Closed gavanderhoorn closed 2 years ago
This was never really a problem until we upgraded rosidl_generator_c
et al. to post https://github.com/ros2/rosidl/pull/593.
I'm torn on this one.
On the one hand, there is no reason that this needs to be Unicode; regular ASCII apostrophe will be just fine, and will fix your use case.
On the other hand, we most definitely should support Unicode inside of comments in .msg
files. This was a test of that, though more of an implicit one rather than an explicit one.
I'm inclined to take this PR as-is, and deal with testing Unicode in comments in .msg
files separately, but I'm not sure. @ros2/team thoughts?
On the other hand, we most definitely should support Unicode inside of comments in
.msg
files. This was a test of that, though more of an implicit one rather than an explicit one.
Didn't https://github.com/ros2/rcl_interfaces/pull/80 already add those?
Edit: oh, hm, you wrote "in comments". #80 does not test comments.
My vote would obviously be: implicit tests are no tests, and fix those missing tests in separate PR(s) ;)
My vote would be opening an issue on https://github.com/ros2/rosidl/issues about adding tests for non-ascii characters in comments in message files.
That works for me. @gavanderhoorn Would you mind opening up that issue? Then we can run CI on this.
Thanks for merging.
Would this be back-portable to Humble as well?
Yeah, that seems reasonable.
@Mergifyio backport humble
@Mergifyio backport humble
backport humble
As per subject.
I'm working on an old platform with an ancient version of GCC which isn't capable of processing files with non-ASCII chars in them.
The unicode apostrophes caused build failures, so 48d66fbb removes them.