* Problem: TIPC availability check is too strict
Solution: at build time only check if the API is available. In the tests
do a first check and a skip if the functionality is not available.
TIPC needs an in-tree but not loaded by default kernel module, tipc.ko
to be loaded, which requires root, so it is unlikely to be available on
any build system by default.
This will allow most distributions to ship with TIPC support built in,
and to avoid tests failure if the module is not there.
* Problem: no Travis tests for TIPC
Solution: mark one job with sudo: required and load the kernel module
* Problem: CMake fails when test returns 77 (skip)
Solution: set property to let it mark the test as skipped as intended
The zero copy decoding strategy implemented for 4.2.0 can lead to a large
increase of main memory usage in some cases (I have seen one program go up to
40G from 10G after upgrading from 4.1.4). This commit adds a new option to
contexts, called ZMQ_ZERO_COPY_RECV, which allows one to switch to the old
decoding strategy.
Solution: use ZMQ_LAST_ENDPOINT in most places. This alllows running
tests in paralle, and on over-booked shared machines where many of
the ports would be already in use.
Keep 3 tests with an hardcoded port, as there are some code paths that
require it (eg: connect before bind), but list those ports in
tests/testutil.hpp as macros so that they do not overlap and still
allow parallel runs.
These changes were inspired by a patch uploaded to Ubuntu by the
package maintainer, Steve Langasek <steve.langasek@ubuntu.com>.
Thank you Steve!
Solution: add helper function is_ipv6_available to testutil.hpp to
test if IPv6 is available on the building platform.
This function will try to open and bind a socket to ::1:*, as it's
the ultimate way of knowing if, at least on the loopback, IPv6 is
enabled.
Solution: it's a lot of work to define the tests in project.gyp
so I did this using gsl to generate the JSON, from a small XML
list of the test cases.
To keep this, and the hundreds of .mk files, away from the root
directory, I've moved the gyp files into builds/gyp, where you
would run them.
It all seems to work now. Next up, OS/X and Windows :)
This is due to the mangled include of platform.h, which was to make
CMake happy.
Solution: in CMakeLists.txt, define USING_CMAKE and then look for
platform.h in current directory if that is defined, else look in
../src/ as one would expect.
Of course people still "can" distributed the sources under the
LGPLv3. However we provide COPYING.LESSER with additional grants.
Solution: specify these grants in the header of each source file.
Symptom is that ZMQ_STREAM sockets in 4.1.0 and 4.1.1 generate zero
sized messages on each new connection, unlike 4.0.x which did not do
this.
Person who made this commit also changed test cases so that contract
breakage did not show. Same person was later banned for persistently
poor form in CZMQ contributions.
Solution: enable connect notifications on ZMQ_STREAM sockets using a
new ZMQ_STREAM_NOTIFY setting. By default, socket does not deliver
notifications, and behaves as in 4.0.x.
Fixes#1316
Users who need e.g. zmq_curve_keypair() have to remember to include
zmq_utils.h, which is counter-intuitive. The whole library should be
represented by a single include file.
Solution: merge all contents of zmq_utils.h into zmq.h, and deprecate
zmq_utils.h. Existing apps can continue unchanged. New apps can ignore
zmq_utils.h completely.
The expect_bounce_fail () helper assumed that messages could always
be sent. However in some cases zmq_send() blocks, due to there not
being any outgoing pipe. This changed in 77f5f7, where previously
there would be a pipe that kept trying to reconnect forever.
Solution: use a send timeout and check for EAGAIN if sending failed.
Specifically:
* zmq_event_t should not be used internally in libzmq, it was
meant to be an outward facing structure.
* In 4.x, zmq_event_t does not correspond to monitor events, so
I removed the structure entirely.
* man page for zmq_socket_monitor is incomplete and the example
code was particularly nasty.
* test_monitor.cpp needed rewriting, it was not clean.
- used msleep (10) in most places instead of zmq_sleep (1)
- may cause failures on slower machines
- to change, modify SETTLE_TIME in testutil.h
- tested down to 1 msec on fast boxes
