rust/compiler/rustc_pattern_analysis/src, branch lcnr/rustc-dev-guidehttps://github.com/rust-lang/rust
http://git.dreamy.place/mirrors/rust/atom?h=lcnr/rustc-dev-guide2025-07-27T23:01:39+00:00In rustc_pattern_analysis, put `true` witnesses before `false` witnesses2025-07-27T23:01:39+00:00Chayim Refael Friedmanchayimfr@gmail.com2025-07-27T21:01:28+00:00urn:sha1:6bf3cbe39e09f4d9187a0b1f6a7bd45101f7aad8
In rustc it doesn't really matter what the order of the witnesses is, but I'm planning to use the witnesses for implementing the "add missing match arms" assist in rust-analyzer, and there `true` before `false` is the natural order (like `Some` before `None`), and also what the current assist does.
The current order doesn't seem to be intentional; the code was created when bool ctors became their own thing, not just int ctors, but for integer, 0 before 1 is indeed the natural order.
Rollup merge of #144171 - Nadrieril:exhaustive-witnesses, r=davidtwco2025-07-26T06:15:04+00:00Trevor Grosst.gross35@gmail.com2025-07-26T06:15:04+00:00urn:sha1:3f7d497c8aba1b415c2b9b6a597eab1faa744416
pattern_analysis: add option to get a full set of witnesses
This adds an option to the rustc_pattern_analysis machinery to have it report a complete set of patterns when a match is non-exhaustive (by default we only guarantee to report _some_ missing patterns). This is for use in rust-analyzer.
Leaving as draft until I'm sure this is what r-a needs.
r? ghost
List all the variants of non-exhaustive enums in exhaustive mode2025-07-20T16:27:30+00:00Nadrierilnadrieril+git@gmail.com2025-07-19T19:18:14+00:00urn:sha1:9b01de20e10376d379ae32baa6a315d8e30cb351pattern_analysis: add option to get a full set of witnesses2025-07-20T16:27:30+00:00Nadrierilnadrieril+git@gmail.com2025-07-19T09:28:31+00:00urn:sha1:2bb00741d463143a10e632bb118d4ed336dbb75fDon't consider unstable fields always-inhabited2025-07-20T16:23:18+00:00Nadrierilnadrieril+git@gmail.com2025-07-20T12:28:58+00:00urn:sha1:3567ab19a0edb4f6c7ca88157d84caa3e1564326
This reverts the hack in https://github.com/rust-lang/rust/pull/133889
now that `Pin`'s field is no longer public.
only check for mixed deref/normal constructors when needed2025-07-05T06:47:31+00:00diannediannes.gm@gmail.com2025-07-05T04:53:03+00:00urn:sha1:bb643159786c789c0b64f819d36a0979c63f9791always check for mixed deref pattern and normal constructors2025-07-05T06:47:31+00:00diannediannes.gm@gmail.com2025-07-05T04:17:40+00:00urn:sha1:50061f3b11f51d7a6e3acd8ce793a1f17f99b597
This makes it work for box patterns and in rust-analyzer.
treat box patterns as deref patterns in THIR and usefulness analysis2025-07-04T08:28:35+00:00diannediannes.gm@gmail.com2025-07-04T08:23:36+00:00urn:sha1:98659a339dd7203e0f23871b327b9305940e0d61
This removes special-casing of boxes from `rustc_pattern_analysis`, as a
first step in replacing `box_patterns` with `deref_patterns`.
Incidentally, it fixes a bug caused by box patterns being represented as
structs rather than pointers, where `exhaustive_patterns` could generate
spurious `unreachable_patterns` lints on arms required for
exhaustiveness; following the lint's advice would result in an error.
setup CI and tidy to use typos for spellchecking and fix few typos2025-07-03T07:51:06+00:00klensyklensy@users.noreply.github.com2024-12-07T10:51:08+00:00urn:sha1:c76d032f0144b650a438ee1efba89c475e0b115bAdd `#[loop_match]` for improved DFA codegen2025-06-23T18:43:04+00:00bjorn317426603+bjorn3@users.noreply.github.com2025-02-18T13:16:57+00:00urn:sha1:ba5556d239c11232dc8d95123ea70a2783019476
Co-authored-by: Folkert de Vries <folkert@folkertdev.nl>