Repository navigation
fix(semantic): number moved lines after the interned ids - #58
Merged
Merged
Conversation
fill gave each moved line a fresh id counting down from u32::MAX, so the line diff sized its per-id tables by that id: two 4 GiB arrays for every region with a moved line, and three 16 GiB ones more when the search ran over budget. They only survived through lazy zeroed pages. Moved lines now get their ids after interning, counting up from the number of interned lines, so the ids stay below the region's line count. Equal ids are still exactly the equal unmoved lines, so the alignment is unchanged.
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Every semantic diff region containing a moved line reserved about 8 GiB of memory, and about 48 GiB more when the line diff ran over budget, because moved lines got ids counting down from
u32::MAXand the line diff sizes its per-id tables by the largest id. Moved lines are now numbered after the interned lines, so ids stay below the region's line count and the alignment comes out the same.