Solution: if a PING message contains a context, echo it back in the
PONG message. In order to do so, create the PONG message when PING
is received and store it in the engine.
After the PING the engine goes straight to encoding and sending, so
there can always be at most one pending PING.
Add tests for various contexts.
Solution: treat the first byte of the command body as the size of the
command name, rather than as an id, to comply with ZMTP 3.1.
This was not an actual problem at runtime since both heartbeat
commands have a size of 4, which was treated like an id.
But once SUBSCRIBE/UNSUBSCRIBE get implemented it needs to be checked.
Solution: replace it with the ip_resolver code shared with the TCP
address code
It simplifies the UDP parsing code and makes it behave more like the
TCP counterpart, in particular it's not possible to connect to hosts
by name and bind by NIC names.
It also adds support for "*" port resolving to 0 (useful to let the OS
allocate the port number).
Solution: return an error in this situation but still allow using an explicit
"0" if somebody really wants to connect to port 0.
This shouldn't break any existing code because a "*" port was already rejected
in an early test in the TCP path in zmq::socket_base_t::connect.
Solution: ignore command messages in the REQ session to avoid
disrupting the state machine.
Commands are handled by the engine before handing off to the session.
event_accepted() already accepts fd_t type and there is no reason to cast it to int type
Moreover, on Windows x64 this leads to truncation memsize -> int
Solution: revert the revert!
Revert "Problem: regression in 4.2.3 went unnoticed, want to release 4.2.5"
This reverts commit 5f17e26fa4c60c3de0282d1b6ad1e8b7037ed57a.
Solution: revert DRAFT -> STABLE API transition so that we can do a
bugfix-only 4.2.5 release.
Will be re-reverted once tagged.
Revert "Problem: ZMQ_BINDTODEVICE has met STABLE conditions"
This reverts commit 3cb79f5042cf32cdb7b1b58d4acf17eba85ec9f7.
Revert "Problem: ZMQ_MSG_GSSAPI_* have met STABLE conditions"
This reverts commit 374da4207b8034b0fcd67a2cc2165d50e09b9387.
Revert "Problem: ZMQ_MSG_T_SIZE has met STABLE conditions"
This reverts commit 6411c4a247c08ead50919d16b30eb030eaf44a7e.
Revert "Problem: docs say STABLE API still in DRAFT"
This reverts commit 9f2f30b7ffa09acc51d3b87251a47e83b435d5d4.
Solution: like for other mechanism, do not enforce strict ZAP protocol
adherence unless the specific socket option is enabled with NULL auth.
Add test to exercise this functionality, and fix ZAP test to set the
socket option when it uses NULL auth.
See: https://github.com/zeromq/pyzmq/pull/1152
Solution: move it from DRAFT to STABLE since it's been in a public
release, committed for 6+ months and has not changed.
Given a new STABLE symbol has been added, bump minor version number.
Solution: restrict it only to the original issue #2632, Tweetnacl on
*NIX when using /dev/urandom, ie: without the new Linux getrandom()
syscall.
Existing applications might use atexit to register cleanup functions
(like CZMQ does), and the current change as-is imposes an ordering
that did not exist before - the context MUST be created BEFORE
registering the cleanup with atexit. This is a backward incompatible
change that is reported to cause aborts in some applications.
Although libsodium's documentation says that its initialisation APIs
is not thread-safe, nobody has ever reported an issue with it, so
avoiding the global init/deinit in the libsodium case is the less
risky option we have.
Tweetnacl users on Windows and on Linux with getrandom (glibc 2.25 and
Linux kernel 3.17) are not affected by the original issue.
Fixes#2991
Lets the application set per-connection metadata.
Metadata is specified as "X-key:value" and set using zmq_setsockopt, eg:
zmq_setsockopt (s, ZMQ_METADATA, "X-key:value", 11);
The peer can then obtain the metadata from a received message:
char *data = zmq_msg_gets(msg, "X-key");
* Problem: Still need to port over more files to VxWorks 6.x
Solution: Port more files to VxWorks 6.x
* Problem: Need to port over remaining files to VxWorks 6.x. Also remove POSIX thread dependency for VxWorks (because of priority inversion problem in POSIX mutexes with VxWorks 6.x processes)
Solution: Port over remaining files to VxWorks 6.x. Also removed POSIX thread dependency for VxWorks
* Problem: Needed to modify TCP, UDP, TIPC classes with #ifdefs to be compatible with VxWorks 6.x.
Solution: Modify TCP, UDP, TIPC classes with #ifdefs to be compatible with VxWorks 6.x
* Tests for different TIPC address types and code cleanup
* Adds tests for binding/connecting with different TIPC address types
using Unity
* Adds error checking for address type misuse
Solution: do not rely __cplusplus >= 201103L to detect whether the
compiler supports thread safe static initialisation, but check only
the proper feature preprocessor macro.
GCC introduced it in version 8, and Clang in version 6.
Solution: initialise variable in options.cpp to dummy value to fix
compiler complaint:
src/options.cpp: In function
'int zmq::do_setsockopt_int_as_bool_strict(const void*, size_t, bool*)':
src/options.cpp:121:5: error: 'value' may be used uninitialized in
this function [-Werror=maybe-uninitialized]
src/options.cpp: In function
'int zmq::do_setsockopt_int_as_bool_relaxed(const void*, size_t, bool*)':
src/options.cpp:135:31: error: 'value' may be used uninitialized in
this function [-Werror=maybe-uninitialized]
