rubydb 0.1.4 → 0.1.6

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (469) hide show
  1. checksums.yaml +4 -4
  2. data/.github/PULL_REQUEST_TEMPLATE.md +15 -15
  3. data/.github/workflows/benchmark.yml +26 -26
  4. data/.github/workflows/compatibility.yml +63 -63
  5. data/.github/workflows/fuzz.yml +33 -33
  6. data/.github/workflows/lint.yml +21 -21
  7. data/.github/workflows/operations.yml +24 -24
  8. data/.github/workflows/production-validation.yml +111 -111
  9. data/.github/workflows/release.yml +77 -77
  10. data/.github/workflows/security.yml +39 -37
  11. data/.github/workflows/test.yml +26 -26
  12. data/.github/workflows/workload.yml +58 -58
  13. data/.gitignore +16 -5
  14. data/.rubocop.yml +50 -44
  15. data/.standard.yml +9 -14
  16. data/ARCHITECTURE.md +21 -21
  17. data/CHANGELOG.md +57 -27
  18. data/CODE_OF_CONDUCT.md +13 -13
  19. data/CONTRIBUTING.md +29 -29
  20. data/GOVERNANCE.md +16 -16
  21. data/Gemfile +18 -17
  22. data/Gemfile.lock +125 -71
  23. data/README.md +168 -12
  24. data/ROADMAP.md +27 -27
  25. data/Rakefile +76 -71
  26. data/SECURITY.md +54 -54
  27. data/SUPPORT.md +14 -14
  28. data/accelerator/bin/SHA256SUMS +6 -0
  29. data/accelerator/bin/rubydb-accelerator-darwin-amd64 +0 -0
  30. data/accelerator/bin/rubydb-accelerator-darwin-arm64 +0 -0
  31. data/accelerator/bin/rubydb-accelerator-linux-amd64 +0 -0
  32. data/accelerator/bin/rubydb-accelerator-linux-arm64 +0 -0
  33. data/accelerator/bin/rubydb-accelerator-windows-amd64.exe +0 -0
  34. data/accelerator/bin/rubydb-accelerator-windows-arm64.exe +0 -0
  35. data/accelerator/cmd/rubydb-accelerator/main.go +11 -0
  36. data/accelerator/go.mod +3 -0
  37. data/accelerator/internal/execution/aggregate.go +94 -0
  38. data/accelerator/internal/execution/distinct.go +22 -0
  39. data/accelerator/internal/execution/filter.go +73 -0
  40. data/accelerator/internal/execution/join.go +79 -0
  41. data/accelerator/internal/execution/operators.go +167 -0
  42. data/accelerator/internal/execution/scan.go +20 -0
  43. data/accelerator/internal/execution/sort.go +62 -0
  44. data/accelerator/internal/execution/types.go +136 -0
  45. data/accelerator/internal/execution/value.go +67 -0
  46. data/accelerator/internal/memory/arena.go +47 -0
  47. data/accelerator/internal/memory/reuse.go +22 -0
  48. data/accelerator/internal/metrics/registry.go +67 -0
  49. data/accelerator/internal/parallel/bounded_queue.go +56 -0
  50. data/accelerator/internal/parallel/scheduler.go +47 -0
  51. data/accelerator/internal/parallel/worker_pool.go +53 -0
  52. data/accelerator/internal/protocol/cancellation.go +48 -0
  53. data/accelerator/internal/protocol/columnar.go +263 -0
  54. data/accelerator/internal/protocol/frame.go +187 -0
  55. data/accelerator/internal/runtime/worker.go +521 -0
  56. data/accelerator/internal/storage/page_reader.go +81 -0
  57. data/accelerator/internal/storage/snapshot_scan.go +539 -0
  58. data/accelerator/internal/wal/checksum.go +13 -0
  59. data/accelerator/internal/wal/compression.go +41 -0
  60. data/accelerator/internal/wal/group_commit.go +24 -0
  61. data/accelerator/internal/wal/record_encoder.go +40 -0
  62. data/adapters/activerecord/Gemfile +11 -11
  63. data/adapters/activerecord/README.md +8 -3
  64. data/adapters/activerecord/lib/active_record/connection_adapters/rubydb_adapter.rb +881 -879
  65. data/adapters/activerecord/rubydb-activerecord.gemspec +21 -21
  66. data/adapters/activerecord/spec/rubydb_adapter_integration_spec.rb +143 -143
  67. data/adapters/ruby/README.md +18 -18
  68. data/adapters/sequel/README.md +11 -11
  69. data/config/monitoring/prometheus-alerts.yml +39 -39
  70. data/config/production.yml +36 -36
  71. data/docs/README.md +77 -71
  72. data/docs/architecture/concurrency.md +14 -14
  73. data/docs/architecture/current-state.md +125 -125
  74. data/docs/architecture/execution-engine.md +25 -25
  75. data/docs/architecture/go-accelerator.md +179 -0
  76. data/docs/architecture/indexes.md +19 -19
  77. data/docs/architecture/mvcc.md +19 -19
  78. data/docs/architecture/overview.md +13 -13
  79. data/docs/architecture/pages.md +11 -11
  80. data/docs/architecture/production-roadmap.md +82 -82
  81. data/docs/architecture/query-planner.md +20 -20
  82. data/docs/architecture/recovery.md +18 -18
  83. data/docs/architecture/sql-engine.md +12 -12
  84. data/docs/architecture/storage-engine.md +14 -14
  85. data/docs/architecture/transactions.md +10 -10
  86. data/docs/architecture/wal.md +28 -28
  87. data/docs/cli-cheatsheet.md +98 -98
  88. data/docs/cli.md +299 -275
  89. data/docs/contributing/architecture.md +9 -9
  90. data/docs/contributing/benchmarking.md +30 -14
  91. data/docs/contributing/development.md +16 -16
  92. data/docs/contributing/release-process.md +49 -49
  93. data/docs/contributing/testing.md +16 -16
  94. data/docs/debugging.md +229 -229
  95. data/docs/developer/branching.md +10 -10
  96. data/docs/developer/database-diff.md +10 -10
  97. data/docs/developer/local-development.md +49 -17
  98. data/docs/developer/snapshots.md +9 -9
  99. data/docs/developer/temporal-data.md +10 -10
  100. data/docs/developer-guide.md +297 -297
  101. data/docs/getting-started/first-database.md +16 -16
  102. data/docs/getting-started/first-query.md +13 -13
  103. data/docs/getting-started/installation.md +19 -19
  104. data/docs/getting-started/local-to-production.md +300 -300
  105. data/docs/getting-started/quickstart.md +17 -17
  106. data/docs/getting-started/rails.md +16 -16
  107. data/docs/hardening_backlog.md +93 -93
  108. data/docs/lessons-learned.md +112 -112
  109. data/docs/operations/backups.md +33 -33
  110. data/docs/operations/disaster-recovery.md +31 -31
  111. data/docs/operations/failover.md +30 -30
  112. data/docs/operations/monitoring.md +25 -25
  113. data/docs/operations/production-guide.md +295 -295
  114. data/docs/operations/production-runbook.md +45 -45
  115. data/docs/operations/replication.md +33 -33
  116. data/docs/operations/restore.md +6 -6
  117. data/docs/operations/runbook.md +34 -34
  118. data/docs/operations/upgrades.md +14 -14
  119. data/docs/operations/workload-testing.md +17 -17
  120. data/docs/production-readiness.md +118 -118
  121. data/docs/production_validation.md +150 -150
  122. data/docs/rails/active-record.md +11 -11
  123. data/docs/rails/compatibility-guide.md +90 -90
  124. data/docs/rails/database-yml.md +92 -92
  125. data/docs/rails/installation.md +17 -17
  126. data/docs/rails/migrations.md +17 -17
  127. data/docs/rails/production.md +82 -82
  128. data/docs/rails/troubleshooting.md +18 -18
  129. data/docs/release.md +25 -25
  130. data/docs/server/architecture.md +10 -10
  131. data/docs/server/authentication.md +10 -10
  132. data/docs/server/configuration.md +16 -16
  133. data/docs/server/connection-pooling.md +10 -10
  134. data/docs/server/deployment.md +10 -10
  135. data/docs/server/protocol.md +12 -12
  136. data/docs/sql/compatibility-guide.md +82 -82
  137. data/docs/sql/compatibility.md +39 -39
  138. data/docs/sql/data-types.md +10 -10
  139. data/docs/sql/functions.md +9 -9
  140. data/docs/sql/joins.md +9 -9
  141. data/docs/sql/operators.md +9 -9
  142. data/docs/sql/sqlite-compatibility.md +21 -21
  143. data/docs/sql/syntax.md +10 -10
  144. data/docs/sql/transactions.md +10 -10
  145. data/docs/troubleshooting.md +244 -244
  146. data/lessons/01-foundations.md +73 -0
  147. data/lessons/02-local-development.md +121 -0
  148. data/lessons/03-embedded-rubydb.md +99 -0
  149. data/lessons/04-rails-complex-apps.md +138 -0
  150. data/lessons/05-rubydb-production-server.md +237 -0
  151. data/lessons/06-postgresql-massive-apps.md +96 -0
  152. data/lessons/07-hybrid-microservices.md +179 -0
  153. data/lessons/08-migrations-backups-recovery.md +86 -0
  154. data/lessons/09-observability-security-scale.md +87 -0
  155. data/lessons/10-release-readiness.md +192 -0
  156. data/lessons/11-community-adapter.md +323 -0
  157. data/lessons/12-rails-ecommerce-pressure.md +263 -0
  158. data/lib/rubydb/accelerator/client.rb +451 -0
  159. data/lib/rubydb/accelerator/error.rb +22 -0
  160. data/lib/rubydb/accelerator/manager.rb +606 -0
  161. data/lib/rubydb/accelerator.rb +13 -0
  162. data/lib/rubydb/backup/archive.rb +332 -334
  163. data/lib/rubydb/backup/backup.rb +400 -401
  164. data/lib/rubydb/backup/incremental.rb +349 -353
  165. data/lib/rubydb/backup/restore.rb +289 -290
  166. data/lib/rubydb/backup/snapshot.rb +265 -267
  167. data/lib/rubydb/backup/verification.rb +276 -279
  168. data/lib/rubydb/branching/branch.rb +181 -181
  169. data/lib/rubydb/branching/branch_manager.rb +307 -311
  170. data/lib/rubydb/branching/branch_metadata.rb +140 -140
  171. data/lib/rubydb/branching/checkout.rb +165 -166
  172. data/lib/rubydb/branching/copy_on_write.rb +272 -272
  173. data/lib/rubydb/branching/diff.rb +137 -138
  174. data/lib/rubydb/branching/merge.rb +282 -285
  175. data/lib/rubydb/build_info.rb +15 -15
  176. data/lib/rubydb/catalog/catalog.rb +391 -391
  177. data/lib/rubydb/catalog/column.rb +112 -112
  178. data/lib/rubydb/catalog/constraint.rb +180 -180
  179. data/lib/rubydb/catalog/database.rb +184 -184
  180. data/lib/rubydb/catalog/index.rb +97 -97
  181. data/lib/rubydb/catalog/schema.rb +103 -103
  182. data/lib/rubydb/catalog/sequence.rb +90 -90
  183. data/lib/rubydb/catalog/system_catalog.rb +698 -698
  184. data/lib/rubydb/catalog/table.rb +178 -178
  185. data/lib/rubydb/catalog/trigger.rb +102 -102
  186. data/lib/rubydb/catalog/view.rb +66 -66
  187. data/lib/rubydb/cli/application.rb +168 -163
  188. data/lib/rubydb/cli/commands/accelerator.rb +72 -0
  189. data/lib/rubydb/cli/commands/backup.rb +80 -81
  190. data/lib/rubydb/cli/commands/branch.rb +72 -72
  191. data/lib/rubydb/cli/commands/checkout.rb +54 -54
  192. data/lib/rubydb/cli/commands/create.rb +58 -58
  193. data/lib/rubydb/cli/commands/diff.rb +76 -77
  194. data/lib/rubydb/cli/commands/doctor.rb +77 -74
  195. data/lib/rubydb/cli/commands/drop.rb +57 -57
  196. data/lib/rubydb/cli/commands/init.rb +101 -102
  197. data/lib/rubydb/cli/commands/inspect.rb +95 -95
  198. data/lib/rubydb/cli/commands/merge.rb +63 -63
  199. data/lib/rubydb/cli/commands/migrate.rb +62 -62
  200. data/lib/rubydb/cli/commands/restart.rb +42 -39
  201. data/lib/rubydb/cli/commands/restore.rb +121 -121
  202. data/lib/rubydb/cli/commands/shell.rb +365 -365
  203. data/lib/rubydb/cli/commands/snapshot.rb +79 -79
  204. data/lib/rubydb/cli/commands/start.rb +88 -82
  205. data/lib/rubydb/cli/commands/status.rb +96 -92
  206. data/lib/rubydb/cli/commands/stop.rb +47 -47
  207. data/lib/rubydb/cli/commands/vacuum.rb +58 -58
  208. data/lib/rubydb/cli/formatter.rb +221 -221
  209. data/lib/rubydb/cli/output.rb +168 -168
  210. data/lib/rubydb/client/client.rb +309 -304
  211. data/lib/rubydb/client/connection.rb +429 -415
  212. data/lib/rubydb/client/connection_pool.rb +168 -168
  213. data/lib/rubydb/client/connection_url.rb +96 -96
  214. data/lib/rubydb/client/prepared_statement.rb +60 -60
  215. data/lib/rubydb/client/result.rb +127 -123
  216. data/lib/rubydb/client/statement.rb +52 -52
  217. data/lib/rubydb/client/transaction.rb +130 -130
  218. data/lib/rubydb/concurrency/concurrency.rb +19 -19
  219. data/lib/rubydb/concurrency/deadlock_detector.rb +148 -150
  220. data/lib/rubydb/concurrency/latch.rb +101 -101
  221. data/lib/rubydb/concurrency/lock_graph.rb +163 -165
  222. data/lib/rubydb/concurrency/mutex.rb +181 -183
  223. data/lib/rubydb/concurrency/rw_lock.rb +180 -180
  224. data/lib/rubydb/concurrency/scheduler.rb +248 -250
  225. data/lib/rubydb/concurrency/worker_pool.rb +145 -143
  226. data/lib/rubydb/configuration/config.rb +170 -170
  227. data/lib/rubydb/configuration/defaults.rb +191 -179
  228. data/lib/rubydb/configuration/environment.rb +152 -152
  229. data/lib/rubydb/configuration/parser.rb +185 -185
  230. data/lib/rubydb/configuration/validation.rb +228 -221
  231. data/lib/rubydb/constants.rb +74 -74
  232. data/lib/rubydb/constraints/check.rb +181 -181
  233. data/lib/rubydb/constraints/constraint.rb +101 -101
  234. data/lib/rubydb/constraints/foreign_key.rb +130 -130
  235. data/lib/rubydb/constraints/not_null.rb +64 -64
  236. data/lib/rubydb/constraints/primary_key.rb +99 -99
  237. data/lib/rubydb/constraints/unique.rb +106 -108
  238. data/lib/rubydb/constraints/validator.rb +349 -350
  239. data/lib/rubydb/errors/authentication_error.rb +10 -10
  240. data/lib/rubydb/errors/authorization_error.rb +23 -23
  241. data/lib/rubydb/errors/client_error.rb +10 -10
  242. data/lib/rubydb/errors/configuration_error.rb +10 -10
  243. data/lib/rubydb/errors/connection_error.rb +10 -10
  244. data/lib/rubydb/errors/constraint_error.rb +23 -23
  245. data/lib/rubydb/errors/corruption_error.rb +10 -10
  246. data/lib/rubydb/errors/database_error.rb +10 -10
  247. data/lib/rubydb/errors/error.rb +20 -20
  248. data/lib/rubydb/errors/execution_error.rb +10 -10
  249. data/lib/rubydb/errors/parser_error.rb +10 -10
  250. data/lib/rubydb/errors/recovery_error.rb +10 -10
  251. data/lib/rubydb/errors/replication_error.rb +10 -10
  252. data/lib/rubydb/errors/server_error.rb +6 -6
  253. data/lib/rubydb/errors/storage_error.rb +10 -10
  254. data/lib/rubydb/errors/transaction_error.rb +10 -10
  255. data/lib/rubydb/execution/accelerator_dispatch.rb +30 -0
  256. data/lib/rubydb/execution/aggregate_executor.rb +134 -138
  257. data/lib/rubydb/execution/cost_model.rb +72 -0
  258. data/lib/rubydb/execution/delete_executor.rb +110 -112
  259. data/lib/rubydb/execution/distinct_executor.rb +131 -135
  260. data/lib/rubydb/execution/executor.rb +1544 -1188
  261. data/lib/rubydb/execution/expression.rb +191 -193
  262. data/lib/rubydb/execution/index_scan.rb +142 -142
  263. data/lib/rubydb/execution/insert_executor.rb +215 -217
  264. data/lib/rubydb/execution/join_executor.rb +243 -249
  265. data/lib/rubydb/execution/limit_executor.rb +83 -85
  266. data/lib/rubydb/execution/operator_selection.rb +57 -0
  267. data/lib/rubydb/execution/optimizer.rb +227 -215
  268. data/lib/rubydb/execution/physical_plan.rb +47 -0
  269. data/lib/rubydb/execution/plan.rb +355 -353
  270. data/lib/rubydb/execution/planner.rb +508 -536
  271. data/lib/rubydb/execution/predicate.rb +235 -235
  272. data/lib/rubydb/execution/scan.rb +49 -49
  273. data/lib/rubydb/execution/sequential_scan.rb +63 -63
  274. data/lib/rubydb/execution/sort_executor.rb +194 -185
  275. data/lib/rubydb/execution/update_executor.rb +160 -162
  276. data/lib/rubydb/functions/aggregate.rb +70 -70
  277. data/lib/rubydb/functions/date_functions.rb +274 -278
  278. data/lib/rubydb/functions/function.rb +85 -85
  279. data/lib/rubydb/functions/json_functions.rb +231 -215
  280. data/lib/rubydb/functions/numeric_functions.rb +346 -346
  281. data/lib/rubydb/functions/scalar.rb +52 -52
  282. data/lib/rubydb/functions/string_functions.rb +383 -383
  283. data/lib/rubydb/functions/system_functions.rb +258 -246
  284. data/lib/rubydb/history/as_of.rb +238 -238
  285. data/lib/rubydb/history/change.rb +105 -105
  286. data/lib/rubydb/history/history.rb +131 -131
  287. data/lib/rubydb/history/history_manager.rb +228 -229
  288. data/lib/rubydb/history/temporal_query.rb +202 -202
  289. data/lib/rubydb/history/timeline.rb +144 -144
  290. data/lib/rubydb/indexes/btree.rb +215 -186
  291. data/lib/rubydb/indexes/btree_cursor.rb +258 -258
  292. data/lib/rubydb/indexes/btree_node.rb +384 -385
  293. data/lib/rubydb/indexes/hash_index.rb +150 -150
  294. data/lib/rubydb/indexes/index.rb +71 -71
  295. data/lib/rubydb/indexes/index_manager.rb +408 -406
  296. data/lib/rubydb/indexes/index_scan.rb +466 -470
  297. data/lib/rubydb/migrations/migration.rb +253 -254
  298. data/lib/rubydb/migrations/migration_lock.rb +146 -146
  299. data/lib/rubydb/migrations/migration_manager.rb +187 -176
  300. data/lib/rubydb/migrations/migration_version.rb +71 -71
  301. data/lib/rubydb/migrations/schema_diff.rb +211 -211
  302. data/lib/rubydb/migrations/schema_version.rb +64 -64
  303. data/lib/rubydb/monitoring/events.rb +155 -160
  304. data/lib/rubydb/monitoring/health.rb +216 -222
  305. data/lib/rubydb/monitoring/logger.rb +188 -193
  306. data/lib/rubydb/monitoring/metrics.rb +363 -359
  307. data/lib/rubydb/monitoring/performance.rb +176 -176
  308. data/lib/rubydb/monitoring/statistics.rb +168 -170
  309. data/lib/rubydb/mvcc/garbage_collector.rb +199 -199
  310. data/lib/rubydb/mvcc/mvcc.rb +16 -16
  311. data/lib/rubydb/mvcc/snapshot.rb +146 -147
  312. data/lib/rubydb/mvcc/vacuum.rb +180 -180
  313. data/lib/rubydb/mvcc/version.rb +106 -106
  314. data/lib/rubydb/mvcc/version_store.rb +396 -398
  315. data/lib/rubydb/mvcc/visibility.rb +107 -109
  316. data/lib/rubydb/protocol/capabilities.rb +125 -125
  317. data/lib/rubydb/protocol/decoder.rb +142 -145
  318. data/lib/rubydb/protocol/encoder.rb +131 -136
  319. data/lib/rubydb/protocol/handshake.rb +306 -305
  320. data/lib/rubydb/protocol/message.rb +121 -121
  321. data/lib/rubydb/protocol/parameter_binder.rb +101 -0
  322. data/lib/rubydb/protocol/protocol.rb +276 -277
  323. data/lib/rubydb/protocol/version.rb +54 -54
  324. data/lib/rubydb/rails/adapter.rb +245 -239
  325. data/lib/rubydb/rails/connection.rb +312 -314
  326. data/lib/rubydb/rails/database_statements.rb +122 -122
  327. data/lib/rubydb/rails/migration.rb +131 -131
  328. data/lib/rubydb/rails/quoting.rb +109 -109
  329. data/lib/rubydb/rails/result.rb +117 -117
  330. data/lib/rubydb/rails/schema_statements.rb +339 -339
  331. data/lib/rubydb/rails/transaction.rb +105 -105
  332. data/lib/rubydb/rails/type.rb +126 -126
  333. data/lib/rubydb/recovery/checkpoint.rb +261 -257
  334. data/lib/rubydb/recovery/consistency.rb +457 -467
  335. data/lib/rubydb/recovery/corruption_detector.rb +5 -5
  336. data/lib/rubydb/recovery/crash_recovery.rb +381 -387
  337. data/lib/rubydb/recovery/recovery_manager.rb +204 -206
  338. data/lib/rubydb/recovery/redo.rb +235 -237
  339. data/lib/rubydb/recovery/undo.rb +204 -206
  340. data/lib/rubydb/replication/failover.rb +5 -5
  341. data/lib/rubydb/replication/fencing.rb +63 -63
  342. data/lib/rubydb/replication/primary.rb +461 -450
  343. data/lib/rubydb/replication/replica.rb +382 -384
  344. data/lib/rubydb/replication/replication_log.rb +194 -200
  345. data/lib/rubydb/replication/replication_manager.rb +307 -308
  346. data/lib/rubydb/replication/replication_slot.rb +293 -295
  347. data/lib/rubydb/replication/replication_stream.rb +198 -201
  348. data/lib/rubydb/rubydb.rb +570 -560
  349. data/lib/rubydb/security/access_control.rb +252 -254
  350. data/lib/rubydb/security/audit_log.rb +209 -213
  351. data/lib/rubydb/security/authentication.rb +302 -302
  352. data/lib/rubydb/security/authorization.rb +282 -282
  353. data/lib/rubydb/security/credentials.rb +192 -196
  354. data/lib/rubydb/security/password.rb +205 -215
  355. data/lib/rubydb/security/permissions.rb +74 -74
  356. data/lib/rubydb/security/role.rb +99 -101
  357. data/lib/rubydb/security/user.rb +86 -86
  358. data/lib/rubydb/server/connection.rb +383 -366
  359. data/lib/rubydb/server/connection_pool.rb +193 -193
  360. data/lib/rubydb/server/lifecycle.rb +227 -228
  361. data/lib/rubydb/server/listener.rb +139 -136
  362. data/lib/rubydb/server/request_handler.rb +277 -276
  363. data/lib/rubydb/server/server.rb +363 -364
  364. data/lib/rubydb/server/session.rb +416 -369
  365. data/lib/rubydb/server/worker.rb +206 -210
  366. data/lib/rubydb/server/worker_pool.rb +168 -168
  367. data/lib/rubydb/sql/ast/alter_table.rb +169 -169
  368. data/lib/rubydb/sql/ast/begin_transaction.rb +47 -47
  369. data/lib/rubydb/sql/ast/commit.rb +37 -37
  370. data/lib/rubydb/sql/ast/constraint.rb +92 -83
  371. data/lib/rubydb/sql/ast/create_database.rb +41 -41
  372. data/lib/rubydb/sql/ast/create_index.rb +61 -61
  373. data/lib/rubydb/sql/ast/create_schema.rb +52 -52
  374. data/lib/rubydb/sql/ast/create_table.rb +187 -187
  375. data/lib/rubydb/sql/ast/delete.rb +54 -54
  376. data/lib/rubydb/sql/ast/drop_database.rb +41 -41
  377. data/lib/rubydb/sql/ast/drop_index.rb +41 -41
  378. data/lib/rubydb/sql/ast/drop_schema.rb +49 -49
  379. data/lib/rubydb/sql/ast/drop_table.rb +49 -49
  380. data/lib/rubydb/sql/ast/explain.rb +64 -64
  381. data/lib/rubydb/sql/ast/expression.rb +617 -604
  382. data/lib/rubydb/sql/ast/insert.rb +66 -66
  383. data/lib/rubydb/sql/ast/node.rb +42 -42
  384. data/lib/rubydb/sql/ast/rollback.rb +63 -63
  385. data/lib/rubydb/sql/ast/savepoint.rb +59 -59
  386. data/lib/rubydb/sql/ast/select.rb +88 -88
  387. data/lib/rubydb/sql/ast/set_operation.rb +22 -20
  388. data/lib/rubydb/sql/ast/trigger.rb +35 -29
  389. data/lib/rubydb/sql/ast/update.rb +88 -88
  390. data/lib/rubydb/sql/ast/vacuum.rb +19 -19
  391. data/lib/rubydb/sql/ast/view.rb +38 -32
  392. data/lib/rubydb/sql/ast/with.rb +32 -32
  393. data/lib/rubydb/sql/grammar.rb +86 -86
  394. data/lib/rubydb/sql/keywords.rb +156 -156
  395. data/lib/rubydb/sql/lexer.rb +209 -214
  396. data/lib/rubydb/sql/operators.rb +100 -100
  397. data/lib/rubydb/sql/parser.rb +1167 -1170
  398. data/lib/rubydb/sql/planner/analyzer.rb +283 -302
  399. data/lib/rubydb/sql/planner/binder.rb +537 -550
  400. data/lib/rubydb/sql/planner/type_checker.rb +427 -431
  401. data/lib/rubydb/sql/token.rb +210 -210
  402. data/lib/rubydb/storage/buffer_frame.rb +44 -44
  403. data/lib/rubydb/storage/buffer_pool.rb +155 -155
  404. data/lib/rubydb/storage/database_lock.rb +74 -74
  405. data/lib/rubydb/storage/deserializer.rb +332 -342
  406. data/lib/rubydb/storage/engine.rb +2409 -2330
  407. data/lib/rubydb/storage/file_manager.rb +191 -187
  408. data/lib/rubydb/storage/free_space_map.rb +79 -81
  409. data/lib/rubydb/storage/page.rb +92 -94
  410. data/lib/rubydb/storage/page_allocator.rb +852 -855
  411. data/lib/rubydb/storage/page_header.rb +63 -67
  412. data/lib/rubydb/storage/page_manager.rb +127 -131
  413. data/lib/rubydb/storage/record.rb +58 -58
  414. data/lib/rubydb/storage/row.rb +78 -78
  415. data/lib/rubydb/storage/serializer.rb +51 -51
  416. data/lib/rubydb/storage/snapshot_reader.rb +167 -0
  417. data/lib/rubydb/storage/storage_layout.rb +151 -151
  418. data/lib/rubydb/storage/storage_manager.rb +114 -114
  419. data/lib/rubydb/storage/tuple.rb +458 -461
  420. data/lib/rubydb/storage/visibility_map.rb +964 -973
  421. data/lib/rubydb/transactions/commit_manager.rb +219 -220
  422. data/lib/rubydb/transactions/isolation.rb +98 -98
  423. data/lib/rubydb/transactions/lock.rb +76 -76
  424. data/lib/rubydb/transactions/lock_manager.rb +359 -362
  425. data/lib/rubydb/transactions/savepoint.rb +142 -143
  426. data/lib/rubydb/transactions/transaction.rb +214 -215
  427. data/lib/rubydb/transactions/transaction_id.rb +84 -84
  428. data/lib/rubydb/transactions/transaction_log.rb +256 -257
  429. data/lib/rubydb/transactions/transaction_manager.rb +434 -435
  430. data/lib/rubydb/types/bigint.rb +36 -36
  431. data/lib/rubydb/types/blob.rb +37 -37
  432. data/lib/rubydb/types/boolean.rb +34 -34
  433. data/lib/rubydb/types/date.rb +39 -39
  434. data/lib/rubydb/types/decimal.rb +48 -48
  435. data/lib/rubydb/types/float.rb +34 -34
  436. data/lib/rubydb/types/integer.rb +36 -36
  437. data/lib/rubydb/types/json.rb +41 -41
  438. data/lib/rubydb/types/null.rb +34 -34
  439. data/lib/rubydb/types/smallint.rb +36 -36
  440. data/lib/rubydb/types/text.rb +37 -37
  441. data/lib/rubydb/types/time.rb +46 -46
  442. data/lib/rubydb/types/timestamp.rb +39 -39
  443. data/lib/rubydb/types/type.rb +119 -119
  444. data/lib/rubydb/types/uuid.rb +47 -47
  445. data/lib/rubydb/types/varchar.rb +37 -37
  446. data/lib/rubydb/version.rb +32 -32
  447. data/lib/rubydb/wal/archive.rb +207 -193
  448. data/lib/rubydb/wal/checkpoint.rb +181 -183
  449. data/lib/rubydb/wal/lsn.rb +94 -94
  450. data/lib/rubydb/wal/reader.rb +259 -260
  451. data/lib/rubydb/wal/record.rb +105 -105
  452. data/lib/rubydb/wal/segment.rb +193 -193
  453. data/lib/rubydb/wal/wal.rb +481 -452
  454. data/lib/rubydb/wal/writer.rb +236 -236
  455. data/lib/rubydb.rb +7 -7
  456. data/packaging/docker/docker-compose.failover.yml +43 -43
  457. data/packaging/homebrew/rubydb.rb +19 -19
  458. data/rubydb.gemspec +70 -57
  459. data/scripts/benchmark +7 -7
  460. data/scripts/build_accelerator +49 -0
  461. data/scripts/durability_drill +37 -37
  462. data/scripts/fuzz +63 -63
  463. data/scripts/release +77 -42
  464. data/scripts/release_check +43 -43
  465. data/scripts/replication_failover_drill +268 -250
  466. data/scripts/replication_network_failover_drill +287 -255
  467. data/scripts/restore_drill +45 -45
  468. data/scripts/security +45 -0
  469. metadata +102 -1
@@ -1,398 +1,396 @@
1
- # frozen_string_literal: true
2
-
3
- require "set"
4
- require "json"
5
- require "fileutils"
6
- require "time"
7
- require "monitor"
8
- require_relative "version"
9
-
10
- module RubyDB
11
- module MVCC
12
- # VersionStore - Stores and manages all versions of rows
13
- class VersionStore
14
- attr_reader :stats, :version_count, :row_count
15
-
16
- def initialize(config = {})
17
- @versions = {} # row_id => [versions]
18
- @version_map = {} # version_id => version
19
- @version_keys = {} # version_id => storage key
20
- @active_versions = Set.new
21
- @stats = {
22
- versions_created: 0,
23
- versions_committed: 0,
24
- versions_aborted: 0,
25
- versions_deleted: 0,
26
- versions_pruned: 0,
27
- cache_hits: 0,
28
- cache_misses: 0
29
- }
30
- @version_count = 0
31
- @row_count = 0
32
- @max_versions_per_row = config[:max_versions_per_row] || 100
33
- @lock = Monitor.new
34
- @persistence_path = config[:persistence_path]
35
- @dirty = false
36
- load if @persistence_path && File.exist?(@persistence_path)
37
- end
38
-
39
- def create_version(row_id, data, transaction_id, key: row_id)
40
- @lock.synchronize do
41
- # Get current latest version
42
- versions = @versions[key] || []
43
- prev_version_id = versions.last&.version_id
44
-
45
- # Create new version
46
- version = Version.new(row_id, data, transaction_id, prev_version_id)
47
-
48
- # Store version
49
- @versions[key] ||= []
50
- @versions[key] << version
51
- @version_map[version.version_id] = version
52
- @version_keys[version.version_id] = key
53
-
54
- # Update active versions
55
- @active_versions.add(version.version_id)
56
-
57
- @version_count += 1
58
- @row_count = @versions.size
59
- @stats[:versions_created] += 1
60
- @dirty = true
61
-
62
- version
63
- end
64
- end
65
-
66
- def commit_version(version, commit_id = nil)
67
- @lock.synchronize do
68
- version.commit(commit_id)
69
- @active_versions.delete(version.version_id)
70
- @stats[:versions_committed] += 1
71
- @dirty = true
72
- end
73
- end
74
-
75
- def abort_version(version)
76
- @lock.synchronize do
77
- version.abort
78
- @active_versions.delete(version.version_id)
79
-
80
- # Remove from version chain if it was the latest
81
- key = @version_keys[version.version_id] || version.row_id
82
- versions = @versions[key]
83
- if versions && versions.last&.version_id == version.version_id
84
- versions.pop
85
- @version_map.delete(version.version_id)
86
- @version_keys.delete(version.version_id)
87
- end
88
-
89
- @stats[:versions_aborted] += 1
90
- @dirty = true
91
- end
92
- end
93
-
94
- def get_latest_version(row_id, transaction_id = nil, snapshot = nil, key: row_id)
95
- @lock.synchronize do
96
- # Check cache for snapshot
97
- cache_key = "#{row_id}:#{transaction_id}:#{snapshot&.id}"
98
-
99
- versions = @versions[key]
100
- return nil unless versions
101
-
102
- # Get the latest committed version visible to transaction
103
- versions.reverse_each do |version|
104
- if snapshot
105
- if version.visible_to?(transaction_id, snapshot)
106
- return nil if version.is_deleted
107
- @stats[:cache_hits] += 1
108
- return version
109
- end
110
- elsif version.is_committed
111
- # A nil transaction ID means "latest committed". Otherwise a
112
- # version is visible only when its commit ID is not newer than
113
- # the reader's transaction ID.
114
- if transaction_id.nil? || version.commit_id.nil? || version.commit_id <= transaction_id
115
- @stats[:cache_hits] += 1
116
- return version.is_deleted ? nil : version
117
- end
118
- end
119
- end
120
-
121
- @stats[:cache_misses] += 1
122
- nil
123
- end
124
- end
125
-
126
- def get_version_at_time(row_id, time)
127
- @lock.synchronize do
128
- versions = @versions[row_id]
129
- return nil unless versions
130
-
131
- # Find version that was active at the given time
132
- versions.reverse_each do |version|
133
- if version.created_at <= time
134
- return version
135
- end
136
- end
137
-
138
- nil
139
- end
140
- end
141
-
142
- def get_all_versions(row_id)
143
- @lock.synchronize do
144
- @versions[row_id] || []
145
- end
146
- end
147
-
148
- def keys(prefix = nil)
149
- @lock.synchronize do
150
- values = @versions.keys
151
- prefix ? values.select { |key| key.to_s.start_with?(prefix.to_s) } : values
152
- end
153
- end
154
-
155
- def validate_serializable!(snapshot, read_keys, write_keys, read_predicates: [])
156
- @lock.synchronize do
157
- dependencies = Set.new(read_keys.to_a + write_keys.to_a)
158
- @versions.each do |key, versions|
159
- predicate_conflict = read_predicates.any? { |predicate| key.to_s.start_with?(predicate.to_s) }
160
- next unless dependencies.include?(key) || predicate_conflict
161
-
162
- versions.each do |version|
163
- next unless version.is_committed
164
- next unless version.commit_id && version.commit_id.to_i > snapshot.transaction_id.to_i
165
- next if snapshot.committed_transactions.include?(version.commit_id)
166
-
167
- raise RubyDB::DatabaseError,
168
- "serialization failure: concurrent commit conflicts with #{key.inspect}"
169
- end
170
- end
171
- true
172
- end
173
- end
174
-
175
- def get_version(version_id)
176
- @lock.synchronize do
177
- @version_map[version_id]
178
- end
179
- end
180
-
181
- def delete_version(version)
182
- @lock.synchronize do
183
- version.mark_deleted
184
- @active_versions.delete(version.version_id)
185
- @stats[:versions_deleted] += 1
186
- @dirty = true
187
- end
188
- end
189
-
190
- def prune_versions(row_id, keep_count = nil)
191
- @lock.synchronize do
192
- versions = @versions[row_id]
193
- return 0 unless versions
194
-
195
- keep_count ||= @max_versions_per_row
196
- pruned = 0
197
-
198
- # Remove old versions (keeping the most recent ones)
199
- while versions.size > keep_count
200
- old_version = versions.shift
201
- next unless old_version
202
-
203
- # Only prune committed versions that are not active
204
- if old_version.is_committed && !@active_versions.include?(old_version.version_id)
205
- @version_map.delete(old_version.version_id)
206
- @version_keys.delete(old_version.version_id)
207
- @stats[:versions_pruned] += 1
208
- pruned += 1
209
- else
210
- # Can't prune active versions, move to end
211
- versions << old_version
212
- break
213
- end
214
- end
215
-
216
- @dirty = true if pruned > 0
217
- pruned
218
- end
219
- end
220
-
221
- def active_version_count
222
- @active_versions.size
223
- end
224
-
225
- def prune_all(max_versions = nil)
226
- @lock.synchronize do
227
- total_pruned = 0
228
- @versions.keys.each do |row_id|
229
- versions = @versions[row_id]
230
- keep_count = max_versions || @max_versions_per_row
231
- while versions && versions.size > keep_count
232
- version = versions.first
233
- break unless version.is_committed && !@active_versions.include?(version.version_id)
234
-
235
- versions.shift
236
- @version_map.delete(version.version_id)
237
- @version_keys.delete(version.version_id)
238
- @stats[:versions_pruned] += 1
239
- total_pruned += 1
240
- end
241
- end
242
- @dirty = true if total_pruned > 0
243
- total_pruned
244
- end
245
- end
246
-
247
- def vacuum(min_active_transaction_id: nil)
248
- @lock.synchronize do
249
- # Remove versions that are no longer needed
250
- removed = 0
251
- @versions.each do |row_id, versions|
252
- # Keep the newest version as the base visible version. Older
253
- # committed history may be removed only before the oldest active
254
- # transaction; active and uncommitted versions are never removed.
255
- newest = versions.last
256
- protected_base = versions.reverse.find(&:is_committed)
257
- versions.delete_if do |version|
258
- removable = version.is_aborted ||
259
- (version != newest && version != protected_base && version.is_committed &&
260
- min_active_transaction_id && version.commit_id &&
261
- version.commit_id < min_active_transaction_id)
262
- if removable && !@active_versions.include?(version.version_id)
263
- @version_map.delete(version.version_id)
264
- removed += 1
265
- true
266
- else
267
- false
268
- end
269
- end
270
-
271
- # Remove empty arrays
272
- if versions.empty?
273
- @versions.delete(row_id)
274
- end
275
- end
276
-
277
- @row_count = @versions.size
278
- @stats[:versions_pruned] += removed
279
- @dirty = true if removed > 0
280
- removed
281
- end
282
- end
283
-
284
- def serialize
285
- @lock.synchronize do
286
- data = {
287
- versions: {},
288
- version_count: @version_count,
289
- row_count: @row_count,
290
- active_versions: @active_versions.to_a,
291
- timestamp: Time.now.iso8601
292
- }
293
-
294
- @versions.each do |row_key, versions|
295
- data[:versions][row_key.to_s] = versions.map do |v|
296
- {
297
- row_id: v.row_id,
298
- version_id: v.version_id,
299
- created_at: v.created_at.iso8601,
300
- transaction_id: v.transaction_id,
301
- prev_version_id: v.prev_version_id,
302
- is_committed: v.is_committed,
303
- is_aborted: v.is_aborted,
304
- is_deleted: v.is_deleted,
305
- commit_id: v.commit_id,
306
- visibility: v.visibility,
307
- data: v.data
308
- }
309
- end
310
- end
311
-
312
- JSON.generate(data)
313
- end
314
- end
315
-
316
- def deserialize(json_data)
317
- @lock.synchronize do
318
- data = JSON.parse(json_data, symbolize_names: true)
319
-
320
- @versions.clear
321
- @version_map.clear
322
- @version_keys.clear
323
- @active_versions.clear
324
-
325
- data[:versions].each do |row_key, versions_data|
326
- row_key = row_key.to_s
327
- row_key = row_key.to_i if row_key.match?(/\A\d+\z/)
328
- @versions[row_key] = versions_data.map do |v_data|
329
- version = Version.new(
330
- v_data[:row_id],
331
- v_data[:data],
332
- v_data[:transaction_id],
333
- v_data[:prev_version_id]
334
- )
335
- version.instance_variable_set(:@version_id, v_data[:version_id])
336
- version.instance_variable_set(:@created_at, Time.parse(v_data[:created_at]))
337
- version.instance_variable_set(:@is_committed, v_data[:is_committed])
338
- version.instance_variable_set(:@is_aborted, v_data[:is_aborted])
339
- version.instance_variable_set(:@is_deleted, v_data[:is_deleted])
340
- version.instance_variable_set(:@commit_id, v_data[:commit_id])
341
- version.instance_variable_set(:@visibility, v_data[:visibility].to_sym)
342
-
343
- @version_map[version.version_id] = version
344
- @version_keys[version.version_id] = row_key
345
- version
346
- end
347
- end
348
-
349
- @version_count = data[:version_count] || 0
350
- @row_count = data[:row_count] || 0
351
- @active_versions = Set.new(data[:active_versions] || [])
352
- @dirty = false
353
- end
354
- end
355
-
356
- def persist
357
- return unless @persistence_path && @dirty
358
-
359
- @lock.synchronize do
360
- FileUtils.mkdir_p(File.dirname(@persistence_path))
361
- temp_path = "#{@persistence_path}.tmp"
362
- File.open(temp_path, "w") do |file|
363
- file.write(serialize)
364
- file.flush
365
- file.fsync
366
- end
367
- File.rename(temp_path, @persistence_path)
368
- @dirty = false
369
- end
370
- end
371
-
372
- def load
373
- return unless @persistence_path && File.exist?(@persistence_path)
374
-
375
- @lock.synchronize do
376
- json_data = File.read(@persistence_path)
377
- deserialize(json_data)
378
- end
379
- end
380
-
381
- def stats
382
- @lock.synchronize do
383
- @stats.merge({
384
- version_count: @version_count,
385
- row_count: @row_count,
386
- active_versions: @active_versions.size,
387
- rows_with_versions: @versions.size,
388
- dirty: @dirty
389
- })
390
- end
391
- end
392
-
393
- def inspect
394
- "#<VersionStore versions=#{@version_count} rows=#{@row_count} active=#{@active_versions.size}>"
395
- end
396
- end
397
- end
398
- end
1
+ # frozen_string_literal: true
2
+
3
+ require "json"
4
+ require "fileutils"
5
+ require "time"
6
+ require "monitor"
7
+ require_relative "version"
8
+
9
+ module RubyDB
10
+ module MVCC
11
+ # VersionStore - Stores and manages all versions of rows
12
+ class VersionStore
13
+ attr_reader :stats, :version_count, :row_count
14
+
15
+ def initialize(config = {})
16
+ @versions = {} # row_id => [versions]
17
+ @version_map = {} # version_id => version
18
+ @version_keys = {} # version_id => storage key
19
+ @active_versions = Set.new
20
+ @stats = {
21
+ versions_created: 0,
22
+ versions_committed: 0,
23
+ versions_aborted: 0,
24
+ versions_deleted: 0,
25
+ versions_pruned: 0,
26
+ cache_hits: 0,
27
+ cache_misses: 0
28
+ }
29
+ @version_count = 0
30
+ @row_count = 0
31
+ @max_versions_per_row = config[:max_versions_per_row] || 100
32
+ @lock = Monitor.new
33
+ @persistence_path = config[:persistence_path]
34
+ @dirty = false
35
+ load if @persistence_path && File.exist?(@persistence_path)
36
+ end
37
+
38
+ def create_version(row_id, data, transaction_id, key: row_id)
39
+ @lock.synchronize do
40
+ # Get current latest version
41
+ versions = @versions[key] || []
42
+ prev_version_id = versions.last&.version_id
43
+
44
+ # Create new version
45
+ version = Version.new(row_id, data, transaction_id, prev_version_id)
46
+
47
+ # Store version
48
+ @versions[key] ||= []
49
+ @versions[key] << version
50
+ @version_map[version.version_id] = version
51
+ @version_keys[version.version_id] = key
52
+
53
+ # Update active versions
54
+ @active_versions.add(version.version_id)
55
+
56
+ @version_count += 1
57
+ @row_count = @versions.size
58
+ @stats[:versions_created] += 1
59
+ @dirty = true
60
+
61
+ version
62
+ end
63
+ end
64
+
65
+ def commit_version(version, commit_id = nil)
66
+ @lock.synchronize do
67
+ version.commit(commit_id)
68
+ @active_versions.delete(version.version_id)
69
+ @stats[:versions_committed] += 1
70
+ @dirty = true
71
+ end
72
+ end
73
+
74
+ def abort_version(version)
75
+ @lock.synchronize do
76
+ version.abort
77
+ @active_versions.delete(version.version_id)
78
+
79
+ # Remove from version chain if it was the latest
80
+ key = @version_keys[version.version_id] || version.row_id
81
+ versions = @versions[key]
82
+ if versions && versions.last&.version_id == version.version_id
83
+ versions.pop
84
+ @version_map.delete(version.version_id)
85
+ @version_keys.delete(version.version_id)
86
+ end
87
+
88
+ @stats[:versions_aborted] += 1
89
+ @dirty = true
90
+ end
91
+ end
92
+
93
+ def get_latest_version(row_id, transaction_id = nil, snapshot = nil, key: row_id)
94
+ @lock.synchronize do
95
+ # Check cache for snapshot
96
+
97
+ versions = @versions[key]
98
+ return nil unless versions
99
+
100
+ # Get the latest committed version visible to transaction
101
+ versions.reverse_each do |version|
102
+ if snapshot
103
+ if version.visible_to?(transaction_id, snapshot)
104
+ return nil if version.is_deleted
105
+ @stats[:cache_hits] += 1
106
+ return version
107
+ end
108
+ elsif version.is_committed
109
+ # A nil transaction ID means "latest committed". Otherwise a
110
+ # version is visible only when its commit ID is not newer than
111
+ # the reader's transaction ID.
112
+ if transaction_id.nil? || version.commit_id.nil? || version.commit_id <= transaction_id
113
+ @stats[:cache_hits] += 1
114
+ return version.is_deleted ? nil : version
115
+ end
116
+ end
117
+ end
118
+
119
+ @stats[:cache_misses] += 1
120
+ nil
121
+ end
122
+ end
123
+
124
+ def get_version_at_time(row_id, time)
125
+ @lock.synchronize do
126
+ versions = @versions[row_id]
127
+ return nil unless versions
128
+
129
+ # Find version that was active at the given time
130
+ versions.reverse_each do |version|
131
+ if version.created_at <= time
132
+ return version
133
+ end
134
+ end
135
+
136
+ nil
137
+ end
138
+ end
139
+
140
+ def get_all_versions(row_id)
141
+ @lock.synchronize do
142
+ @versions[row_id] || []
143
+ end
144
+ end
145
+
146
+ def keys(prefix = nil)
147
+ @lock.synchronize do
148
+ values = @versions.keys
149
+ prefix ? values.select { |key| key.to_s.start_with?(prefix.to_s) } : values
150
+ end
151
+ end
152
+
153
+ def validate_serializable!(snapshot, read_keys, write_keys, read_predicates: [])
154
+ @lock.synchronize do
155
+ dependencies = Set.new(read_keys.to_a + write_keys.to_a)
156
+ @versions.each do |key, versions|
157
+ predicate_conflict = read_predicates.any? { |predicate| key.to_s.start_with?(predicate.to_s) }
158
+ next unless dependencies.include?(key) || predicate_conflict
159
+
160
+ versions.each do |version|
161
+ next unless version.is_committed
162
+ next unless version.commit_id && version.commit_id.to_i > snapshot.transaction_id.to_i
163
+ next if snapshot.committed_transactions.include?(version.commit_id)
164
+
165
+ raise RubyDB::DatabaseError,
166
+ "serialization failure: concurrent commit conflicts with #{key.inspect}"
167
+ end
168
+ end
169
+ true
170
+ end
171
+ end
172
+
173
+ def get_version(version_id)
174
+ @lock.synchronize do
175
+ @version_map[version_id]
176
+ end
177
+ end
178
+
179
+ def delete_version(version)
180
+ @lock.synchronize do
181
+ version.mark_deleted
182
+ @active_versions.delete(version.version_id)
183
+ @stats[:versions_deleted] += 1
184
+ @dirty = true
185
+ end
186
+ end
187
+
188
+ def prune_versions(row_id, keep_count = nil)
189
+ @lock.synchronize do
190
+ versions = @versions[row_id]
191
+ return 0 unless versions
192
+
193
+ keep_count ||= @max_versions_per_row
194
+ pruned = 0
195
+
196
+ # Remove old versions (keeping the most recent ones)
197
+ while versions.size > keep_count
198
+ old_version = versions.shift
199
+ next unless old_version
200
+
201
+ # Only prune committed versions that are not active
202
+ if old_version.is_committed && !@active_versions.include?(old_version.version_id)
203
+ @version_map.delete(old_version.version_id)
204
+ @version_keys.delete(old_version.version_id)
205
+ @stats[:versions_pruned] += 1
206
+ pruned += 1
207
+ else
208
+ # Can't prune active versions, move to end
209
+ versions << old_version
210
+ break
211
+ end
212
+ end
213
+
214
+ @dirty = true if pruned > 0
215
+ pruned
216
+ end
217
+ end
218
+
219
+ def active_version_count
220
+ @active_versions.size
221
+ end
222
+
223
+ def prune_all(max_versions = nil)
224
+ @lock.synchronize do
225
+ total_pruned = 0
226
+ @versions.keys.each do |row_id|
227
+ versions = @versions[row_id]
228
+ keep_count = max_versions || @max_versions_per_row
229
+ while versions && versions.size > keep_count
230
+ version = versions.first
231
+ break unless version.is_committed && !@active_versions.include?(version.version_id)
232
+
233
+ versions.shift
234
+ @version_map.delete(version.version_id)
235
+ @version_keys.delete(version.version_id)
236
+ @stats[:versions_pruned] += 1
237
+ total_pruned += 1
238
+ end
239
+ end
240
+ @dirty = true if total_pruned > 0
241
+ total_pruned
242
+ end
243
+ end
244
+
245
+ def vacuum(min_active_transaction_id: nil)
246
+ @lock.synchronize do
247
+ # Remove versions that are no longer needed
248
+ removed = 0
249
+ @versions.each do |row_id, versions|
250
+ # Keep the newest version as the base visible version. Older
251
+ # committed history may be removed only before the oldest active
252
+ # transaction; active and uncommitted versions are never removed.
253
+ newest = versions.last
254
+ protected_base = versions.reverse.find(&:is_committed)
255
+ versions.delete_if do |version|
256
+ removable = version.is_aborted ||
257
+ (version != newest && version != protected_base && version.is_committed &&
258
+ min_active_transaction_id && version.commit_id &&
259
+ version.commit_id < min_active_transaction_id)
260
+ if removable && !@active_versions.include?(version.version_id)
261
+ @version_map.delete(version.version_id)
262
+ removed += 1
263
+ true
264
+ else
265
+ false
266
+ end
267
+ end
268
+
269
+ # Remove empty arrays
270
+ if versions.empty?
271
+ @versions.delete(row_id)
272
+ end
273
+ end
274
+
275
+ @row_count = @versions.size
276
+ @stats[:versions_pruned] += removed
277
+ @dirty = true if removed > 0
278
+ removed
279
+ end
280
+ end
281
+
282
+ def serialize
283
+ @lock.synchronize do
284
+ data = {
285
+ versions: {},
286
+ version_count: @version_count,
287
+ row_count: @row_count,
288
+ active_versions: @active_versions.to_a,
289
+ timestamp: Time.now.iso8601
290
+ }
291
+
292
+ @versions.each do |row_key, versions|
293
+ data[:versions][row_key.to_s] = versions.map do |v|
294
+ {
295
+ row_id: v.row_id,
296
+ version_id: v.version_id,
297
+ created_at: v.created_at.iso8601,
298
+ transaction_id: v.transaction_id,
299
+ prev_version_id: v.prev_version_id,
300
+ is_committed: v.is_committed,
301
+ is_aborted: v.is_aborted,
302
+ is_deleted: v.is_deleted,
303
+ commit_id: v.commit_id,
304
+ visibility: v.visibility,
305
+ data: v.data
306
+ }
307
+ end
308
+ end
309
+
310
+ JSON.generate(data)
311
+ end
312
+ end
313
+
314
+ def deserialize(json_data)
315
+ @lock.synchronize do
316
+ data = JSON.parse(json_data, symbolize_names: true)
317
+
318
+ @versions.clear
319
+ @version_map.clear
320
+ @version_keys.clear
321
+ @active_versions.clear
322
+
323
+ data[:versions].each do |row_key, versions_data|
324
+ row_key = row_key.to_s
325
+ row_key = row_key.to_i if row_key.match?(/\A\d+\z/)
326
+ @versions[row_key] = versions_data.map do |v_data|
327
+ version = Version.new(
328
+ v_data[:row_id],
329
+ v_data[:data],
330
+ v_data[:transaction_id],
331
+ v_data[:prev_version_id]
332
+ )
333
+ version.instance_variable_set(:@version_id, v_data[:version_id])
334
+ version.instance_variable_set(:@created_at, Time.parse(v_data[:created_at]))
335
+ version.instance_variable_set(:@is_committed, v_data[:is_committed])
336
+ version.instance_variable_set(:@is_aborted, v_data[:is_aborted])
337
+ version.instance_variable_set(:@is_deleted, v_data[:is_deleted])
338
+ version.instance_variable_set(:@commit_id, v_data[:commit_id])
339
+ version.instance_variable_set(:@visibility, v_data[:visibility].to_sym)
340
+
341
+ @version_map[version.version_id] = version
342
+ @version_keys[version.version_id] = row_key
343
+ version
344
+ end
345
+ end
346
+
347
+ @version_count = data[:version_count] || 0
348
+ @row_count = data[:row_count] || 0
349
+ @active_versions = Set.new(data[:active_versions] || [])
350
+ @dirty = false
351
+ end
352
+ end
353
+
354
+ def persist
355
+ return unless @persistence_path && @dirty
356
+
357
+ @lock.synchronize do
358
+ FileUtils.mkdir_p(File.dirname(@persistence_path))
359
+ temp_path = "#{@persistence_path}.tmp"
360
+ File.open(temp_path, "w") do |file|
361
+ file.write(serialize)
362
+ file.flush
363
+ file.fsync
364
+ end
365
+ File.rename(temp_path, @persistence_path)
366
+ @dirty = false
367
+ end
368
+ end
369
+
370
+ def load
371
+ return unless @persistence_path && File.exist?(@persistence_path)
372
+
373
+ @lock.synchronize do
374
+ json_data = File.read(@persistence_path)
375
+ deserialize(json_data)
376
+ end
377
+ end
378
+
379
+ def stats
380
+ @lock.synchronize do
381
+ @stats.merge({
382
+ version_count: @version_count,
383
+ row_count: @row_count,
384
+ active_versions: @active_versions.size,
385
+ rows_with_versions: @versions.size,
386
+ dirty: @dirty
387
+ })
388
+ end
389
+ end
390
+
391
+ def inspect
392
+ "#<VersionStore versions=#{@version_count} rows=#{@row_count} active=#{@active_versions.size}>"
393
+ end
394
+ end
395
+ end
396
+ end