393 commits

Author	SHA1	Message	Date
Lukas Wirth	679d82c4e7	Gate accumulator feature behind a feature flag (#946)	2025-08-02 08:41:17 +00:00
Lukas Wirth	211bc158df	Do manual trait casting (#922) ... * Do manual trait upcasting instead of downcasting * Remove another dynamic `zalsa` call * Rename UpCaster back to DownCaster * Address reviews	2025-07-31 15:56:13 +00:00
Ibraheem Ahmed	53cd6b15ba	remove bounds and type checks from IngredientCache (#937)	2025-07-21 16:10:10 +00:00
Ibraheem Ahmed	0e1df67ec4	Avoid dynamic dispatch to access memo tables (#941) ... * avoid dynamic dispatch to access memo tables * update comment	2025-07-21 16:09:11 +00:00
Ibraheem Ahmed	dba66f1a37	Use inventory for static ingredient registration (#934) ... * use `inventory` for static ingredient registration * remove unnecessary synchronization from memo tables * use global ingredient caches for database-independent ingredients * add manual ingredient registration API * remove static ingredient index optimization when manual registration is in use * fix atomic imports * simplify ingredient caches	2025-07-18 04:55:50 +00:00
Lukas Wirth	f384ab538e	Emit self ty for query debug name of assoc function queries (#927)	2025-06-27 09:23:25 +00:00
Ibraheem Ahmed	d44f638408	Replace ingredient cache with faster ingredient map (#921) ... * replace ingredient cache with faster ingredient map * avoid checking for downcasters in ingredient cache slow-path * pre-size ingredient map * avoid double lookup in ingredient creation slow-path	2025-06-27 07:19:58 +00:00
Ibraheem Ahmed	0666e2018b	add option to track heap memory usage of memos (#925)	2025-06-25 20:26:49 +00:00
Ibraheem Ahmed	c145596ef8	Add API to dump memory usage (#916) ... * add memory usage information hooks * gate memory usage API under `salsa_unstable` feature * use snapshot tests for memory usage API	2025-06-20 17:15:43 +00:00
Lukas Wirth	8528bab609	feat: Update derive field overwrite support (#747) ... * `Update` derive field overwrite support * Allow overwriting tracked struct field update functions	2025-06-13 17:17:39 +00:00
Ibraheem Ahmed	b8cdd300a0	add an option to tune interned garbage collection (#911)	2025-06-13 06:17:45 +00:00
Lukas Wirth	6ced42b334	Use explicit discriminants for QueryOriginKind for better comparisons (#913)	2025-06-12 18:32:33 +00:00
Ibraheem Ahmed	dc9066d667	use latest revision for dependencies on interned values (#908)	2025-06-11 06:03:30 +00:00
Ibraheem Ahmed	4161bd727f	remove high-durability values from interned LRU (#907)	2025-06-11 05:28:08 +00:00
puuuuh	8924db1441	Preserve attributes on interned/tracked struct fields (#905) ... * Preserve attributes on interned/tracked structs * Pass unknown attributes from fields to getter functions in inputs	2025-06-10 06:34:54 +00:00
Carl Meyer	a95bae5119	panic with string message again for cycle panics (#898)	2025-06-04 15:15:22 +00:00
Lukas Wirth	2c57c5628b	Fix flaky parallel_join test (#900)	2025-06-04 12:24:15 +00:00
Ibraheem Ahmed	15c28cdf88	Simple LRU garbage collection for interned values (#839) ... * simple LRU garbage collection for interned values * shard interned values * gate memo size assertion to 64-bit platforms * extend internal documentation for interned structs * remove `first_interned_at` field from interned values * clarify difference between tracked and interned structs * only keep track of low durability interned values for garbage collection	2025-06-04 08:09:04 +00:00
Micha Reiser	2b5188778e	fix: multithreaded nested fixpoint iteration (#882) ... * Set `validate_final` in `execute` after removing the last cycle head * Add runaway query repro * Add tracing * Fix part 1 * Fix `cycle_head_kinds` to always return provisional for memos that aren't verified final (They should be validated by `validate_same_iteration` or wait for the cycle head * Fix cycle error * Documentation * Fix await for queries depending on initial value * correctly initialize queued * Cleanup * Short circuit if entire query runs on single thread * Move parallel code into its own method * Rename method, add self_key to queued * Revert self-key changes * Move check *after* `deep_verify_memo` * Add a test for a cycle with changing cycle heads * Short circuit more often * Consider iteration in `validate_provisional` * Only yield if all heads result in a cycle. Retry if even just one inner cycle made progress (in which case there's a probably a new memo) * Fix hangs * Cargo fmt * clippy * Fix hang if cycle initial panics * Rename `cycle_head_kind` enable `cycle_a_t1_b_t2_fallback` shuttle test * Cleanup * Docs	2025-06-01 08:45:37 +00:00
puuuuh	f9a9bb2a6a	Don't report stale outputs if there is newer generation in new_outputs (#879) ... * Don't report stale outputs if there is newer generation in new_outputs * Remove outdated comment, add explanation why we ignore generation in diff_outputs	2025-05-27 17:44:15 +00:00
Micha Reiser	e1fe3698ce	Fix hang in nested fixpoint iteration (#871)	2025-05-27 17:06:34 +00:00
Chayim Refael Friedman	2a54667121	Allow lifetimes in arguments in tracked fns with >1 parameters (#880)	2025-05-26 04:50:43 +00:00
Ibraheem Ahmed	0414d89327	Replace loom with shuttle (#876) ... * replace loom with shuttle * inline `empty_cycle_heads` * ignore failing shuttle test	2025-05-23 15:28:51 +00:00
Ibraheem Ahmed	f7b08562ed	Use generational identifiers for tracked structs (#864) ... * use generational identifiers for tracked structs * increase ID generations to 32-bits * remove `created_at` field from tracked structs * clean up tracked struct IDs to handle overflow * log tracing message for leaked tracked structs	2025-05-23 14:18:03 +00:00
Vincent Esche	e859218dea	Allow creation of tracked associated functions (without self) (#859) ... * Rename file "setup_method_body.rs" to "setup_tracked_method_body.rs" * Rename macro `setup_method_body!` to `setup_tracked_method_body!` * Add support for tracked associated functions * Rename struct `MethodArguments` to `AssociatedFunctionArguments` (since every method is an associated function, but not every associated function is a method)	2025-05-23 10:19:12 +00:00
Lukas Wirth	4327d6bb94	Unwind with specific type when encountering an unexpected cycle (#856)	2025-05-22 07:40:11 +00:00
Lukas Wirth	2d4321e989	Implement an !Update bound escape hatch for tracked fn (#867)	2025-05-20 15:36:34 +00:00
Micha Reiser	96eeecb5af	Remove default PartialOrd and Ord derives for salsa-structs (#868)	2025-05-19 11:57:23 +00:00
Micha Reiser	7edce6e248	Fix returns(deref | as_ref | as_deref) in tracked methods (#857)	2025-05-09 09:09:16 +00:00
CheaterCodes	13a2bd7461	Changed return_ref syntax to returns(as_ref) and returns(cloned) (#772) ... * Changed `return_ref` syntax to `returns(as_ref)` and `returns(cloned)` * Implement * renamed module for return_mode * Rename macro, fix docs, add tests, validate return modes * Cargo fmt --------- Co-authored-by: Micha Reiser <micha@reiser.io>	2025-05-09 07:28:54 +00:00
Micha Reiser	a6793e783a	Lazy finalization of cycle participants in maybe_changed_after (#854) ... * Lazy verification of cycle participants in maybe_changed_after * Update comment	2025-05-09 07:03:54 +00:00
Lukas Wirth	af69cc1114	Do not re-verify already verified memoized value in cycle verification (#851)	2025-05-08 16:33:58 +00:00
Lukas Wirth	00dbf01e87	Pass cycle heads as out parameter for maybe_changed_after (#852)	2025-05-08 08:32:27 +00:00
Lukas Wirth	cfa88e9a4d	Move salsa event system into Zalsa (#849) ... * Move salsa event system into `Zalsa` * Encode `None` into ShallowUpdate This shrinks the return value from 16 to 8 bytes	2025-05-07 11:45:49 +00:00
Ibraheem Ahmed	54020469a3	Add loom support (#842) ... * add support for loom tests * add just script and fix comment	2025-05-06 15:36:38 +00:00
Micha Reiser	b2b82bccdb	fix: change detection for fixpoint queries (#836) ... * bug: Fix missing cycle inputs * Pass provisional as old memo to `execute_query` during fixpoint * Remove `provisional` from remove stale output * Revert debug code * Update test * Some documentation * Clean up tests * Fix for direct enclosing query * Format * More comment fiddling * Revert copy only outputs change * Revert `fixpoint_initial` start revision change (worth its own PR) * Align fixpoint handling with Derived * Preserve cycle heads when returning fixpoint initial * Always return changed for `Fixpoint::Initial` * Keep returning unchanged in some cases	2025-05-06 08:11:31 +00:00
Micha Reiser	ad3dff6cb3	better debug name for interned query arguments (#837) ... * feat: Improve debug name for interned query arguments * Change name again	2025-04-30 17:20:49 +00:00
Lukas Wirth	89347c8fc9	Implement a query stack Backtrace analog (#827)	2025-04-29 17:01:15 +00:00
Micha Reiser	b27e3927e9	fix: Access to tracked-struct that was freed during fixpoint (#817) ... * Add test for untracked read on tracked struct created in previous cycle * Initial fix * Restrict seeding to memos from the same revision * Reduce changes * seed_outputs * Cleanup test * Add assertion * Try * Try merging outputs after query executed * Assert logs from first execution * Enable trace level logging * Use `FxIndexSet` in `diff_outputs` * Log more events * Cleanup * Append outputs only once	2025-04-28 14:24:33 +00:00
Lukas Wirth	445b627277	refactor: Simplify ID conversions (#822)	2025-04-28 12:00:16 +00:00
Lukas Wirth	9a9fb4e51b	Add a compile-fail test for a 'static !Update struct (#820)	2025-04-28 07:09:05 +00:00
Micha Reiser	f7eea9d69b	Use DatabaseKey for interned events (#813) ... * Use `DatabaseKey` for interned events * Tame clippy	2025-04-24 10:38:46 +00:00
Chayim Refael Friedman	cf9efae0da	Make interned's last_interned_at equal Revision::MAX if they are interned outside a query (#804) ... There is an assert that `last_interned_at >= last_changed_revision`, and it can fail without this, see the added test.	2025-04-22 10:42:17 +00:00
Chayim Refael Friedman	05b4faddb5	Add a third cycle mode, equivalent to old Salsa cycle behavior (#801) ... That is, directly set a value for all queries that have fallbacks, and ignore all other queries in the cycle. Unlike old Salsa, we still need all cycle heads to be marked, and we still execute the queries to completion, but we throw their result.	2025-04-22 10:35:43 +00:00
Lukas Wirth	883baaf8de	refactor: Inline/Outline more cold and slow paths (#805)	2025-04-21 10:40:22 +00:00
Lukas Wirth	3820cac9a4	Remove incorrect parallel_scope API (#797) ... Rayon does not allow joining on the spawns, meaning we might end up with results in the tests before reading out the panics	2025-04-16 14:09:35 +00:00
Carl Meyer	87bf6b6c2d	add WillIterateCycle event (#790) ... * add WillIterateCycle event * add iteration_count and fell_back to event	2025-04-10 13:13:17 +00:00
Micha Reiser	69272a8fe6	fix: Dereferencing freed memos when verifying provisional memos (#788) ... * Add test * Fix `fetch_hot` and `maybe_changed_after`	2025-04-09 16:39:23 +00:00
Carl Meyer	296a8c78da	allow reuse of cached provisional memos within the same cycle iteration (#786) ... * test for caching provisional values * add iteration-count to cycle heads * CycleHeads insert/extend checks iteration count match * update iteration count in cycle heads * all tests passing * remove debug prints * just walk active query stack once * switch to tracking active cycle iterations on ZalsaLocal * Revert "switch to tracking active cycle iterations on ZalsaLocal" This reverts commit 4ea3d850b53a449dda7af7ceb4cc565d3a64001c. * Revert "just walk active query stack once" This reverts commit 2d7948612be3f411c7787829ff2db74e3a575f0c. * make ActiveQuery::iteration_count private with accessor * iterate active query stack in reverse * use tracing::trace! in hot path * try a cold annotation on validate_same_iteration * Revert "try a cold annotation on validate_same_iteration" This reverts commit 49ceb84bb7.	2025-04-03 15:10:50 +00:00
Lukas Wirth	395b29d08a	fix: #[doc(hidden)] plumbing module (#781)	2025-04-03 10:55:23 +00:00