Stage shadow tables with their virtual table through named add#1650
Merged
Conversation
Virtual tables carry no catalog entry of their own — their commit
content IS the shadow tables — so the entry-based named-add flow
silently ignored them: dolt_add('ft') returned success while staging
nothing, and committed fts/rtree content was whatever the last -Am
commit happened to capture. Staged drops of a vtab errored the same
silent way, leaving the vtab and its shadows in the commit.
Named add of a virtual table now stages every working entry that
sqlite3IsShadowTableOf identifies as its shadow (with their indexes)
and adopts the working master so the vtab's schema row travels;
staging a dropped vtab resolves through the staged catalog's schema
rows and removes the shadow entries with it. Add-time snapshot
semantics hold: rows written after dolt_add stay out of the commit.
The index x VC matrix grows vtab scenarios: named add, -am, staged
drop, old-branch checkout, and clone, over fts5 and rtree.
Fixes #1647
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Version-Control Performance CeilingsRuns: median of 3 executions per benchmark, excluding fixture setup. The
|
Sysbench-Style Benchmark: Doltlite vs SQLiteIn-MemoryReads
Writes
File-BackedReads
Writes
File-Backed (autocommit)Each statement runs as its own transaction — exposes per-commit ReadsReads have no commit cost; these are the same SQL files as the
Writes
100000 rows, median of 5 invocations per test; autocommit writes use 9, workload-only timing via host monotonic clock when available. Performance Ceiling Check (2.5x individual, 2x average; autocommit writes: 6x / 5x)All tests within ceilings. |
Sysbench-Style Benchmark (BLOB PK): Doltlite vs SQLiteCompanion to the classic Sysbench-Style Benchmark. Every workload here In-MemoryReads
Writes
File-BackedReads
Writes
File-Backed (autocommit)Each statement runs as its own transaction — exposes per-commit ReadsReads have no commit cost; these are the same SQL files as the
Writes
100000 rows, median of 5 invocations per test; autocommit writes use 9, workload-only timing via host monotonic clock when available. Performance Ceiling Check (2.5x individual, 2x average; autocommit writes: 6x / 5x)All tests within ceilings. |
Sysbench-Style Benchmark (TEXT PK): Doltlite vs SQLiteCompanion to the classic Sysbench-Style Benchmark. Every workload here In-MemoryReads
Writes
File-BackedReads
Writes
File-Backed (autocommit)Each statement runs as its own transaction — exposes per-commit ReadsReads have no commit cost; these are the same SQL files as the
Writes
100000 rows, median of 5 invocations per test; autocommit writes use 9, workload-only timing via host monotonic clock when available. Performance Ceiling Check (2.5x individual, 2x average; autocommit writes: 6x / 5x)All tests within ceilings. |
Sysbench-Style Benchmark (composite PK): Doltlite vs SQLiteCompanion to the classic Sysbench-Style Benchmark. Every workload here In-MemoryReads
Writes
File-BackedReads
Writes
File-Backed (autocommit)Each statement runs as its own transaction — exposes per-commit ReadsReads have no commit cost; these are the same SQL files as the
Writes
100000 rows, median of 5 invocations per test; autocommit writes use 9, workload-only timing via host monotonic clock when available. Performance Ceiling Check (2.5x individual, 2x average; autocommit writes: 6x / 5x)All tests within ceilings. |
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
1 participant
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.
Fixes #1647.
The bug ran deeper than filed: virtual tables have no catalog entry at all (schema row only), so the entry-based named-add flow silently ignored them entirely —
dolt_add('ft')returned success while staging nothing, and staged drops of a vtab failed just as silently, leaving the vtab and shadows in the commit. Committed fts/rtree content was whatever the last-Amcommit happened to capture.The fix, in the named-add path (same functions as #1643's index staging):
sqlite3IsShadowTableOfidentifies as its shadow (plus their indexes), and adopts the working master so the vtab's schema row travels.dolt_addstay out of the commit (verified).The index×VC matrix grows five vtab scenarios — named add,
-am, staged drop, old-branch checkout, and clone, over both fts5 and rtree. Fail-before/pass-after: pre-fix binary fails the three staging scenarios (35/38), fixed binary passes 38/38.Validation: matrix 38/38, shell battery 88/88, C regression 309,830/309,830 (
DOLTLITE_PROLLY_CHECK=1).Remaining sibling in the class: #1648 (view/trigger ride-along on master adoption) stays filed with analysis — its fix is a staging-model design change, not a patch.
🤖 Generated with Claude Code