Improve MIR match generation for ranges #56810

sinkuu · 2018-12-14T10:11:11Z

Improves MIR match generation to rule out ranges/values distinct from the range that has been tested. e.g., for this code:

match x {
    0..=5 if b => 0,
    6..=10 => 1,
    _ => 2,
}

MIR (before):

bb0: { ...; _4 = Le(const 0i32, _1); switchInt(move _4) -> [false: bb6, otherwise: bb5]; }
bb1: { _3 = const 0i32; goto -> bb8; }
bb2: { _6 = _2; switchInt(move _6) -> [false: bb6, otherwise: bb1]; } // If `!b`, jumps to test if `6 <= x <= 10`.
bb3: { _3 = const 1i32; goto -> bb8; }
bb4: { _3 = const 2i32; goto -> bb8; }
bb5: { _5 = Le(_1, const 5i32); switchInt(move _5) -> [false: bb6, otherwise: bb2]; }
bb6: { _7 = Le(const 6i32, _1); switchInt(move _7) -> [false: bb4, otherwise: bb7]; }
bb7: { _8 = Le(_1, const 10i32); switchInt(move _8) -> [false: bb4, otherwise: bb3]; }

MIR (after):

bb0: { ...; _4 = Le(const 0i32, _1); switchInt(move _4) -> [false: bb5, otherwise: bb6]; }
bb1: { _3 = const 0i32; goto -> bb8; }
bb2: { _6 = _2; switchInt(move _6) -> [false: bb4, otherwise: bb1]; } // If `!b`, jumps to `_ =>` arm.
bb3: { _3 = const 1i32; goto -> bb8; }
bb4: { _3 = const 2i32; goto -> bb8; }
bb5: { _7 = Le(const 6i32, _1); switchInt(move _7) -> [false: bb4, otherwise: bb7]; }
bb6: { _5 = Le(_1, const 5i32); switchInt(move _5) -> [false: bb5, otherwise: bb2]; }
bb7: { _8 = Le(_1, const 10i32); switchInt(move _8) -> [false: bb4, otherwise: bb3]; }

cc #29623

rust-highfive · 2018-12-14T10:11:14Z

r? @cramertj

(rust_highfive has picked a reviewer for you, use r? to override)

bors · 2018-12-14T18:52:21Z

☔ The latest upstream changes (presumably #56818) made this pull request unmergeable. Please resolve the merge conflicts.

cramertj · 2018-12-14T19:30:41Z

r? @nikomatsakis

Makes testing a range rule out ranges/constant covered by the range that is being tested

matthewjasper · 2018-12-15T10:27:27Z

Do we generate a false edge from the !b path to the test for 6..=10 now?

sinkuu · 2018-12-15T11:06:51Z

Imaginary part of a false edge always points to Candidate.next_candidate_pre_binding_block, which is populated in Builder::match_expr, so it won't be affected.

Essential^* diff of SimplifyBranch-initial.before for that example:

@@ -58,7 +58,7 @@
         switchInt(move _6) -> [false: bb10, otherwise: bb1];
     }
     bb10: {
-        falseEdges -> [real: bb11, imaginary: bb5];
+        falseEdges -> [real: bb6, imaginary: bb5];
     }
     bb11: {
         _7 = Le(const 6i32, _1);

(* this PR changes the order of BBs, so I swapped a few of them to reduce amount of diff)

oli-obk

I think it would be a good idea to add a mir-opt test ensuring that we don't regress this in the future, but these tests are such a pain to keep updated that we should punt this to the future.

This is probably not going to affect the perf tests as it's unlikely that we have any doing such range matches in a perf critical path. Have you done any benchmarks locally? Just curious, does not affect this PR at all.

src/librustc_mir/build/matches/test.rs

oli-obk · 2018-12-15T12:16:07Z

src/librustc_mir/build/matches/test.rs

+                    return true;
+                }
+
+                let no_overlap = (|| {


we can use the try syntax inside the compiler.

try syntax seems to require not only #![feature(try_block)] but also --edition 2018 for try keyword.

sinkuu · 2018-12-15T14:36:38Z

Have you done any benchmarks locally?

No.

rust-highfive · 2018-12-17T08:09:20Z

The job x86_64-gnu-llvm-6.0 of your PR failed on Travis (raw log). Through arcane magic we have determined that the following fragments from the build log may contain information about the problem.

Click to expand the log.

travis_time:end:19fde8d3:start=1545033388039394459,finish=1545033444296127243,duration=56256732784
$ git checkout -qf FETCH_HEAD
travis_fold:end:git.checkout

Encrypted environment variables have been removed for security reasons.
See https://docs.travis-ci.com/user/pull-requests/#pull-requests-and-security-restrictions
$ export SCCACHE_BUCKET=rust-lang-ci-sccache2
$ export SCCACHE_REGION=us-west-1
Setting environment variables from .travis.yml
$ export IMAGE=x86_64-gnu-llvm-6.0
---

[00:03:08] travis_fold:start:tidy
travis_time:start:tidy
tidy check
[00:03:08] tidy error: /checkout/src/test/mir-opt/match_test.rs: missing trailing newline
[00:03:10] some tidy checks failed
[00:03:10] 
[00:03:10] 
[00:03:10] command did not execute successfully: "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0-tools-bin/tidy" "/checkout/src" "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0/bin/cargo" "--no-vendor" "--quiet"
[00:03:10] 
[00:03:10] 
[00:03:10] failed to run: /checkout/obj/build/bootstrap/debug/bootstrap test src/tools/tidy
[00:03:10] Build completed unsuccessfully in 0:00:49
[00:03:10] Build completed unsuccessfully in 0:00:49
[00:03:10] make: *** [tidy] Error 1
[00:03:10] Makefile:79: recipe for target 'tidy' failed
The command "stamp sh -x -c "$RUN_SCRIPT"" exited with 2.
travis_time:start:1ca91566
$ date && (curl -fs --head https://google.com | grep ^Date: | sed 's/Date: //g' || true)
Mon Dec 17 08:00:43 UTC 2018
---
travis_time:end:05f89200:start=1545033643754175432,finish=1545033643760154801,duration=5979369
travis_fold:end:after_failure.3
travis_fold:start:after_failure.4
travis_time:start:0a5425a0
$ ln -s . checkout && for CORE in obj/cores/core.*; do EXE=$(echo $CORE | sed 's|obj/cores/core\.[0-9]*\.!checkout!\(.*\)|\1|;y|!|/|'); if [ -f "$EXE" ]; then printf travis_fold":start:crashlog\n\033[31;1m%s\033[0m\n" "$CORE"; gdb --batch -q -c "$CORE" "$EXE" -iex 'set auto-load off' -iex 'dir src/' -iex 'set sysroot .' -ex bt -ex q; echo travis_fold":"end:crashlog; fi; done || true
travis_fold:end:after_failure.4
travis_fold:start:after_failure.5
travis_time:start:098aab04
travis_time:start:098aab04
$ cat ./obj/build/x86_64-unknown-linux-gnu/native/asan/build/lib/asan/clang_rt.asan-dynamic-i386.vers || true
cat: ./obj/build/x86_64-unknown-linux-gnu/native/asan/build/lib/asan/clang_rt.asan-dynamic-i386.vers: No such file or directory
travis_fold:end:after_failure.5
travis_fold:start:after_failure.6
travis_time:start:22d5eff3
$ dmesg | grep -i kill

I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact @TimNN. (Feature Requests)

oli-obk · 2018-12-17T08:21:04Z

@bors try

bors · 2018-12-17T08:21:16Z

⌛ Trying commit d66a55e with merge f8b8ae3...

Improve MIR match generation for ranges Improves MIR match generation to rule out ranges/values distinct from the range that has been tested. e.g., for this code: ```rust match x { 0..=5 if b => 0, 6..=10 => 1, _ => 2, } ``` MIR (before): ```rust bb0: { ...; _4 = Le(const 0i32, _1); switchInt(move _4) -> [false: bb6, otherwise: bb5]; } bb1: { _3 = const 0i32; goto -> bb8; } bb2: { _6 = _2; switchInt(move _6) -> [false: bb6, otherwise: bb1]; } // If `!b`, jumps to test if `6 <= x <= 10`. bb3: { _3 = const 1i32; goto -> bb8; } bb4: { _3 = const 2i32; goto -> bb8; } bb5: { _5 = Le(_1, const 5i32); switchInt(move _5) -> [false: bb6, otherwise: bb2]; } bb6: { _7 = Le(const 6i32, _1); switchInt(move _7) -> [false: bb4, otherwise: bb7]; } bb7: { _8 = Le(_1, const 10i32); switchInt(move _8) -> [false: bb4, otherwise: bb3]; } ``` MIR (after): ```rust bb0: { ...; _4 = Le(const 0i32, _1); switchInt(move _4) -> [false: bb5, otherwise: bb6]; } bb1: { _3 = const 0i32; goto -> bb8; } bb2: { _6 = _2; switchInt(move _6) -> [false: bb4, otherwise: bb1]; } // If `!b`, jumps to `_ =>` arm. bb3: { _3 = const 1i32; goto -> bb8; } bb4: { _3 = const 2i32; goto -> bb8; } bb5: { _7 = Le(const 6i32, _1); switchInt(move _7) -> [false: bb4, otherwise: bb7]; } bb6: { _5 = Le(_1, const 5i32); switchInt(move _5) -> [false: bb5, otherwise: bb2]; } bb7: { _8 = Le(_1, const 10i32); switchInt(move _8) -> [false: bb4, otherwise: bb3]; } ``` cc #29623

oli-obk · 2018-12-17T08:21:37Z

@rust-timer build f8b8ae3

rust-timer · 2018-12-17T08:21:38Z

Success: Queued f8b8ae3 with parent 63f8e6e, comparison URL.

bors · 2018-12-17T10:42:11Z

☀️ Test successful - status-travis
State: approved= try=True

rust-timer · 2018-12-17T12:15:40Z

Finished benchmarking try commit f8b8ae3

oli-obk · 2018-12-17T12:25:17Z

@bors r+

bors · 2018-12-17T12:25:18Z

📌 Commit d66a55e has been approved by oli-obk

bors · 2018-12-17T13:26:46Z

⌛ Testing commit d66a55e with merge 54f3cd6...

Improve MIR match generation for ranges Improves MIR match generation to rule out ranges/values distinct from the range that has been tested. e.g., for this code: ```rust match x { 0..=5 if b => 0, 6..=10 => 1, _ => 2, } ``` MIR (before): ```rust bb0: { ...; _4 = Le(const 0i32, _1); switchInt(move _4) -> [false: bb6, otherwise: bb5]; } bb1: { _3 = const 0i32; goto -> bb8; } bb2: { _6 = _2; switchInt(move _6) -> [false: bb6, otherwise: bb1]; } // If `!b`, jumps to test if `6 <= x <= 10`. bb3: { _3 = const 1i32; goto -> bb8; } bb4: { _3 = const 2i32; goto -> bb8; } bb5: { _5 = Le(_1, const 5i32); switchInt(move _5) -> [false: bb6, otherwise: bb2]; } bb6: { _7 = Le(const 6i32, _1); switchInt(move _7) -> [false: bb4, otherwise: bb7]; } bb7: { _8 = Le(_1, const 10i32); switchInt(move _8) -> [false: bb4, otherwise: bb3]; } ``` MIR (after): ```rust bb0: { ...; _4 = Le(const 0i32, _1); switchInt(move _4) -> [false: bb5, otherwise: bb6]; } bb1: { _3 = const 0i32; goto -> bb8; } bb2: { _6 = _2; switchInt(move _6) -> [false: bb4, otherwise: bb1]; } // If `!b`, jumps to `_ =>` arm. bb3: { _3 = const 1i32; goto -> bb8; } bb4: { _3 = const 2i32; goto -> bb8; } bb5: { _7 = Le(const 6i32, _1); switchInt(move _7) -> [false: bb4, otherwise: bb7]; } bb6: { _5 = Le(_1, const 5i32); switchInt(move _5) -> [false: bb5, otherwise: bb2]; } bb7: { _8 = Le(_1, const 10i32); switchInt(move _8) -> [false: bb4, otherwise: bb3]; } ``` cc #29623

bors · 2018-12-17T16:01:39Z

☀️ Test successful - status-appveyor, status-travis
Approved by: oli-obk
Pushing 54f3cd6 to master...

rust-highfive assigned cramertj Dec 14, 2018

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Dec 14, 2018

sinkuu closed this Dec 14, 2018

This comment has been minimized.

Sign in to view

sinkuu reopened this Dec 14, 2018

sinkuu force-pushed the build_match branch from 8e25aaf to 27c8864 Compare December 14, 2018 10:17

sinkuu changed the title ~~Improve match MIR generation for ranges~~ Improve MIR match generation for ranges Dec 14, 2018

This comment has been minimized.

Sign in to view

sinkuu closed this Dec 14, 2018

sinkuu reopened this Dec 14, 2018

sinkuu force-pushed the build_match branch from 27c8864 to 856e3b1 Compare December 14, 2018 14:38

rust-highfive assigned nikomatsakis and unassigned cramertj Dec 14, 2018

sinkuu added 2 commits December 15, 2018 07:56

Cleanups

d2ac11c

Improve match MIR generation for ranges

9daa823

Makes testing a range rule out ranges/constant covered by the range that is being tested

sinkuu force-pushed the build_match branch from 856e3b1 to 9daa823 Compare December 14, 2018 23:40

oli-obk reviewed Dec 15, 2018

View reviewed changes

sinkuu added 2 commits December 15, 2018 23:14

Factor out

2c1b1c2

Add common struct for range

059bbd9

Add MIR test

0aab437

sinkuu force-pushed the build_match branch from 95594ae to 0aab437 Compare December 17, 2018 07:53

tidy

d66a55e

bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Dec 17, 2018

bors merged commit d66a55e into rust-lang:master Dec 17, 2018

This was referenced Dec 17, 2018

NLL: User type annotations refactor, associated constant patterns and ref bindings. #55937

Merged

Don't emit Unevaluated from const_eval #56723

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Improve MIR match generation for ranges #56810

Improve MIR match generation for ranges #56810

sinkuu commented Dec 14, 2018 •

edited

Loading

rust-highfive commented Dec 14, 2018

This comment has been minimized.

This comment has been minimized.

bors commented Dec 14, 2018

cramertj commented Dec 14, 2018

matthewjasper commented Dec 15, 2018

sinkuu commented Dec 15, 2018

oli-obk left a comment

oli-obk Dec 15, 2018

sinkuu Dec 15, 2018

sinkuu commented Dec 15, 2018

rust-highfive commented Dec 17, 2018

oli-obk commented Dec 17, 2018

bors commented Dec 17, 2018

oli-obk commented Dec 17, 2018

rust-timer commented Dec 17, 2018

bors commented Dec 17, 2018

rust-timer commented Dec 17, 2018

oli-obk commented Dec 17, 2018

bors commented Dec 17, 2018

bors commented Dec 17, 2018

bors commented Dec 17, 2018

Improve MIR match generation for ranges #56810

Improve MIR match generation for ranges #56810

Conversation

sinkuu commented Dec 14, 2018 • edited Loading

rust-highfive commented Dec 14, 2018

This comment has been minimized.

This comment has been minimized.

bors commented Dec 14, 2018

cramertj commented Dec 14, 2018

matthewjasper commented Dec 15, 2018

sinkuu commented Dec 15, 2018

oli-obk left a comment

Choose a reason for hiding this comment

oli-obk Dec 15, 2018

Choose a reason for hiding this comment

sinkuu Dec 15, 2018

Choose a reason for hiding this comment

sinkuu commented Dec 15, 2018

rust-highfive commented Dec 17, 2018

oli-obk commented Dec 17, 2018

bors commented Dec 17, 2018

oli-obk commented Dec 17, 2018

rust-timer commented Dec 17, 2018

bors commented Dec 17, 2018

rust-timer commented Dec 17, 2018

oli-obk commented Dec 17, 2018

bors commented Dec 17, 2018

bors commented Dec 17, 2018

bors commented Dec 17, 2018

sinkuu commented Dec 14, 2018 •

edited

Loading