Conversation
rcl_yaml_param_parser/src/parser.c
Outdated
| } | ||
| } else { | ||
| #pragma GCC diagnostic push | ||
| #pragma GCC diagnostic ignored "-Wmaybe-uninitialized" |
There was a problem hiding this comment.
This needs to be conditionally done, as it will not work on Windows. For example: https://github.com/ros2/rviz/blob/f594d9f00ecf87f75a0a3909f538bd463f7b7375/rviz_rendering/include/rviz_rendering/objects/point_cloud.hpp#L40-L63
We target only platforms using gcc 5.4 at the moment.
Actually it seem to show up only on the last comparison of Regarding the patch, this should be applied only if _WIN32 is not defined like we do in other places CI without and with this patch: Without: |
Ah. The console output of the aarch64 job has the line |
ec55617 to
e4d68e4
Compare
|
Thanks for the build type @mikaelarguedas. Fixed warning rather than suppressed it in e4d68e4 New CI (build up to and test only
|
|
thanks @sloretz, this is a better resolution indeed 👍 Can you please run CI using the |
|
Ooops. CI RelWithDebInfo
|
This suppresses an occurrance of a
maybe-uninitializedwarning on gcc 4.x. I don't see a warning using gcc 5.4It seems to be a false positive. Searching for false positive
maybe-uninitializedgcc bugs returns too many to reasonably search for a match. It's strange to me that the warning is not also present on lines 941, 921, 883, etc.In progress while CI runs