mikros/rust
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
rust/src/test/mir-opt/simplify_arm.id_try.SimplifyArmIdentity.diff

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

95 lines
6.2 KiB
Diff
Raw Normal View History

Modify SimplifyArmIdentity so it can trigger on mir-opt-level=1 I also added test cases to make sure the optimization can fire on all of these cases: ```rust fn case_1(o: Option<u8>) -> Option<u8> { match o { Some(u) => Some(u), None => None, } } fn case2(r: Result<u8, i32>) -> Result<u8, i32> { match r { Ok(u) => Ok(u), Err(i) => Err(i), } } fn case3(r: Result<u8, i32>) -> Result<u8, i32> { let u = r?; Ok(u) } ``` Without MIR inlining, this still does not completely optimize away the `?` operator because the `Try::into_result()`, `From::from()` and `Try::from_error()` calls still exist. This does move us a bit closer to that goal though because: - We can now run the pass on mir-opt-level=1 - We no longer depend on the copy propagation pass running which is unlikely to stabilize anytime soon.
2020-05-11 20:13:15 -04:00
- // MIR for `id_try` before SimplifyArmIdentity
+ // MIR for `id_try` after SimplifyArmIdentity
Fix tests from rebase
2020-09-07 10:01:45 +01:00
path trimming: ignore type aliases
2021-01-28 18:01:36 +02:00
fn id_try(_1: Result<u8, i32>) -> Result<u8, i32> {
mir-opt & codegen test updates `SimplifyArm` and such are currently in `-Zunsound-mir-opts` and thus weren't running by default, so having something like them for the new desugar shouldn't be necessary for switching.
2021-04-25 22:02:48 -07:00
debug r => _1; // in scope 0 at $DIR/simplify-arm.rs:35:11: 35:12
let mut _0: std::result::Result<u8, i32>; // return place in scope 0 at $DIR/simplify-arm.rs:35:34: 35:49
let _2: u8; // in scope 0 at $DIR/simplify-arm.rs:36:9: 36:10
let mut _3: std::result::Result<u8, i32>; // in scope 0 at $DIR/simplify-arm.rs:36:19: 36:33
let mut _4: std::result::Result<u8, i32>; // in scope 0 at $DIR/simplify-arm.rs:36:31: 36:32
let mut _5: isize; // in scope 0 at $DIR/simplify-arm.rs:37:9: 37:15
let _6: i32; // in scope 0 at $DIR/simplify-arm.rs:37:13: 37:14
let mut _7: !; // in scope 0 at $DIR/simplify-arm.rs:37:19: 37:51
let mut _8: i32; // in scope 0 at $DIR/simplify-arm.rs:37:37: 37:50
let mut _9: i32; // in scope 0 at $DIR/simplify-arm.rs:37:48: 37:49
let _10: u8; // in scope 0 at $DIR/simplify-arm.rs:38:12: 38:13
let mut _11: u8; // in scope 0 at $DIR/simplify-arm.rs:40:8: 40:9
Modify SimplifyArmIdentity so it can trigger on mir-opt-level=1 I also added test cases to make sure the optimization can fire on all of these cases: ```rust fn case_1(o: Option<u8>) -> Option<u8> { match o { Some(u) => Some(u), None => None, } } fn case2(r: Result<u8, i32>) -> Result<u8, i32> { match r { Ok(u) => Ok(u), Err(i) => Err(i), } } fn case3(r: Result<u8, i32>) -> Result<u8, i32> { let u = r?; Ok(u) } ``` Without MIR inlining, this still does not completely optimize away the `?` operator because the `Try::into_result()`, `From::from()` and `Try::from_error()` calls still exist. This does move us a bit closer to that goal though because: - We can now run the pass on mir-opt-level=1 - We no longer depend on the copy propagation pass running which is unlikely to stabilize anytime soon.
2020-05-11 20:13:15 -04:00
scope 1 {
mir-opt & codegen test updates `SimplifyArm` and such are currently in `-Zunsound-mir-opts` and thus weren't running by default, so having something like them for the new desugar shouldn't be necessary for switching.
2021-04-25 22:02:48 -07:00
- debug x => _2; // in scope 1 at $DIR/simplify-arm.rs:36:9: 36:10
+ debug x => ((_0 as Ok).0: u8); // in scope 1 at $DIR/simplify-arm.rs:36:9: 36:10
Modify SimplifyArmIdentity so it can trigger on mir-opt-level=1 I also added test cases to make sure the optimization can fire on all of these cases: ```rust fn case_1(o: Option<u8>) -> Option<u8> { match o { Some(u) => Some(u), None => None, } } fn case2(r: Result<u8, i32>) -> Result<u8, i32> { match r { Ok(u) => Ok(u), Err(i) => Err(i), } } fn case3(r: Result<u8, i32>) -> Result<u8, i32> { let u = r?; Ok(u) } ``` Without MIR inlining, this still does not completely optimize away the `?` operator because the `Try::into_result()`, `From::from()` and `Try::from_error()` calls still exist. This does move us a bit closer to that goal though because: - We can now run the pass on mir-opt-level=1 - We no longer depend on the copy propagation pass running which is unlikely to stabilize anytime soon.
2020-05-11 20:13:15 -04:00
}
scope 2 {
mir-opt & codegen test updates `SimplifyArm` and such are currently in `-Zunsound-mir-opts` and thus weren't running by default, so having something like them for the new desugar shouldn't be necessary for switching.
2021-04-25 22:02:48 -07:00
- debug e => _6; // in scope 2 at $DIR/simplify-arm.rs:37:13: 37:14
+ debug e => ((_0 as Err).0: i32); // in scope 2 at $DIR/simplify-arm.rs:37:13: 37:14
scope 5 (inlined <i32 as From<i32>>::from) { // at $DIR/simplify-arm.rs:37:37: 37:50
Only create a single expansion for each inline integration.
2022-02-27 16:09:52 +01:00
- debug t => _9; // in scope 5 at $SRC_DIR/core/src/convert/mod.rs:LL:COL
+ debug t => ((_0 as Err).0: i32); // in scope 5 at $SRC_DIR/core/src/convert/mod.rs:LL:COL
Modify SimplifyArmIdentity so it can trigger on mir-opt-level=1 I also added test cases to make sure the optimization can fire on all of these cases: ```rust fn case_1(o: Option<u8>) -> Option<u8> { match o { Some(u) => Some(u), None => None, } } fn case2(r: Result<u8, i32>) -> Result<u8, i32> { match r { Ok(u) => Ok(u), Err(i) => Err(i), } } fn case3(r: Result<u8, i32>) -> Result<u8, i32> { let u = r?; Ok(u) } ``` Without MIR inlining, this still does not completely optimize away the `?` operator because the `Try::into_result()`, `From::from()` and `Try::from_error()` calls still exist. This does move us a bit closer to that goal though because: - We can now run the pass on mir-opt-level=1 - We no longer depend on the copy propagation pass running which is unlikely to stabilize anytime soon.
2020-05-11 20:13:15 -04:00
}
mir-opt & codegen test updates `SimplifyArm` and such are currently in `-Zunsound-mir-opts` and thus weren't running by default, so having something like them for the new desugar shouldn't be necessary for switching.
2021-04-25 22:02:48 -07:00
scope 6 (inlined from_error::<u8, i32>) { // at $DIR/simplify-arm.rs:37:26: 37:51
Only create a single expansion for each inline integration.
2022-02-27 16:09:52 +01:00
- debug e => _8; // in scope 6 at $DIR/simplify-arm.rs:27:21: 27:22
+ debug e => ((_0 as Err).0: i32); // in scope 6 at $DIR/simplify-arm.rs:27:21: 27:22
Modify SimplifyArmIdentity so it can trigger on mir-opt-level=1 I also added test cases to make sure the optimization can fire on all of these cases: ```rust fn case_1(o: Option<u8>) -> Option<u8> { match o { Some(u) => Some(u), None => None, } } fn case2(r: Result<u8, i32>) -> Result<u8, i32> { match r { Ok(u) => Ok(u), Err(i) => Err(i), } } fn case3(r: Result<u8, i32>) -> Result<u8, i32> { let u = r?; Ok(u) } ``` Without MIR inlining, this still does not completely optimize away the `?` operator because the `Try::into_result()`, `From::from()` and `Try::from_error()` calls still exist. This does move us a bit closer to that goal though because: - We can now run the pass on mir-opt-level=1 - We no longer depend on the copy propagation pass running which is unlikely to stabilize anytime soon.
2020-05-11 20:13:15 -04:00
}
}
mir-opt & codegen test updates `SimplifyArm` and such are currently in `-Zunsound-mir-opts` and thus weren't running by default, so having something like them for the new desugar shouldn't be necessary for switching.
2021-04-25 22:02:48 -07:00
scope 3 {
- debug v => _10; // in scope 3 at $DIR/simplify-arm.rs:38:12: 38:13
+ debug v => ((_0 as Ok).0: u8); // in scope 3 at $DIR/simplify-arm.rs:38:12: 38:13
}
scope 4 (inlined into_result::<u8, i32>) { // at $DIR/simplify-arm.rs:36:19: 36:33
Only create a single expansion for each inline integration.
2022-02-27 16:09:52 +01:00
debug r => _4; // in scope 4 at $DIR/simplify-arm.rs:23:22: 23:23
Implement 'considered equal' for statements, so that for example `_0 = _1` and `discriminant(_0) = discriminant(0)` are considered equal if 0 is a fieldless variant of an enum
2020-07-25 17:46:11 +02:00
}
Fix tests from rebase
2020-09-07 10:01:45 +01:00
Modify SimplifyArmIdentity so it can trigger on mir-opt-level=1 I also added test cases to make sure the optimization can fire on all of these cases: ```rust fn case_1(o: Option<u8>) -> Option<u8> { match o { Some(u) => Some(u), None => None, } } fn case2(r: Result<u8, i32>) -> Result<u8, i32> { match r { Ok(u) => Ok(u), Err(i) => Err(i), } } fn case3(r: Result<u8, i32>) -> Result<u8, i32> { let u = r?; Ok(u) } ``` Without MIR inlining, this still does not completely optimize away the `?` operator because the `Try::into_result()`, `From::from()` and `Try::from_error()` calls still exist. This does move us a bit closer to that goal though because: - We can now run the pass on mir-opt-level=1 - We no longer depend on the copy propagation pass running which is unlikely to stabilize anytime soon.
2020-05-11 20:13:15 -04:00
bb0: {
mir-opt & codegen test updates `SimplifyArm` and such are currently in `-Zunsound-mir-opts` and thus weren't running by default, so having something like them for the new desugar shouldn't be necessary for switching.
2021-04-25 22:02:48 -07:00
StorageLive(_2); // scope 0 at $DIR/simplify-arm.rs:36:9: 36:10
StorageLive(_3); // scope 0 at $DIR/simplify-arm.rs:36:19: 36:33
StorageLive(_4); // scope 0 at $DIR/simplify-arm.rs:36:31: 36:32
_4 = _1; // scope 0 at $DIR/simplify-arm.rs:36:31: 36:32
Only create a single expansion for each inline integration.
2022-02-27 16:09:52 +01:00
_3 = move _4; // scope 4 at $DIR/simplify-arm.rs:24:5: 24:6
mir-opt & codegen test updates `SimplifyArm` and such are currently in `-Zunsound-mir-opts` and thus weren't running by default, so having something like them for the new desugar shouldn't be necessary for switching.
2021-04-25 22:02:48 -07:00
StorageDead(_4); // scope 0 at $DIR/simplify-arm.rs:36:32: 36:33
Fix debugger stepping behavior around `match` expressions Previously, we would set up the source lines for `match` expressions so that the code generated to perform the test of the scrutinee was matched to the line of the arm that required the test and then jump from the arm block to the "next" block was matched to all of the lines in the `match` expression. While that makes sense, it has the side effect of causing strange stepping behavior in debuggers. I've changed the source information so that all of the generated tests are sourced to `match {scrutinee}` and the jumps are sourced to the last line of the block they are inside. This resolves the weird stepping behavior in all debuggers and resolves some instances of "ambiguous symbol" errors in WinDbg preventing the user from setting breakpoints at `match` expressions.
2021-07-23 18:55:36 -04:00
_5 = discriminant(_3); // scope 0 at $DIR/simplify-arm.rs:36:19: 36:33
switchInt(move _5) -> [0_isize: bb1, 1_isize: bb3, otherwise: bb2]; // scope 0 at $DIR/simplify-arm.rs:36:13: 36:33
Modify SimplifyArmIdentity so it can trigger on mir-opt-level=1 I also added test cases to make sure the optimization can fire on all of these cases: ```rust fn case_1(o: Option<u8>) -> Option<u8> { match o { Some(u) => Some(u), None => None, } } fn case2(r: Result<u8, i32>) -> Result<u8, i32> { match r { Ok(u) => Ok(u), Err(i) => Err(i), } } fn case3(r: Result<u8, i32>) -> Result<u8, i32> { let u = r?; Ok(u) } ``` Without MIR inlining, this still does not completely optimize away the `?` operator because the `Try::into_result()`, `From::from()` and `Try::from_error()` calls still exist. This does move us a bit closer to that goal though because: - We can now run the pass on mir-opt-level=1 - We no longer depend on the copy propagation pass running which is unlikely to stabilize anytime soon.
2020-05-11 20:13:15 -04:00
}
Fix tests from rebase
2020-09-07 10:01:45 +01:00
Implement 'considered equal' for statements, so that for example `_0 = _1` and `discriminant(_0) = discriminant(0)` are considered equal if 0 is a fieldless variant of an enum
2020-07-25 17:46:11 +02:00
bb1: {
mir-opt & codegen test updates `SimplifyArm` and such are currently in `-Zunsound-mir-opts` and thus weren't running by default, so having something like them for the new desugar shouldn't be necessary for switching.
2021-04-25 22:02:48 -07:00
- StorageLive(_10); // scope 0 at $DIR/simplify-arm.rs:38:12: 38:13
- _10 = ((_3 as Ok).0: u8); // scope 0 at $DIR/simplify-arm.rs:38:12: 38:13
- _2 = _10; // scope 3 at $DIR/simplify-arm.rs:38:18: 38:19
- StorageDead(_10); // scope 0 at $DIR/simplify-arm.rs:38:18: 38:19
+ _0 = move _3; // scope 1 at $DIR/simplify-arm.rs:40:5: 40:10
StorageDead(_3); // scope 0 at $DIR/simplify-arm.rs:39:6: 39:7
- StorageLive(_11); // scope 1 at $DIR/simplify-arm.rs:40:8: 40:9
- _11 = _2; // scope 1 at $DIR/simplify-arm.rs:40:8: 40:9
- ((_0 as Ok).0: u8) = move _11; // scope 1 at $DIR/simplify-arm.rs:40:5: 40:10
- discriminant(_0) = 0; // scope 1 at $DIR/simplify-arm.rs:40:5: 40:10
- StorageDead(_11); // scope 1 at $DIR/simplify-arm.rs:40:9: 40:10
StorageDead(_2); // scope 0 at $DIR/simplify-arm.rs:41:1: 41:2
goto -> bb4; // scope 0 at $DIR/simplify-arm.rs:41:2: 41:2
Modify SimplifyArmIdentity so it can trigger on mir-opt-level=1 I also added test cases to make sure the optimization can fire on all of these cases: ```rust fn case_1(o: Option<u8>) -> Option<u8> { match o { Some(u) => Some(u), None => None, } } fn case2(r: Result<u8, i32>) -> Result<u8, i32> { match r { Ok(u) => Ok(u), Err(i) => Err(i), } } fn case3(r: Result<u8, i32>) -> Result<u8, i32> { let u = r?; Ok(u) } ``` Without MIR inlining, this still does not completely optimize away the `?` operator because the `Try::into_result()`, `From::from()` and `Try::from_error()` calls still exist. This does move us a bit closer to that goal though because: - We can now run the pass on mir-opt-level=1 - We no longer depend on the copy propagation pass running which is unlikely to stabilize anytime soon.
2020-05-11 20:13:15 -04:00
}
Fix tests from rebase
2020-09-07 10:01:45 +01:00
Implement 'considered equal' for statements, so that for example `_0 = _1` and `discriminant(_0) = discriminant(0)` are considered equal if 0 is a fieldless variant of an enum
2020-07-25 17:46:11 +02:00
bb2: {
mir-opt & codegen test updates `SimplifyArm` and such are currently in `-Zunsound-mir-opts` and thus weren't running by default, so having something like them for the new desugar shouldn't be necessary for switching.
2021-04-25 22:02:48 -07:00
unreachable; // scope 0 at $DIR/simplify-arm.rs:36:19: 36:33
Modify SimplifyArmIdentity so it can trigger on mir-opt-level=1 I also added test cases to make sure the optimization can fire on all of these cases: ```rust fn case_1(o: Option<u8>) -> Option<u8> { match o { Some(u) => Some(u), None => None, } } fn case2(r: Result<u8, i32>) -> Result<u8, i32> { match r { Ok(u) => Ok(u), Err(i) => Err(i), } } fn case3(r: Result<u8, i32>) -> Result<u8, i32> { let u = r?; Ok(u) } ``` Without MIR inlining, this still does not completely optimize away the `?` operator because the `Try::into_result()`, `From::from()` and `Try::from_error()` calls still exist. This does move us a bit closer to that goal though because: - We can now run the pass on mir-opt-level=1 - We no longer depend on the copy propagation pass running which is unlikely to stabilize anytime soon.
2020-05-11 20:13:15 -04:00
}
Fix tests from rebase
2020-09-07 10:01:45 +01:00
Implement 'considered equal' for statements, so that for example `_0 = _1` and `discriminant(_0) = discriminant(0)` are considered equal if 0 is a fieldless variant of an enum
2020-07-25 17:46:11 +02:00
bb3: {
mir-opt & codegen test updates `SimplifyArm` and such are currently in `-Zunsound-mir-opts` and thus weren't running by default, so having something like them for the new desugar shouldn't be necessary for switching.
2021-04-25 22:02:48 -07:00
- StorageLive(_6); // scope 0 at $DIR/simplify-arm.rs:37:13: 37:14
- _6 = ((_3 as Err).0: i32); // scope 0 at $DIR/simplify-arm.rs:37:13: 37:14
- StorageLive(_8); // scope 2 at $DIR/simplify-arm.rs:37:37: 37:50
- StorageLive(_9); // scope 2 at $DIR/simplify-arm.rs:37:48: 37:49
- _9 = _6; // scope 2 at $DIR/simplify-arm.rs:37:48: 37:49
Only create a single expansion for each inline integration.
2022-02-27 16:09:52 +01:00
- _8 = move _9; // scope 5 at $SRC_DIR/core/src/convert/mod.rs:LL:COL
mir-opt & codegen test updates `SimplifyArm` and such are currently in `-Zunsound-mir-opts` and thus weren't running by default, so having something like them for the new desugar shouldn't be necessary for switching.
2021-04-25 22:02:48 -07:00
- StorageDead(_9); // scope 2 at $DIR/simplify-arm.rs:37:49: 37:50
Only create a single expansion for each inline integration.
2022-02-27 16:09:52 +01:00
- ((_0 as Err).0: i32) = move _8; // scope 6 at $DIR/simplify-arm.rs:28:9: 28:10
- discriminant(_0) = 1; // scope 6 at $DIR/simplify-arm.rs:28:5: 28:11
mir-opt & codegen test updates `SimplifyArm` and such are currently in `-Zunsound-mir-opts` and thus weren't running by default, so having something like them for the new desugar shouldn't be necessary for switching.
2021-04-25 22:02:48 -07:00
- StorageDead(_8); // scope 2 at $DIR/simplify-arm.rs:37:50: 37:51
- StorageDead(_6); // scope 0 at $DIR/simplify-arm.rs:37:50: 37:51
Only create a single expansion for each inline integration.
2022-02-27 16:09:52 +01:00
+ _0 = move _3; // scope 6 at $DIR/simplify-arm.rs:28:5: 28:11
mir-opt & codegen test updates `SimplifyArm` and such are currently in `-Zunsound-mir-opts` and thus weren't running by default, so having something like them for the new desugar shouldn't be necessary for switching.
2021-04-25 22:02:48 -07:00
StorageDead(_3); // scope 0 at $DIR/simplify-arm.rs:39:6: 39:7
StorageDead(_2); // scope 0 at $DIR/simplify-arm.rs:41:1: 41:2
goto -> bb4; // scope 0 at $DIR/simplify-arm.rs:41:2: 41:2
Modify SimplifyArmIdentity so it can trigger on mir-opt-level=1 I also added test cases to make sure the optimization can fire on all of these cases: ```rust fn case_1(o: Option<u8>) -> Option<u8> { match o { Some(u) => Some(u), None => None, } } fn case2(r: Result<u8, i32>) -> Result<u8, i32> { match r { Ok(u) => Ok(u), Err(i) => Err(i), } } fn case3(r: Result<u8, i32>) -> Result<u8, i32> { let u = r?; Ok(u) } ``` Without MIR inlining, this still does not completely optimize away the `?` operator because the `Try::into_result()`, `From::from()` and `Try::from_error()` calls still exist. This does move us a bit closer to that goal though because: - We can now run the pass on mir-opt-level=1 - We no longer depend on the copy propagation pass running which is unlikely to stabilize anytime soon.
2020-05-11 20:13:15 -04:00
}
Fix tests from rebase
2020-09-07 10:01:45 +01:00
Implement 'considered equal' for statements, so that for example `_0 = _1` and `discriminant(_0) = discriminant(0)` are considered equal if 0 is a fieldless variant of an enum
2020-07-25 17:46:11 +02:00
bb4: {
mir-opt & codegen test updates `SimplifyArm` and such are currently in `-Zunsound-mir-opts` and thus weren't running by default, so having something like them for the new desugar shouldn't be necessary for switching.
2021-04-25 22:02:48 -07:00
return; // scope 0 at $DIR/simplify-arm.rs:41:2: 41:2
further bless tests.
2020-06-04 12:26:13 -04:00
}
Modify SimplifyArmIdentity so it can trigger on mir-opt-level=1 I also added test cases to make sure the optimization can fire on all of these cases: ```rust fn case_1(o: Option<u8>) -> Option<u8> { match o { Some(u) => Some(u), None => None, } } fn case2(r: Result<u8, i32>) -> Result<u8, i32> { match r { Ok(u) => Ok(u), Err(i) => Err(i), } } fn case3(r: Result<u8, i32>) -> Result<u8, i32> { let u = r?; Ok(u) } ``` Without MIR inlining, this still does not completely optimize away the `?` operator because the `Try::into_result()`, `From::from()` and `Try::from_error()` calls still exist. This does move us a bit closer to that goal though because: - We can now run the pass on mir-opt-level=1 - We no longer depend on the copy propagation pass running which is unlikely to stabilize anytime soon.
2020-05-11 20:13:15 -04:00
}
Fix tests from rebase
2020-09-07 10:01:45 +01:00
Reference in New Issue Copy Permalink