Make roslaunch-check respect if/unless attribute on <include>#998
Make roslaunch-check respect if/unless attribute on <include>#998dirk-thomas merged 1 commit intolunar-develfrom
Conversation
|
@dirk-thomas Getting a dependency failure here on the Lunar buildfarm: I'm assuming we should be targeting Lunar with new changes now, and considering Kinetic for backport? Oh blah, I see that there are a number of broken/missing packages at the moment, including rosbuild: http://repositories.ros.org/status_page/ros_lunar_default.html?q=rosbuild |
| elif tag.attributes.has_key('unless'): | ||
| val = resolve_args(tag.attributes['unless'].value, context) | ||
| if convert_value(val, 'bool'): | ||
| continue |
There was a problem hiding this comment.
Another possibility would be moving this logic up a level so that it would also do the right thing on <group> and <node>— not currently the case.
There was a problem hiding this comment.
Is this block still necessary with _check_ifunless being called at the beginning of the loop?
There was a problem hiding this comment.
Oh— no it isn't, but I forgot to remove it. Fixed now.
Looks like it's still rospack 2.4.0 in the shadow repo, so this isn't going to pass yet.
There was a problem hiding this comment.
rospack 2.4.1 has propagated to shadow-fixed already: http://repositories.ros.org/status_page/ros_lunar_default.html?q=rospack
Other jobs have turned green.
781aea4 to
7ffd338
Compare
|
@dirk-thomas I think this is good to go, the failed tests appear to be unrelated to this change. |
|
The failing tests might be related to ros/rospack#70. I will retrigger the PR jobs once the new version of rospack has been synced to the testing repo. |
|
@ros-pull-request-builder retest this please |
|
Most recent test got rospack 2.4.1 and ran clean. |
| elif tag.attributes.has_key('unless'): | ||
| if not convert_value(val, 'bool'): | ||
| return False | ||
| if tag.attributes.has_key('unless'): |
There was a problem hiding this comment.
Previously if a tag had both - an if as well as an unless attribute the second one wasn't considered. Is there a reason to change this?
There was a problem hiding this comment.
Oh I suppose not; I'll change it back.
0efa6ca to
83fb7c9
Compare
|
@ros/ros_comm-maintainers This should be good to go in. Anyone got thoughts? |
|
Not a maintainer, but changes look good on my end. Is this a candidate for backporting into Kinetic? |
|
@mikepurvis Looks good to me 👍 @IanTheEngineer First it will be released into Lunar to give it some soak time. It can be backported if there is a strong need for it, the chance for regressions are low, and it doesn't change API. |
…comm#953 could not load launch file with args directory
…_comm#953 could not load launch file with args directory
As discussed in #993. With the fix removed, the updated test XML fails with: