[df] Consider jitted nodes in Snapshot with variations #20352
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Test Results 21 files 21 suites 3d 16h 4m 10s ⏱️ Results for commit 42a43ad. ♻️ This comment has been updated with latest results. |
hageboeck
approved these changes
Nov 10, 2025
Member
hageboeck
left a comment
hageboeck
left a comment
There was a problem hiding this comment.
Nice, thanks!
I had started condensing a reproducer, too, but this change fixed my test. 🙂
This commit introduces two enhancements to the Snapshot with variations implementation: - Snapshot with variations must call RLoopManager::Jit before trying to request varied filters from upstream nodes. This is necessary because the upstream node could be a jitted node itself. - Snapshot with variations must deal with the possibility of having a different nominal upstream node type than the other varied upstream node types. This aligns the implementation witht the one of RVariedAction for what concerns dealing with jitted nodes. Fixes root-project#20320
hageboeck
added a commit
to hageboeck/root
that referenced
this pull request
Nov 24, 2025
This is a followup of the fix root-project#20352 for issue root-project#20320: When snapshot with variations is run with JIT-ted filters, the computation graph first needs to be JIT-ted before reading Defines or Filters. In the fix for that case, an iterator invalidation on push_back caused crashes, because vector.reserve() had been called with the wrong size. Here, the size is corrected, but more importantly, instead of saving the iterator, the shared_ptr it points to is saved. This makes the code independent of the underlying storage.
hageboeck
added a commit
that referenced
this pull request
Nov 25, 2025
This is a followup of the fix #20352 for issue #20320: When snapshot with variations is run with JIT-ted filters, the computation graph first needs to be JIT-ted before reading Defines or Filters. In the fix for that case, an iterator invalidation on push_back caused crashes, because vector.reserve() had been called with the wrong size. Here, the size is corrected, but more importantly, instead of saving the iterator, the shared_ptr it points to is saved. This makes the code independent of the underlying storage.
hageboeck
added a commit
to hageboeck/root
that referenced
this pull request
Nov 25, 2025
This is a followup of the fix root-project#20352 for issue root-project#20320: When snapshot with variations is run with JIT-ted filters, the computation graph first needs to be JIT-ted before reading Defines or Filters. In the fix for that case, an iterator invalidation on push_back caused crashes, because vector.reserve() had been called with the wrong size. Here, the size is corrected, but more importantly, instead of saving the iterator, the shared_ptr it points to is saved. This makes the code independent of the underlying storage. (cherry picked from commit 886abae)
hageboeck
added a commit
to hageboeck/root
that referenced
this pull request
Nov 25, 2025
This is a followup of the fix root-project#20352 for issue root-project#20320: When snapshot with variations is run with JIT-ted filters, the computation graph first needs to be JIT-ted before reading Defines or Filters. In the fix for that case, an iterator invalidation on push_back caused crashes, because vector.reserve() had been called with the wrong size. Here, the size is corrected, but more importantly, instead of saving the iterator, the shared_ptr it points to is saved. This makes the code independent of the underlying storage. (cherry picked from commit 886abae)
dpiparo
pushed a commit
that referenced
this pull request
Nov 25, 2025
This is a followup of the fix #20352 for issue #20320: When snapshot with variations is run with JIT-ted filters, the computation graph first needs to be JIT-ted before reading Defines or Filters. In the fix for that case, an iterator invalidation on push_back caused crashes, because vector.reserve() had been called with the wrong size. Here, the size is corrected, but more importantly, instead of saving the iterator, the shared_ptr it points to is saved. This makes the code independent of the underlying storage. (cherry picked from commit 886abae)
KoljaFrahm
pushed a commit
to KoljaFrahm/root
that referenced
this pull request
Nov 29, 2025
This is a followup of the fix root-project#20352 for issue root-project#20320: When snapshot with variations is run with JIT-ted filters, the computation graph first needs to be JIT-ted before reading Defines or Filters. In the fix for that case, an iterator invalidation on push_back caused crashes, because vector.reserve() had been called with the wrong size. Here, the size is corrected, but more importantly, instead of saving the iterator, the shared_ptr it points to is saved. This makes the code independent of the underlying storage.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This commit introduces two enhancements to the Snapshot with variations implementation:
This aligns the implementation witht the one of RVariedAction for what concerns dealing with jitted nodes.
Fixes #20320