Change PCL smart pointers from boost to std#3750
Change PCL smart pointers from boost to std#3750kunaltyagi merged 2 commits intoPointCloudLibrary:masterfrom
boost to std#3750Conversation
|
Windows error (link): |
|
I might be completely off, but from what I gather from here it's not so much an issue coming from the switch, but more like coming from a missing There is only one pcl/filters/src/radius_outlier_removal.cpp Line 167 in 20a289b If that's the case, their suggestion is pretty straightforward though: add at line 127 Lines 125 to 130 in 20a289b Thoughts? |
|
You can see I edited the windows error message because I realized it's a much longer error. It is of course picking up on the |
kunaltyagi
left a comment
There was a problem hiding this comment.
Maybe split this PR in 2 parts. One for the shift from boost::shared_ptr to std::shared_ptr and another for boost::->pcl::. The second one can be merged faster, maybe in time for 1.10.1 and the first one afterwards.
registration/include/pcl/registration/correspondence_rejection_distance.h
Outdated
Show resolved
Hide resolved
|
There is also a single boost::weak_ptr -> std::weak_ptr to decide upon. I can see the whole codebase has 3 weak ptrs though, and two of them are already std somehow. I'm rebasing and regrouping the commits to ease my work afterwards. |
They probably were not a part of the public interface and got swapped :) |
|
They are not, but neither is this one. Separate PR, I guess. |
|
Using the old definition of the label |
This is false. It's part of the public API because it's locked and returned pcl/io/src/openni_camera/openni_driver.cpp Lines 254 to 291 in e6e7377 |
|
What's the status of this one? |
|
Last night I was working a bit on the |
|
Also I'm still waiting on |
|
Sorry for that. My opinion is to try and go unqualified. |
|
Ouch! So the flag isn't correctly propagated AND it was completely fine with boost pointers! |
|
Maintainer note: needs rebase post #3770 merge. |
|
And also post #3753. We have a test failure on Windows x86: |
|
Flaky test. Rerunning it |
207eab8 to
540f673
Compare
|
Some changes seem to be duplicated across the PRs |
|
This one checks if everything works after the whole transition is complete, the extracted PRs check the single changes? I can cancel the build if CI time is precious. I'll remove the commits as soon as the other PRs are merged. EDIT: and this PR has the smaller |
|
No need to cancel CI. I meant that some parts of the changeset were duplicated in other PR too, like editing the include files, etc. Maybe add a preferred sequence to the PRs submitted, so we can prioritize the ones that need to go first |
|
This won't compile until #3753 is merged. Should be the last remaining dependency PR-wise. EDIT: it'll also show unrelated changes until a rebase after that PR is merged |
|
Time to rebase and test 😄 |
boost to std
|
Flagging it an API break. Just because we provided a Good to merge from my side once CI is done. |
Hoping to close #3735
Also closes #2792