Conversation
cppcheck 2.0 was released to chocolatey yesterday and is causing major havoc in our Windows builds. ros2/ros2#942 is open to track making that work but this will reduce the failure count on Windows so it's easier to triage and manage builds in the interim.
|
Build up to rosidl_typesupport_cpp since it exhibits the errors that pinning is intended to resolve. Windows |
| RUN choco install -y cmake curl git vcredist2013 vcredist140 cppcheck patch | ||
| RUN choco install -y cmake curl git vcredist2013 vcredist140 patch | ||
| # Pin cppcheck until our builds are compatible with 2.0 https://github.com/ros2/ros2/issues/942 | ||
| RUN choco install -y cppcheck --version 1.90 |
There was a problem hiding this comment.
Line 112 also needs this. Similar to the rigmarole we had to go through for 1.89->1.9. I'll commit this after dinner if you're not around.
There was a problem hiding this comment.
Would it make more sense to pin cppcheck. I'll propose something.
There was a problem hiding this comment.
What do you think of 26da4cb ? I like pinning because it keeps both lines local rather than needing to split attention?
There was a problem hiding this comment.
Pinning works for me, definitely makes things simpler.
| RUN choco install -y cmake curl git vcredist2013 vcredist140 patch | ||
| # Pin cppcheck until our builds are compatible with 2.0 https://github.com/ros2/ros2/issues/942 | ||
| RUN choco install -y cppcheck --version 1.90 | ||
| RUN choco pin add --name=cppcheck --version 1.90 |
There was a problem hiding this comment.
For whatever reason last time I pinned with chocolatey neither --name cppcheck nor --version=1.90 worked so this command uses two different argument forms.
There was a problem hiding this comment.
Does choco pin by itself just apply to all packages?
There was a problem hiding this comment.
Does
choco pinby itself just apply to all packages?
Nope. It lists existing pins. The --name argument is required for choco pin add https://chocolatey.org/docs/commands-pin
|
Testing the pin: |
brawner
left a comment
There was a problem hiding this comment.
Looks good to me. Was pinning actually necessary? Why did the first build succeed?
|
Err, it did install 2.0 during the upgrade, so I'm not entirely sure what's happening. |
|
Building up to std_msgs is enough to trigger the cppcheck issues on mainline. So if |
This reverts commit 42a0519.
This reverts #472 Signed-off-by: Jacob Perron <jacob@openrobotics.org>
This reverts #472 Signed-off-by: Jacob Perron <jacob@openrobotics.org>
This reverts #472 Signed-off-by: Jacob Perron <jacob@openrobotics.org>
This reverts #472 Signed-off-by: Jacob Perron <jacob@openrobotics.org>
cppcheck 2.0 was released to chocolatey yesterday and is causing major
havoc in our Windows builds. ros2/ros2#942 is
open to track making that work but this will reduce the failure count on
Windows so it's easier to triage and manage builds in the interim.