Remove comments from mir-opt MIR dumps #112346

saethlin · 2023-06-06T13:47:32Z

See https://rust-lang.zulipchat.com/#narrow/stream/189540-t-compiler.2Fwg-mir-opt/topic/Line.20numbers.20in.20mir-opt.20tests/near/363849874

In #99780 there is mention that "there has been a zulip conversation about disabling line numbers with mixed opinions" which to me means that some people opposed this. I can't find the referenced conversation so... here we go.

The current situation is quite chaotic. It's not hard to find MIR diffs which contain

Absolute line numbers
Relative line numbers

Substituted line numbers (LL)
For example:

rust/tests/mir-opt/inline/inline_shims.drop.Inline.diff

Lines 10 to 17 in 408bbd0

    
                 let mut _5: *mut std::option::Option<B>; // in scope 0 at $DIR/inline_shims.rs:+2:38: +2:39 
        
                 scope 1 { 
        
                 } 
        
                 scope 2 { 
        
           +         scope 3 (inlined std::ptr::drop_in_place::<Option<B>> - shim(Some(Option<B>))) { // at $DIR/inline_shims.rs:12:14: 12:40 
        
           +             let mut _6: isize;           // in scope 3 at $SRC_DIR/core/src/ptr/mod.rs:LL:COL 
        
           +             let mut _7: isize;           // in scope 3 at $SRC_DIR/core/src/ptr/mod.rs:LL:COL 
        
           +         }

And sometimes adding a comment at the top of a mir-opt test generates a diff in the test because a line number changed: https://github.com/rust-lang/rust/pull/98112/files#diff-b8cf4bcce95078e6a3faf075e9abf6864872fb28a64d95c04f04513b9e3bbd81

And irrelevant changes to the standard library can generate diffs in mir-opt tests: https://github.com/rust-lang/rust/pull/110694/files#diff-bf96b0e7c67b8b272814536888fd9428c314991e155beae1f0a2a67f0ac47b2c
769886c

I think we should, specifically in mir-opt tests, completely remove the comments, or insert placeholders for all line and column numbers.

rustbot · 2023-06-06T13:47:40Z

r? @eholk

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

bjorn3 · 2023-06-06T13:49:11Z

tests/mir-opt/building/custom/consts.statics.built.after.mir

+        _2 = const {alloc2: *mut i32};
                                         // mir::Constant
                                         // + span: $DIR/consts.rs:28:38: 28:39
                                         // + literal: Const { ty: *mut i32, val: Value(Scalar(alloc2)) }


Can you put the // mir::Constant on the same line as the _2 = const {alloc2: *mut i32};?

Ah I see I missed a few more places the comments are emitted.

saethlin · 2023-06-06T21:28:13Z

r? mir-opt

bors · 2023-06-14T18:20:16Z

☔ The latest upstream changes (presumably #112418) made this pull request unmergeable. Please resolve the merge conflicts.

oli-obk · 2023-06-15T06:39:17Z

@bors r+ p=1 bitrotty

bors · 2023-06-15T06:39:19Z

📌 Commit e2fab8f0dcca015d83c83a06f5f25eb368b7bf5e has been approved by oli-obk

It is now in the queue for this repository.

bors · 2023-06-15T09:13:56Z

⌛ Testing commit e2fab8f0dcca015d83c83a06f5f25eb368b7bf5e with merge e72dff92be7b99d9222f4fdb685df5ca261c257a...

saethlin · 2023-06-15T19:21:46Z

I goofed. I forgot that even with all the awesome improvements to blessing that @pietroalbini built, blessing with debug = true is still not effective.

oli-obk · 2023-06-15T19:47:01Z

@bors r+

bors · 2023-06-15T19:47:03Z

📌 Commit 0a1fa41 has been approved by oli-obk

It is now in the queue for this repository.

scottmcm · 2023-06-15T23:30:57Z

blessing with debug = true is still not effective.

Not this PR, but maybe emitting a warning for that could make sense?

(And I assume it's debug-std = true that's the major issue? debug = true; debug-std = false should be fine for blessing these?)

saethlin · 2023-06-15T23:36:59Z

The problem is that the tests don't get blessed at all because they are // ignore-debug.

Yes, it's the standard library code that's the problem, and I think this can be gracefully solved by piggybacking on the work Pietro did.

bors · 2023-06-16T01:55:37Z

⌛ Testing commit 0a1fa41 with merge c84d5e7...

bors · 2023-06-16T04:44:53Z

☀️ Test successful - checks-actions
Approved by: oli-obk
Pushing c84d5e7 to master...

rust-timer · 2023-06-16T06:00:47Z

Finished benchmarking commit (c84d5e7): comparison URL.

Overall result: no relevant changes - no action needed

@rustbot label: -perf-regression

Instruction count

This benchmark run did not return any relevant results for this metric.

Max RSS (memory usage)

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	2.2%	[2.2%, 2.2%]	1
Improvements ✅ (primary)	-2.2%	[-2.2%, -2.2%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-2.2%	[-2.2%, -2.2%]	1

Cycles

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	3.4%	[1.1%, 4.5%]	6
Regressions ❌ (secondary)	1.4%	[1.4%, 1.4%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	3.4%	[1.1%, 4.5%]	6

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 649.004s -> 647.175s (-0.28%)

lqd · 2023-07-18T18:33:28Z

It seems this PR removed comments from -Zdump-mir dumps that were not mir-opt tests. I'll open a PR.

shepmaster · 2023-09-06T02:04:23Z

This also removed functionality from the playground. I'll delete that code now.

Expand NLL MIR dumps This PR is a first step to clean up and expand NLL MIR dumps: - by restoring the "mir-include-spans" comments which are useful for `-Zdump-mir=nll` - by adding the list of borrows to NLL MIR dumps, where they are introduced in the CFG and in which region Comments in MIR dumps were turned off in rust-lang#112346, but as shown in rust-lang#114652 they were still useful for us working with NLL MIR dumps. So this PR pulls `-Z mir-include-spans` into its own options struct, so that passes dumping MIR can override them if need be. The rest of the compiler is not affected, only the "nll" pass dumps have these comments enabled again. The CLI still has priority when specifying the flag, so that we can explicitly turn them off in the `mir-opt` tests to keep blessed dumps easier to work with (which was one of the points of rust-lang#112346). Then, as part of a couple steps to improve NLL/polonius MIR dumps and `.dot` visualizations, I've also added the list of borrows and where they're introduced. I'm doing all this to help debug some polonius scope issues in my prototype location-sensitive analysis :3. I'll probably add member constraints soon.

Rollup merge of rust-lang#129711 - lqd:nll-mir-dumps, r=compiler-errors Expand NLL MIR dumps This PR is a first step to clean up and expand NLL MIR dumps: - by restoring the "mir-include-spans" comments which are useful for `-Zdump-mir=nll` - by adding the list of borrows to NLL MIR dumps, where they are introduced in the CFG and in which region Comments in MIR dumps were turned off in rust-lang#112346, but as shown in rust-lang#114652 they were still useful for us working with NLL MIR dumps. So this PR pulls `-Z mir-include-spans` into its own options struct, so that passes dumping MIR can override them if need be. The rest of the compiler is not affected, only the "nll" pass dumps have these comments enabled again. The CLI still has priority when specifying the flag, so that we can explicitly turn them off in the `mir-opt` tests to keep blessed dumps easier to work with (which was one of the points of rust-lang#112346). Then, as part of a couple steps to improve NLL/polonius MIR dumps and `.dot` visualizations, I've also added the list of borrows and where they're introduced. I'm doing all this to help debug some polonius scope issues in my prototype location-sensitive analysis :3. I'll probably add member constraints soon.

rustbot assigned eholk Jun 6, 2023

bjorn3 reviewed Jun 6, 2023

View reviewed changes