregory Junker [Sat, 9 Mar 2013 23:16:26 +0000 (07:16 +0800)]
android include gregorys example makefile

Signed-off-by: Gregory Junker <ggjunker@gmail.com>
11 years agoupdate version 1.3 soname 4
Andy Green [Sat, 9 Mar 2013 22:59:01 +0000 (06:59 +0800)]
update version 1.3 soname 4

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agocmake some cosmetic improvementss
Edwin van den Oetelaar [Sat, 9 Mar 2013 22:53:57 +0000 (06:53 +0800)]
cmake some cosmetic improvementss

Signed-off-by: Edwin van den Oetelaar <oetelaar.automatisering@gmail.com>
11 years agoautotools remove INSTALL m4 COPYING
Edwin van der Oetelaar [Sat, 9 Mar 2013 22:43:29 +0000 (06:43 +0800)]
autotools remove INSTALL m4 COPYING

Also get rid of m4/ and COPYING, it's a copy of LGPL2.1 but without the
exceptions we added in LICENSE.

Signed-off-by: Edwin van der Oetelaar <oetelaar.automatisering@gmail.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoremove autotools build
Andy Green [Sat, 9 Mar 2013 05:06:37 +0000 (13:06 +0800)]
remove autotools build

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agodocument revent clearing scheme
Andy Green [Sat, 9 Mar 2013 05:03:53 +0000 (13:03 +0800)]
document revent clearing scheme

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoclear down wsi_lookup on allocation and clear pollfd revents if serviced
Andy Green [Sat, 9 Mar 2013 04:34:30 +0000 (12:34 +0800)]
clear down wsi_lookup on allocation and clear pollfd revents if serviced

This should get rid of a valgrind uninitialized report when using extpoll,
and gives a new way to share the poll loop with external sockets.

If a pollfd says it has something, you can just pass it to
libwebsocket_serice_fd() whether it is a socket handled by lws or not.
If it sees it is a lws socket, the traffic will be handled and
pollfd->revents will be zeroed now.

If the socket is foreign to lws, it leaves revents alone.  So you can see
if you should service by checking the pollfd revents after letting
lws try to service it.

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoreduce httpd close debug spew
Andy Green [Sat, 9 Mar 2013 04:54:14 +0000 (12:54 +0800)]
reduce httpd close debug spew

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agocmake fix install path
Andy Green [Sat, 9 Mar 2013 04:51:43 +0000 (12:51 +0800)]
cmake fix install path

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agocorrect example lws_write in header
Alex Rhatushnyak [Sat, 9 Mar 2013 04:01:47 +0000 (12:01 +0800)]
correct example lws_write in header

Signed-off-by: Alex Rhatyshnyak <a.rhatushnyak@klervi.com>
11 years agodont leak when protocol agreement probs
Andy Green [Sat, 9 Mar 2013 03:52:18 +0000 (11:52 +0800)]
dont leak when protocol agreement probs

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoFix printf format error
Henrik Abelsson [Thu, 7 Mar 2013 14:52:53 +0000 (15:52 +0100)]
Fix printf format error

Signed-off-by: Henrik Abelsson <henrik@abelsson.com>
11 years agocmake win32 change helper includes to HDR_PUBLIC so they install
Andy Green [Sat, 9 Mar 2013 01:16:41 +0000 (09:16 +0800)]
cmake win32 change helper includes to HDR_PUBLIC so they install

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agodont sweat unkknown method for client
Andy Green [Sat, 9 Mar 2013 01:09:05 +0000 (09:09 +0800)]
dont sweat unkknown method for client

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoPOLLOUT during wait for server response not an error
Andy Green [Sat, 9 Mar 2013 01:09:46 +0000 (09:09 +0800)]
POLLOUT during wait for server response not an error

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agofix win32helpers gettimeofday epoch
Arokux B [Tue, 5 Mar 2013 00:41:47 +0000 (08:41 +0800)]
fix win32helpers gettimeofday epoch

Signed-off-by: Arokux B <arokux@gmail.com>
11 years agotrace 22 fix zero length close from client
Andy Green [Thu, 28 Feb 2013 09:11:29 +0000 (17:11 +0800)]
trace 22 fix zero length close from client

close() from javascript api in Chrome and Firefox doesn't do the
right thing.  It's because the payload is zero-length (with a frame
key...)  This fixes it.

Reported-by: 巫书轶
Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agocmake mingw no need for websock w32
Arokux B [Thu, 28 Feb 2013 01:25:02 +0000 (09:25 +0800)]
cmake mingw no need for websock w32

2.  Comment out this line

http://git.warmcat.com/cgi-bin/cgit/libwebsockets/tree/CMakeLists.txt#n279

the file websock-w32.c is already compiled indirectly from libwebsockets.c

Signed-off-by: Arokux B <arokux@gmail.com>
11 years agonormalize formatting in gettimeofday
Andy Green [Thu, 28 Feb 2013 01:22:38 +0000 (09:22 +0800)]
normalize formatting in gettimeofday

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agofix win32 gettimeofday for mingw
Arokux B [Thu, 28 Feb 2013 01:18:38 +0000 (09:18 +0800)]
fix win32 gettimeofday for mingw

here are the problems that I found while compiling libwebsockets with
CMake using as a generator MinGW Makefiles.

1. In http://git.warmcat.com/cgi-bin/cgit/libwebsockets/plain/win32port/win32helpers/gettimeofday.c

Remove spaces around time.h, otherwise it looks for " time.h " and
fails with "No such file ..."

I also needed to comment out

struct timezone
{
  int  tz_minuteswest; /* minutes W of Greenwich */
  int  tz_dsttime;     /* type of dst correction */
};

It is already declared in mingw's time.h

Signed-off-by: Arokux B <arokux@gmail.com>
11 years agocmake lib lib64 problem
Andy Green [Tue, 26 Feb 2013 03:58:45 +0000 (11:58 +0800)]
cmake lib lib64 problem

As discussed here, there's a semi-standardized way to inform
cmake that the install target wants its libs in .../lib64

https://code.google.com/p/bullet/issues/detail?id=174

this patch adds the necessary fiddling

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agofix mime type on leaf header
Luc [Sat, 23 Feb 2013 03:01:21 +0000 (11:01 +0800)]
fix mime type on leaf header

Signed-off-by: Luc <luc@honk-honk.com>
11 years agoreflect send completeness in lws_write return
Andy Green [Sat, 23 Feb 2013 02:50:10 +0000 (10:50 +0800)]
reflect send completeness in lws_write return

under load, writing packet sizes to the socket that are normally fine
can do partial writes, eg asking to write 4096 may only take 2800 of
it and return 2800 from the actual send.

Until now lws assumed that if it was safe to send, it could take any
size buffer, that's not the case under load.

This patch changes lws_write to return the amount actually taken...
that and the meaning of it becomes tricky when dealing with
compressed links, the amount taken and the amount sent differ.  Also
there is no way to recover at the moment from a protocol-encoded
frame only being partially accepted... however for http file send
content it can and does recover now.

Small frames don't have to take any care about it but large atomic
sends (> 2K) have been seen to fail under load.

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agomove ssl cipher list from compiletime to context creation time option
Andy Green [Fri, 22 Feb 2013 01:54:35 +0000 (09:54 +0800)]
move ssl cipher list from compiletime to context creation time option

More flexible this way... NULL for the new member means use
the ssl library default set of ciphers, so as long as your info
struct is zerod by bss or memset, you don't need to do anything
about this change unless you want to set the cipher list.

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoremove MAX_HEADER_NAME_LENGTH from header
Andy Green [Fri, 22 Feb 2013 01:38:54 +0000 (09:38 +0800)]
remove MAX_HEADER_NAME_LENGTH from header

Still lurking aound as a ghost after all real usess of it
were removed with the minilex changes.

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoFix running test programs from within visual studio.
Joakim Soderberg [Fri, 22 Feb 2013 01:28:17 +0000 (09:28 +0800)]
Fix running test programs from within visual studio.

Copy test-server data to a path so that we can run the test-server from
within visual studio without having to copy stuff around.

11 years agoPrint SSL error codes as well as the string.
Joakim Soderberg [Fri, 22 Feb 2013 01:28:15 +0000 (09:28 +0800)]
Print SSL error codes as well as the string.

CYASSL doesn't compile with the error strings in some configurations, so
it is nice to have the error code on SSL errors as well.

11 years agoSet the _DEBUG macro for CMake also.
Joakim Söderberg [Fri, 22 Feb 2013 01:28:13 +0000 (09:28 +0800)]
Set the _DEBUG macro for CMake also.

11 years agoFixed ssl cert generation on Windows.
Joakim Soderberg [Fri, 22 Feb 2013 01:28:11 +0000 (09:28 +0800)]
Fixed ssl cert generation on Windows.

11 years agoAdded CPack support + some more.
Joakim Soderberg [Fri, 22 Feb 2013 01:28:08 +0000 (09:28 +0800)]
Added CPack support + some more.

- "make dist" will now produce a tar.gz file. This includes everything (not really what we want but a start).
- Got rid of a bunch of defines and variables that aren't used.
- Added the option LINK_TESTAPPS_DYNAMIC that enables linking to the shared library version of the lib for the test apps, instead of doing it statically like the default is.
- Fixed proper support for the --prefix stuff (-DCMAKE_INSTALL_PREFIX for cmake). (Don't specify /usr/local explicitly, that is the default anyway and will break other platforms).

- Note: I noticed a problem with the "INSTALL_DATADIR" define used by the test-apps. Since we hard code the path to the certs using this, doing "DESTDIR=/bla make install" will result in not being able to use the SSL mode for the test-apps since they won't find the certs. (This also applies to the autoconf project). Fixed this by setting "../share" as the default location instead of using the prefix.

11 years agoBUGFIX: Fixed bug for installing test-apps for CMake.
Joakim Soderberg [Fri, 22 Feb 2013 01:28:06 +0000 (09:28 +0800)]
BUGFIX: Fixed bug for installing test-apps for CMake.

Appending to the TEST_APP_LIST variable function did not work from within the create_test_app function (It would create a function scope variable with the same name and discard it). We instead need to use a macro for this.

The previous commit also removed the duplicate install for the test-apps (because we want all the install steps in one place in the file for clarity). That change made it obvious that the TEST_APP_LIST never gets populated.

11 years agoFixed CMake compile options.
Joakim Soderberg [Fri, 22 Feb 2013 01:28:04 +0000 (09:28 +0800)]
Fixed CMake compile options.

Fixed so that the build options for the CMake project works:

- The test apps used the LWS_NO_EXTENSIONS define, so they needed lws_config.h included when building using CMake.
- Rename some options so that individual test apps can be turned off.
- Separate building the test-client/test-server and compiling the server/client parts into the lib.
- Don't include server or client specific sources into the build if they shouldn't be built.
- Added an error if both client and server parts are excluded at the same time (makes no sense).
- Removed duplicate install targets for the test apps.
- Commented out the WITH_LIBCRYPTO option since it isn't used at the moment.

11 years agogetifaddrs missing more user friendly in CMake.
Joakim Soderberg [Fri, 22 Feb 2013 01:28:02 +0000 (09:28 +0800)]
getifaddrs missing more user friendly in CMake.

Don't require the user to enable using the built-in BSD getifaddrs implementation on systems such as uclibc that lacks it manually.

Instead if getifaddrs doesn't exist, use the BSD one automatically, except if the user explicitly tells the user not to do this using WITHOUT_BUILTIN_GETIFADDRS (which will result in a compilation error, but at least with a nice error message explaining why).

11 years agoFix LWS_NO_SERVER build.
Joakim Soderberg [Fri, 22 Feb 2013 01:27:59 +0000 (09:27 +0800)]
Fix LWS_NO_SERVER build.

"int n" was used also when server parts are not compiled, so don't put it inside the LWS_NO_SERVER ifdef.

11 years agoBugfix compiling for cross compiling.
Joakim Soderberg [Fri, 22 Feb 2013 01:27:57 +0000 (09:27 +0800)]
Bugfix compiling for cross compiling.

! does not mean NOT, use NOT instead :)

11 years agokeepalive swap interval and probes
Larry Hayes [Fri, 22 Feb 2013 01:16:20 +0000 (09:16 +0800)]
keepalive swap interval and probes

The interval and probes members for setting keepalive options
per-connection on Linux were swapped.

Signed-off-by: Larry Hayes <lhayesg@gmail.com>
11 years agoprovide socketfd at in param for LWS_CALLBACK_FILTER_NETWORK_CONNECTION
Edwin van den Oetelaar [Wed, 20 Feb 2013 12:56:59 +0000 (20:56 +0800)]
provide socketfd at in param for LWS_CALLBACK_FILTER_NETWORK_CONNECTION

Signed-off-by: Edwin van den Oetelaar <oetelaar.automatisering@gmail.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agotrac 18 deal with service connection timing out
Andy Green [Wed, 20 Feb 2013 11:11:31 +0000 (19:11 +0800)]
trac 18 deal with service connection timing out

While looking at http://libwebsockets.org/trac/ticket/18
noticed the flow for timeout in service_fd will do bad things
if the fd we came to service has timed out.  It gets freed and
then "serviced'.

Reported-by: Joakim Soderberg <joakim.soderberg@gmail.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agocmake fix for older cmake
Edwin van den Oetelaar [Tue, 19 Feb 2013 18:08:51 +0000 (02:08 +0800)]
cmake fix for older cmake

AG this wasn't needed on cmake 2.8.10.2 but did not make any
problem either

Signed-off-by: Edwin van den Oetelaar <oetelaar.automatisering@gmail.com>
11 years agoupdate missed extpoll calls to use correct args for ssl
Andy Green [Tue, 19 Feb 2013 11:19:51 +0000 (19:19 +0800)]
update missed extpoll calls to use correct args for ssl

http://libwebsockets.org/trac/ticket/15#comment:16

Reported-by: Joakim Soderberg <joakim.soderberg@gmail.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agocmake docs add note about prefix
Andy Green [Tue, 19 Feb 2013 02:26:39 +0000 (10:26 +0800)]
cmake docs add note about prefix

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoCreate the "doc" dir before generating docs.
Joakim Söderberg [Tue, 19 Feb 2013 02:14:54 +0000 (10:14 +0800)]
Create the "doc" dir before generating docs.

The documentation wouldn't be generated if the path didn't exist.

11 years agoAdded "make install" support to the CMake project.
Joakim Söderberg [Tue, 19 Feb 2013 02:14:37 +0000 (10:14 +0800)]
Added "make install" support to the CMake project.

- Does everything as the autoconf script does (I think).
- Generates SSL certs for the test-server
- Installs all test apps.
- Defaults to /usr/local/
- "DESTDIR=/bla make install" works
- Append "libwebsockets" to start of the test-app names like the autoconf stuff does.
- Only tested on OSX so far...

11 years agotrac 17 update set_log_level api notice
Andy Green [Tue, 19 Feb 2013 02:01:48 +0000 (10:01 +0800)]
trac 17 update set_log_level api notice

http://libwebsockets.org/trac/ticket/17

Reported-by: dbrnz
Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agomove cms cmake addition to cmake module paths earlier
Edwin van den Oetelaar [Mon, 18 Feb 2013 12:20:05 +0000 (20:20 +0800)]
move cms cmake addition to cmake module paths earlier

Signed-off-by: Edwin van den Oetelaar <oetelaar.automatisering@gmail.com>
11 years agoadd FindGit.cmake
Andy Green [Mon, 18 Feb 2013 10:34:10 +0000 (18:34 +0800)]
add FindGit.cmake

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agodocument ensure_user_space going private
Andy Green [Mon, 18 Feb 2013 08:54:26 +0000 (16:54 +0800)]
document ensure_user_space going private

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoremove lws_ensure_user_space from public api change return
Andy Green [Mon, 18 Feb 2013 08:30:10 +0000 (16:30 +0800)]
remove lws_ensure_user_space from public api change return

The function has a logical problem when the size of the requested
allocation is 0, it will return NULL which is overloaded as
failure.

Actually the whole function is evil as an api, this patch moves
it out of the public API space and fixes it to return 0 for
success or 1 for fail.  Private code does not need to to return
wsi->user_space and public code should only get that from the
callback as discussed on trac recently.

Thanks to Edwin for debugging the problem.

Reported-by: Edwin van den Oetelaar <oetelaar.automatisering@gmail.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoupdate arm build stats
Andy Green [Mon, 18 Feb 2013 04:08:15 +0000 (12:08 +0800)]
update arm build stats

Clean out the historical numbers, update info

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agotrim proxy ads plus hostname stg
Andy Green [Mon, 18 Feb 2013 04:02:18 +0000 (12:02 +0800)]
trim proxy ads plus hostname stg

128 should be fine for hostname and proxy ads

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agorestrict http send buffer to 4096
Andy Green [Mon, 18 Feb 2013 03:32:49 +0000 (11:32 +0800)]
restrict http send buffer to 4096

send() is able to complete but not send everything asked...
we don't deal with that yet.  On an ARM9 board, leaf.jpg
failed because the send() only accepted 7240 of 8192... not
a failure but it implies reporting partial send to the
caller.  Work around by keeping the send buffer at 4096.

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agofix no extensions build
Andy Green [Mon, 18 Feb 2013 02:43:18 +0000 (10:43 +0800)]
fix no extensions build

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agomigrate client hs and c_port into ah
Andy Green [Mon, 18 Feb 2013 02:38:45 +0000 (10:38 +0800)]
migrate client hs and c_port into ah

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agominor type optimizations
Andy Green [Mon, 18 Feb 2013 02:34:25 +0000 (10:34 +0800)]
minor type optimizations

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoremove current_alloc_len
Andy Green [Mon, 18 Feb 2013 02:29:22 +0000 (10:29 +0800)]
remove current_alloc_len

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoremove header name buffer
Andy Green [Mon, 18 Feb 2013 02:22:42 +0000 (10:22 +0800)]
remove header name buffer

The header name buffer and its max length handling has actually
been unused since the minilex parser was introduced.  We hold
parsing state in the lex-type parts and don't need to store or
worry about max length, since the parser will let us know as
soon as it can't be a match for the valid header names.

This strips it out reducing the per-connection allocation for
x86_64 with default configure from 224 to 160.

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agohandle http union transition properly
Andy Green [Mon, 18 Feb 2013 01:48:31 +0000 (09:48 +0800)]
handle http union transition properly

There's no proper transition to http union state until now.

It only becomes apparant there's a problem when you try to
return -1 from the HTTP callback, during the close action
it will try to close() a nonsense, nonzero fd pointer in
the uninitialized u.http union member.

This patch takes a copy of the allocated headers struct from
the u.hdr union state, transitions to u.http clearing down u
and then calls the HTTP callback with URI args pointing to
the still-in-scope ah allocation.  After the call, the copy ah
is freed.

That makes sure we are in the correct union state while still
giving the HTTP callback access to the URI without having to
copy it around.

Reported-by: Edwin can den Oetelaar <oetelaar.automatisering@gmail.com>
Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agotest server kill skt with minus 1
Andy Green [Sun, 17 Feb 2013 01:14:08 +0000 (09:14 +0800)]
test server kill skt with minus 1

The client spams the server with mirror protocol content
and then closes the connection.  However the server wants to
send that content back to all the mirror protocol
connections, including the one that closed.

The result is during the send back phase we usually see an
error trying to send to the close client connection. Because
we don't return -1, we don't close it on server side and the
error repeats a lot before finally closing.  This is a side-
effect of a recent patch to drain pending input before
really closing a dead connection.

This patch just changes it to return -1 on the failed send
attempt assertively closing the connection.

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agohandle any POLLIN before error
Andy Green [Sat, 16 Feb 2013 02:17:52 +0000 (10:17 +0800)]
handle any POLLIN before error

See if this can impact http://libwebsockets.org/trac/ticket/16

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoclose if we tried to close politely just close next time
Andy Green [Fri, 15 Feb 2013 14:48:58 +0000 (22:48 +0800)]
close if we tried to close politely just close next time

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoapi change deliver socket fd to in param of extpoll callbacks
Andy Green [Fri, 15 Feb 2013 14:36:30 +0000 (22:36 +0800)]
api change deliver socket fd to in param of extpoll callbacks

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoproblems in lws_handle_POLLOUT_event should just close
Andy Green [Fri, 15 Feb 2013 14:32:53 +0000 (22:32 +0800)]
problems in lws_handle_POLLOUT_event should just close

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agorobustness protect and document ensure_user_space
Andy Green [Fri, 15 Feb 2013 14:31:55 +0000 (22:31 +0800)]
robustness protect and document ensure_user_space

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoREADME.coding add note about closing connections
Andy Green [Fri, 15 Feb 2013 02:25:58 +0000 (10:25 +0800)]
README.coding add note about closing connections

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoadd by hand http send example
Andy Green [Thu, 14 Feb 2013 15:06:37 +0000 (23:06 +0800)]
add by hand http send example

This adds a demonstration of how to use the http nonblocking
send action to the test server.

If you ask for /leaf.jpg from the test server, it will send
"by hand" a 2.4MB jpeg in HTTP, including the headers.  See
the test server sources to see how it is done.

Although it's a file, and a jpeg image, actually the exact
same scheme will work for any data or mime type.

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agocreate user alloc for http if requested
Andy Green [Thu, 14 Feb 2013 15:18:10 +0000 (23:18 +0800)]
create user alloc for http if requested

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agointroduce LWS_CALLBACK_HTTP_WRITEABLE
Andy Green [Thu, 14 Feb 2013 14:23:54 +0000 (22:23 +0800)]
introduce LWS_CALLBACK_HTTP_WRITEABLE

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agojust get hostname into canonical_hostname
Andy Green [Thu, 14 Feb 2013 09:11:22 +0000 (17:11 +0800)]
just get hostname into canonical_hostname

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agoremove getnameinfo
Andy Green [Thu, 14 Feb 2013 09:02:26 +0000 (17:02 +0800)]
remove getnameinfo

Signed-off-by: Andy Green <andy.green@linaro.org>
11 years agofragge able to send chunks larger than rx buffer
Andy Green [Thu, 14 Feb 2013 06:15:58 +0000 (14:15 +0800)]
fragge able to send chunks larger than rx buffer

This now exercises the code for dealing with rx that spills over
the rx buffer.  Single fraggle uses the default rx buffer size
of 4096, it also now exercises code around that.

Signed-off-by: Andy Green <andy.green@linaro.org>