1391 commits

Author	SHA1	Message	Date
Chayim Refael Friedman	1c8a07cfd6	Cache trait solving across queries in the same revision ... Caching trait solving can do a lot to speed. Unfortunately it also consume a huge amount of memory. Therefore, as part of the migration to the new solver Jack Huey disabled caching of trait solving (he made the query transparent). The PR proposes a middle ground: do cache trait solving, but only for the same revision. This allows us to be safe because during a revision the inputs cannot change. The result is hopefully much better performance to features that tend to do a bulk of trait solving, and also repeat the same query (e.g. inference then IDE features). There is another limitation: results are only cached in the same thread, to remove the need for synchronization which will be expensive. More measurements are required to check whether it's better to use a synchronized global cache, or maybe stay with a thread-local cache but batch multiple feature requests (highlighting, inlay hints etc.) of the same file to the same thread. Alongside the actual cache we store the revision, because we need to verify it (we can't eagerly clear caches when incrementing the revision), and also the address of the db to prevent multiple dbs from interleaving (this is mostly relevant in tests, although injected highlighting also uses a new db, therefore maybe it's better to move it to a separate thread). This "games" analysis-stats to both be way faster and use way more memory; the former is because analysis-stats doesn't increment revisions, therefore all queries share the cache and hit ratio is way too good, the latter is because analysis-stats doesn't increment revisions and therefore the cache isn't cleared. Both are not representative of a typical IDE scenario.	2025-08-25 17:17:31 +03:00
Shoyu Vanilla	15ac6a21dd	fix: Make lang items query properly filter out overwritten/excluded sysroots	2025-08-17 22:28:04 +09:00
Deadbeef	d9f67325ab	fix errors after rebase	2025-08-13 15:33:08 +08:00
Deadbeef	82f174fbd9	Merge Trait and TraitAlias handling	2025-08-13 15:28:08 +08:00
jackh726	9418a3f2df	Implement next trait solver	2025-08-09 16:08:58 +00:00
Lukas Wirth	8ce30264c8	cargo clippy --fix	2025-07-31 10:55:10 +02:00
Lukas Wirth	97593ea9c6	Merge pull request #20289 from ChayimFriedman2/expr-store-diags-macros ... internal: Remove `ExpressionStoreDiagnostics::MacroError`, instead recreate it from the `MacroCallId`	2025-07-23 13:58:10 +00:00
Chayim Refael Friedman	963a1317e7	Remove ExpressionStoreDiagnostics::MacroError, instead recreate it from the MacroCallId ... This simplifies the code and also makes us report parse error in macros too.	2025-07-23 16:27:07 +03:00
Chayim Refael Friedman	c7ceb39f67	Parse for<'a> [const] ... And also refactor parsing of HRTB.	2025-07-22 16:24:42 +03:00
John Paul Adrian Glaubitz	25122d9289	hir-def: Don't apply x86_64-specific asserts on x32 ... This fixes the rustc build on x32 for which struct sizes differ.	2025-07-21 17:03:07 +02:00
Chayim Refael Friedman	b1914e420f	Support cfg_select!	2025-07-20 11:56:45 +03:00
Andrey Nikolaev	2b861a7410	remove now useless [allow(unused_lifetimes)]	2025-07-13 05:43:05 -04:00
Chayim Refael Friedman	4b29a9d6ea	Merge pull request #20219 from ChayimFriedman2/expr-store-mem ... perf: Put the expression stuff in the expression store behind an `Option<Box>`	2025-07-11 14:25:39 +00:00
Shoyu Vanilla (Flint)	e9968fc555	Merge pull request #20210 from ChayimFriedman2/naked-asm-safe ... fix: Inline asm fixes	2025-07-10 06:28:49 +00:00
Chayim Refael Friedman	ad708fdbd1	Put the expression stuff in the expression store behind an Option<Box> ... And leave only the type stuff without it. This is because most expression stores don't have anything but types (e.g. generics, fields, signatures) so this saves a lot of memory. This saves 58mb on `analysis-stats .`.	2025-07-10 05:18:43 +03:00
Chayim Refael Friedman	95c04c4503	Make global_asm!() work ... Because apparently, we were not accepting inline asm in item position, completely breaking it.	2025-07-09 18:55:27 +03:00
Chayim Refael Friedman	bd8087e86e	Differentiate between asm!(), global_asm!() and naked_asm!(), and make only asm!() unsafe	2025-07-09 17:37:27 +03:00
Chayim Refael Friedman	edb804a100	Don't hash the SyntaxKind in the ast id ... There is no need to, it's already stored in the `kind`.	2025-07-09 17:21:45 +03:00
Chayim Refael Friedman	87cddda2d8	Always bump in the parser in err_and_bump() ... Even when at curly braces, otherwise the parser can get stuck. This has happened in the past in #18625, but it was just worked around instead of handling the root of the problem. Now this happened again in #20171. IMO we can't let `err_and_bump()` not bump, that's too confusing and invites errors. We can (as I did) workaround the worse recovery instead.	2025-07-06 03:21:43 +03:00
Lukas Wirth	638e49b0fc	Skip unnecessary eq work in BodySourceMap	2025-07-04 11:36:06 +02:00
Chayim Refael Friedman	9c4a7705b1	Fix some things with builtin derives ... 1. Err on unions on derive where it's required. 2. Err on `#[derive(Default)]` on enums without `#[default]` variant. 3. Don't add where bounds `T: Default` when expanding `Default` on enums (structs need that, enums not). Also, because I was annoyed by that, in minicore, add a way to filter on multiple flags in the line-filter (`// :`). This is required for the `Debug` and `Hash` derives, because the derive should be in the prelude but the trait not.	2025-07-03 23:05:56 +03:00
Lukas Wirth	8029c731ed	Bump salsa	2025-07-03 10:05:16 +02:00
Chayim Refael Friedman	a6c1fa01d9	Merge pull request #20100 from ShoyuVanilla/ignore-sized-hierarchy ... Backport new sized-hierarchy trait bounds in old ways	2025-06-26 20:37:18 +00:00
Shoyu Vanilla	953e9d1c36	Backport new sized-hierarchy trait bounds in old ways	2025-06-26 23:22:26 +09:00
Lukas Wirth	468726178f	Bring back the firewall query	2025-06-26 14:12:57 +02:00
Lukas Wirth	889d84a1be	Salsa idiomize VariantFields query	2025-06-26 12:55:55 +02:00
Lukas Wirth	ad3a2d7280	Merge pull request #20088 from Veykril/push-qrslvzzlqqoz ... De-arc trait items query	2025-06-25 08:19:32 +00:00
Lukas Wirth	f25912c6f9	De-arc trait items query	2025-06-25 10:09:21 +02:00
Lukas Wirth	c43e7c71de	Short circuit a couple hir-ty/lower queries	2025-06-24 20:04:43 +02:00
Lukas Wirth	95dce2be51	Merge pull request #20062 from ChayimFriedman2/doctests ... minor: Don't run doctests	2025-06-24 06:47:24 +00:00
Shoyu Vanilla	254c6ec8e1	fix: Use ROOT hygiene for args inside new format_args! expansion	2025-06-24 01:10:32 +09:00
Chayim Refael Friedman	de312d0c71	Don't run doctests	2025-06-23 00:50:22 +03:00
Shoyu Vanilla	4f8767d790	Minic rustc's new format_args! expansion	2025-06-22 13:22:28 +09:00
Lukas Wirth	6301c35cf3	Revert "Turn BlockId into a #[salsa::tracked]" ... This reverts commit 8643a858db.	2025-06-19 07:40:01 +02:00
Chayim Refael Friedman	a207299344	Merge pull request #20000 from tadeokondrak/lifetime-repeat-macro ... Allow lifetime repeats in macros: $($x)'a*	2025-06-15 16:34:42 +00:00
Tadeo Kondrak	2070e9a8f3	Use is_any_identifier in pretty_print_macro_expansion	2025-06-15 18:23:18 +02:00
Tadeo Kondrak	12226b704a	Fix spacing for LIFETIME_IDENT near keywords and literals in test output	2025-06-15 17:43:45 +02:00
Lukas Wirth	b3768cdc0e	Optimize private visibility resolution	2025-06-15 10:25:45 +02:00
Lukas Wirth	e129cdc202	Optimize pub(crate) visibility resolution	2025-06-15 10:25:45 +02:00
Lukas Wirth	4b38ea5abd	Unify assoc item visibility queries	2025-06-15 10:01:34 +02:00
Lukas Wirth	bceeb6a3c7	Idiomatic salsa use for extern block abi query	2025-06-15 10:01:04 +02:00
Lukas Wirth	4e392f82cb	Idiomatic salsa use for impl items query	2025-06-15 09:40:02 +02:00
Lukas Wirth	155ef3de60	Replace unnecessary use of MacroDefId with MacroId	2025-06-15 09:16:45 +02:00
Lukas Wirth	aa2d234426	Idiomatic salsa use for enum variants query	2025-06-15 09:03:44 +02:00
Lukas Wirth	f68512af65	Cleanup incremental tests	2025-06-15 08:07:49 +02:00
Tadeo Kondrak	a7c09532a0	Allow lifetime repeats in macros: $($x)'a* ... This works in rustc. This change isn't motivated by any real code. I just learned about it and decided to see why it doesn't work with rust-analyzer.	2025-06-13 18:13:01 -06:00
Lukas Wirth	8643a858db	Turn BlockId into a #[salsa::tracked]	2025-06-13 17:30:04 +02:00
Lukas Wirth	9f051ee104	Use ThinVec in ItemScope in a couple places	2025-06-13 15:34:03 +02:00
Lukas Wirth	b9ce647cf1	Bring back EMPTY item tree deduplication	2025-06-13 14:51:23 +02:00
Lukas Wirth	bd002fe65c	Merge pull request #19982 from Veykril/push-uptnmqtlylsx ... Simplify and optimize `ItemTree`	2025-06-13 11:16:24 +00:00