Running tests while writing failed assertions to a text file means the
little information there is that causes test failures becomes pretty
much inaccessible on Travis. Sending the output to stdout/stderr means
it can be trivially shown for failed tests using the --output-on-failure
option of ctest.
Signed-off-by: Erik Boasson <eb@ilities.com>
Connext (sometimes, for some versions?) sends a pre-emptive ACKNACK with
a base sequence number of 0, which were rejected following the DDSI
specification (8.3.4.1, 8.3.5.5 and 8.3.7.1.3). That doesn't really
help anyone. (https://github.com/ros2/ros2/issues/824)
Signed-off-by: Erik Boasson <eb@ilities.com>
Installing Cyclone DDS for multiple architectures with the same
installation prefix gave problems because the CMake configuration files
ended up overwriting each other. This brings it in line with the
recommended locations. Thanks to @hansfn.
Signed-off-by: Erik Boasson <eb@ilities.com>
PGUIDFMT creates invalid C++ code.
In file included from /opt/ros/master/src/ros2/rmw_cyclonedds/rmw_cyclonedds_cpp/src/serdata.cpp:23:
In file included from /opt/ros/master/install/include/dds/ddsi/q_radmin.h:17:
In file included from /opt/ros/master/install/include/dds/ddsi/ddsi_tran.h:19:
In file included from /opt/ros/master/install/include/dds/ddsi/q_protocol.h:19:
/opt/ros/master/install/include/dds/ddsi/q_rtps.h:30:21: error: invalid suffix on literal; C++11 requires a space between literal and identifier [-Wreserved-user-defined-literal]
^
/opt/ros/master/install/include/dds/ddsi/q_rtps.h:30:31: error: invalid suffix on literal; C++11 requires a space between literal and identifier [-Wreserved-user-defined-literal]
Signed-off-by: Dan Rose <dan@digilabs.io>
This allows you to express a dependency on it as `find_package(CycloneDDS REQUIRED COMPONENTS idlc)`
Also added a warning if CycloneDDS was built with IDLC but IDLC is not requested.
Signed-off-by: Dan Rose <dan@digilabs.io>
This will modify the default behavior to select BUILD_IDLC based on
whether Maven is discovered or not.
The behavior when `-DBUILD_IDLC` is specified on the command line remains
unchanged - the build will maintain the current behavior of failing to
configure if `BUILD_IDLC=ON` and Maven was later not found.
This reverts (part of) commit 860a6aadae.
Signed-off-by: Scott K Logan <logans@cottsay.net>
This adds a bunch of steps to the build process that verify
cyclonedds.xsd, cyclonedds.rnc and options.md as committed match the
configuration tables in the source.
The cyclonedds.rnc and options.md depend on having perl available,
cyclonedds.xsd on having Java and the "trang" conversion tool. Not
having these tools simply means some of the checks are skipped.
Signed-off-by: Erik Boasson <eb@ilities.com>
makernc: more forgiving of line endings input
Signed-off-by: Erik Boasson <eb@ilities.com>
Ignore line endings comparing cyclonedds XSD, RNC
Signed-off-by: Erik Boasson <eb@ilities.com>
This adds two things to the XSD (and the RNC file and options.md):
* attributes previously missing because of a bug in the conversion
script
* a name space (https://cdds.io/config)
Adding the name spacing requires a different set of attributes at the
top of a configuration file, which in turn need to be ignored by the
configuration parser, and which should be reflected in the configuration
example in the README.
Signed-off-by: Erik Boasson <eb@ilities.com>
The Xcode 9 builds often fail because of some nasty interaction between
it and the Travis build cache -- but not caching means they always take
forever because of homebrew updates. This commit changes the build to
use a new Xcode but with a deployment target of 10.12 instead.
This means that macOS version macros for versions later than 10.12 are
available and that the tests run on a later version of macOS as well.
Therefore, it doesn't prove that a "real" Xcode 9 / macOS 10.12 build
will work.
ROS2 relies on macOS 10.12 so this introduces a (small) risk of build
failures over there. To mitigate that, a pure Xcode 9 build is included
in the cron job.
Signed-off-by: Erik Boasson <eb@ilities.com>
Sign extension causes some character values to be interpreted as special
tokens, potentially causing the parser to hang, pass a null pointer as
element name to the "elem_open" callback (which would crash the
configuration processing).
Signed-off-by: Erik Boasson <eb@ilities.com>
When scanning buckets while looking for a specific element, it is faster
to only inspect those buckets for which the corresponding bit in
"hopinfo" is set. Sadly, simple scanning is faster than the far more
elegant perfect hash based on a De Bruijn sequence.
Add a simple test/performance measurement.
Signed-off-by: Erik Boasson <eb@ilities.com>
* Moved bswap functions to ddsrt
Moved the byte swapping functions from ddsi to ddsrt so that
these can be re-used in e.g. the security plugins and tests.
Signed-off-by: Dennis Potman <dennis.potman@adlinktech.com>
* Moved decarations for bswap functions to ddsrt
Signed-off-by: Dennis Potman <dennis.potman@adlinktech.com>
This will eliminate two large dependencies when building CycloneDDS with
colcon for ROS 2, which doesn't appear to need the IDLC.
Signed-off-by: Scott K Logan <logans@cottsay.net>
This is useful when building CycloneDDS on platforms that provide CMake
configuration for CUnit outside of Conan.
If no configuration is found in the system CMake registry the module
continues as before.
Co-authored-by: Steven! Ragnarök <steven@nuclearsandwich.com>
Signed-off-by: Scott K Logan <logans@cottsay.net>
These conditionals may fail if the variable they're checking isn't
defined at all. Adding quotes makes the comparison against an empty
string in this case, which avoids the syntax error.
Signed-off-by: Scott K Logan <logans@cottsay.net>
CMAKE_PROJECT_NAME refers to the top-level project name, not the most recent project. So any CMake project that pulls this in as a dependency was in for a nasty surprise.
https://cmake.org/cmake/help/latest/variable/CMAKE_PROJECT_NAME.html
Signed-off-by: Dan Rose <dan@digilabs.io>
