Fix a couple of compile-fail tests #376

bjorn3 · 2018-05-09T14:06:08Z

failures:

---- [compile-fail] compile-fail/match_char.rs stdout ----

error: compile-fail test compiled successfully!
status: exit code: 0
command: "target/debug/miri" "tests/compile-fail/match_char.rs" "-L" "/tmp/compiletestpMJMuw" "--target=x86_64-unknown-linux-gnu" "--error-format" "json" "-C" "prefer-dynamic" "-o" "/tmp/compiletestpMJMuw/match_char.stage-id" "--sysroot" "/home/bjorn/.xargo/HOST" "-Zmir-emit-validate=1" "-L" "/tmp/compiletestpMJMuw/match_char.stage-id.aux" "-A" "unused"
stdout:
------------------------------------------

------------------------------------------
stderr:
------------------------------------------

------------------------------------------

thread '[compile-fail] compile-fail/match_char.rs' panicked at 'explicit panic', /home/bjorn/.cargo/registry/src/github.com-1ecc6299db9ec823/compiletest_rs-0.3.11/src/runtest.rs:2544:9
note: Run with `RUST_BACKTRACE=1` for a backtrace.

---- [compile-fail] compile-fail/memleak.rs stdout ----

error: compile-fail test compiled successfully!
status: exit code: 0
command: "target/debug/miri" "tests/compile-fail/memleak.rs" "-L" "/tmp/compiletestpMJMuw" "--target=x86_64-unknown-linux-gnu" "-C" "prefer-dynamic" "-o" "/tmp/compiletestpMJMuw/memleak.stage-id" "--sysroot" "/home/bjorn/.xargo/HOST" "-Zmir-emit-validate=1" "-L" "/tmp/compiletestpMJMuw/memleak.stage-id.aux" "-A" "unused"
stdout:
------------------------------------------

------------------------------------------
stderr:
------------------------------------------

------------------------------------------

thread '[compile-fail] compile-fail/memleak.rs' panicked at 'explicit panic', /home/bjorn/.cargo/registry/src/github.com-1ecc6299db9ec823/compiletest_rs-0.3.11/src/runtest.rs:2544:9

---- [compile-fail] compile-fail/memleak_rc.rs stdout ----

error: compile-fail test compiled successfully!
status: exit code: 0
command: "target/debug/miri" "tests/compile-fail/memleak_rc.rs" "-L" "/tmp/compiletestpMJMuw" "--target=x86_64-unknown-linux-gnu" "-C" "prefer-dynamic" "-o" "/tmp/compiletestpMJMuw/memleak_rc.stage-id" "--sysroot" "/home/bjorn/.xargo/HOST" "-Zmir-emit-validate=1" "-L" "/tmp/compiletestpMJMuw/memleak_rc.stage-id.aux" "-A" "unused"
stdout:
------------------------------------------

------------------------------------------
stderr:
------------------------------------------

------------------------------------------

thread '[compile-fail] compile-fail/memleak_rc.rs' panicked at 'explicit panic', /home/bjorn/.cargo/registry/src/github.com-1ecc6299db9ec823/compiletest_rs-0.3.11/src/runtest.rs:2544:9

---- [compile-fail] compile-fail/overflowing-rsh-2.rs stdout ----

error: compile-fail test compiled successfully!
status: exit code: 0
command: "target/debug/miri" "tests/compile-fail/overflowing-rsh-2.rs" "-L" "/tmp/compiletestpMJMuw" "--target=x86_64-unknown-linux-gnu" "--error-format" "json" "-C" "prefer-dynamic" "-o" "/tmp/compiletestpMJMuw/overflowing-rsh-2.stage-id" "--sysroot" "/home/bjorn/.xargo/HOST" "-Zmir-emit-validate=1" "-L" "/tmp/compiletestpMJMuw/overflowing-rsh-2.stage-id.aux" "-A" "unused"
stdout:
------------------------------------------

------------------------------------------
stderr:
------------------------------------------

------------------------------------------

thread '[compile-fail] compile-fail/overflowing-rsh-2.rs' panicked at 'explicit panic', /home/bjorn/.cargo/registry/src/github.com-1ecc6299db9ec823/compiletest_rs-0.3.11/src/runtest.rs:2544:9

---- [compile-fail] compile-fail/static_memory_modification.rs stdout ----

error: compile-fail test compiled successfully!
status: exit code: 0
command: "target/debug/miri" "tests/compile-fail/static_memory_modification.rs" "-L" "/tmp/compiletestpMJMuw" "--target=x86_64-unknown-linux-gnu" "--error-format" "json" "-C" "prefer-dynamic" "-o" "/tmp/compiletestpMJMuw/static_memory_modification.stage-id" "--sysroot" "/home/bjorn/.xargo/HOST" "-Zmir-emit-validate=1" "-L" "/tmp/compiletestpMJMuw/static_memory_modification.stage-id.aux" "-A" "unused"
stdout:
------------------------------------------

------------------------------------------
stderr:
------------------------------------------

------------------------------------------

thread '[compile-fail] compile-fail/static_memory_modification.rs' panicked at 'explicit panic', /home/bjorn/.cargo/registry/src/github.com-1ecc6299db9ec823/compiletest_rs-0.3.11/src/runtest.rs:2544:9


failures:
    [compile-fail] compile-fail/match_char.rs
    [compile-fail] compile-fail/memleak.rs
    [compile-fail] compile-fail/memleak_rc.rs
    [compile-fail] compile-fail/overflowing-rsh-2.rs
    [compile-fail] compile-fail/static_memory_modification.rs

test result: FAILED. 59 passed; 5 failed; 14 ignored; 0 measured; 0 filtered out

Edit: updated failure summary

bjorn3 · 2018-05-09T14:06:35Z

tests/compile-fail/never_transmute_humans.rs

@@ -1,17 +0,0 @@
-// This should fail even without validation


This causes an ICE

That's a bug, isn't it? Why remove the test?

it causes an ICE in rustc during compilation, not miri: http://play.integer32.com/?gist=76ba0e33563d046120e629846294f4a8&version=nightly&mode=debug

still... ignoring it for now is better so we can add it once upstream fixes it

Ah, I see.

Is this reported upstream? The ignore-test comment should link to the issue.

rust-lang/rust#50570

bjorn3 · 2018-05-09T14:06:56Z

tests/compile-fail/reallocate-bad-alignment-2.rs

@@ -1,16 +0,0 @@
-#![feature(alloc, allocator_api)]


This is simply not possible with the new api

How so? What's changed?

realloc doesn't have a new alignment param anymore.

unsafe fn realloc( &self, ptr: *mut Opaque, layout: Layout, new_size: usize ) -> *mut Opaque;

bjorn3 · 2018-05-09T14:07:00Z

tests/compile-fail/reallocate-bad-alignment.rs

@@ -1,16 +0,0 @@
-#![feature(alloc, allocator_api)]


This is simply not possible with the new api

bjorn3 · 2018-05-09T14:07:25Z

tests/compile-fail/repeat2.rs

@@ -1,3 +1,5 @@
+// ignore-test


Takes too long

Is it possible to make the array smaller and still get the failure?

If not I'd vote for just removing the test instead.

fun fact: we don't have any limits except stack frame depth anymore. This test won't ever trigger the error

oli-obk · 2018-05-09T14:12:31Z

how were these not failing before? are they disabled by default?

bjorn3 · 2018-05-09T15:23:30Z

They are disabled by me at https://github.com/solson/miri/blob/0eb3c18565070e4af530e0ec0f707d99be7c5cef/tests/compiletest.rs#L206-L207 because most of them were just failing, because of the new error format

oli-obk · 2018-05-09T15:26:33Z

Can we turn them back on now? Or what are the next steps for doing that?

bjorn3 · 2018-05-09T15:29:31Z

Can we turn them back on now? Or what are the next steps for doing that?

Once all errors have been converted to the new error format. Sadly compile-test only detects the constant evaluation error [E0080] message, but not what caused the error like for example "a memory access tried to interpret some bytes as a pointer".

oli-obk · 2018-05-09T15:31:08Z

you should be able to add //~^ NOTE a memory access tried to interpret... in the line after the error

bjorn3 · 2018-05-09T15:31:53Z

Will try

Edit: it works, will convert them all

bjorn3 · 2018-05-09T15:46:03Z

Don't have time to convert them all right now. Some errors have either strange placement or different message.

bjorn3 · 2018-05-09T17:08:38Z

tests/compile-fail/invalid_enum_discriminant.rs

        Foo::B => {},
        Foo::C => {},
        Foo::D => {},
    }
-}
+} //~ ERROR constant evaluation error [E0080]


Not sure why exactly here, but it's delayed because we don't have validation anymore

bjorn3

memleak.

bjorn3 · 2018-05-09T17:25:15Z

tests/compile-fail/match_char.rs

@@ -1,6 +1,7 @@
 fn main() {
    assert!(std::char::from_u32(-1_i32 as u32).is_none());
-    match unsafe { std::mem::transmute::<i32, char>(-1) } { //~ERROR tried to interpret an invalid 32-bit value as a char: 4294967295
+    match unsafe { std::mem::transmute::<i32, char>(-1) } { //~ ERROR constant evaluation error [E0080]


This test doesn't error

bjorn3 · 2018-05-09T17:25:59Z

tests/compile-fail/overflowing-rsh-2.rs

@@ -12,5 +12,6 @@

 fn main() {
    // Make sure we catch overflows that would be hidden by first casting the RHS to u32
-    let _n = 1i64 >> (u32::max_value() as i64 + 1); //~ Overflow(Shr)
+    let _n = 1i64 >> (u32::max_value() as i64 + 1); //~ ERROR constant evaluation error [E0080]


This test doesn't error

bjorn3 · 2018-05-09T17:26:35Z

tests/compile-fail/static_memory_modification.rs

@@ -3,7 +3,8 @@ static X: usize = 5;
 #[allow(mutable_transmutes)]
 fn main() {
    unsafe {
-        *std::mem::transmute::<&usize, &mut usize>(&X) = 6; //~ ERROR: tried to modify constant memory
+        *std::mem::transmute::<&usize, &mut usize>(&X) = 6; //~ ERROR constant evaluation error [E0080]


Statics aren't marked static yet.

bjorn3 · 2018-05-09T17:27:33Z

src/lib.rs

@@ -171,6 +171,7 @@ pub fn eval_main<'a, 'tcx: 'a>(
        Ok(()) => {
            let leaks = ecx.memory().leak_report();
            if leaks != 0 {
+                // TODO: Prevent leaks which aren't supposed to be there
                //tcx.sess.err("the evaluated program leaked memory");


Needs to be uncommented to allow memleak.rs and memleak_rc.rs to pass.

bjorn3 · 2018-05-09T17:28:09Z

tests/compiletest.rs

@@ -203,7 +203,7 @@ fn run_pass_rustc() {
 }

 #[test]
-#[ignore] // TODO: update test errors
+#[should_panic] // TODO: update test errors


Remove when all tests pass

bjorn3 · 2018-06-01T11:05:55Z

Rebased on top of #376

oli-obk · 2018-06-03T11:09:23Z

can you rebase again?

bjorn3 · 2018-06-05T16:13:01Z

Rebased

oli-obk · 2018-06-05T16:38:20Z

Thanks! I'm tracking the leftover failing tests in #382

bjorn3 commented May 9, 2018

View reviewed changes

bjorn3 force-pushed the fix_some_tests branch from 906b75e to 1ad9b90 Compare May 9, 2018 17:09

bjorn3 commented May 9, 2018

View reviewed changes

bjorn3 force-pushed the fix_some_tests branch 2 times, most recently from 698d619 to 623a445 Compare May 11, 2018 12:26

bjorn3 force-pushed the fix_some_tests branch from 623a445 to f64e5bc Compare May 20, 2018 13:08

bjorn3 force-pushed the fix_some_tests branch from f64e5bc to 4790ed7 Compare June 1, 2018 11:05

bjorn3 added 5 commits June 5, 2018 18:08

Fix a couple of tests

3db0568

Address some review comments

569792a

Convert some of the tests to the new format

8ae66db

Convert legitimate failing errors to the new error format

94754de

Re-enable never_transmute_humans.rs

ac25a51

bjorn3 force-pushed the fix_some_tests branch from e529637 to ac25a51 Compare June 5, 2018 16:09

oli-obk merged commit 0755fd6 into rust-lang:master Jun 5, 2018

bjorn3 deleted the fix_some_tests branch June 5, 2018 17:45

Fix a couple of compile-fail tests #376

Fix a couple of compile-fail tests #376

Uh oh!

Conversation

bjorn3 commented May 9, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bjorn3 May 9, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

oli-obk commented May 9, 2018

Uh oh!

bjorn3 commented May 9, 2018

Uh oh!

oli-obk commented May 9, 2018

Uh oh!

bjorn3 commented May 9, 2018

Uh oh!

oli-obk commented May 9, 2018

Uh oh!

bjorn3 commented May 9, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

bjorn3 commented May 9, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bjorn3 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bjorn3 commented Jun 1, 2018

Uh oh!

oli-obk commented Jun 3, 2018

Uh oh!

bjorn3 commented Jun 5, 2018

Uh oh!

oli-obk commented Jun 5, 2018

Uh oh!

Uh oh!

bjorn3 commented May 9, 2018 •

edited

Loading

bjorn3 May 9, 2018 •

edited

Loading

bjorn3 commented May 9, 2018 •

edited

Loading