Future-proof MIR for dedicated debuginfo. #56278

eddyb · 2018-11-27T12:09:09Z

This is #56231 without the last commit (the one that actually moves to VarDebuginfo).
Nothing should be broken, but it should no longer depend on debuginfo for anything else.

r? @nikomatsakis

nikomatsakis

these changes look fine, though I'd be happier if we could make some nice queries for getting info about the Nth upvar of a closure, so that the code would be a bit cleaner

nikomatsakis · 2018-11-29T18:25:08Z

src/librustc_mir/borrow_check/mod.rs

@@ -141,6 +142,40 @@ fn do_mir_borrowck<'a, 'gcx, 'tcx>(
        .as_local_node_id(def_id)
        .expect("do_mir_borrowck: non-local DefId");

+    // Gather the upvars of a closure, if any.
+    let upvars: Vec<_> = tcx.with_freevars(id, |freevars| {


This really feels like information that could be encoded in the MIR, no? Well, I guess I can see an argument both ways.

At minimum I would prefer if we can factor it out from borrow check into some kind of helper table that would be more re-usable.

But I don't know that I want to block the PR on it. In general I feel like the borrow check code is kind of noodly and could use some extractions of this kind.

nikomatsakis · 2018-11-29T18:28:28Z

src/librustc/mir/mod.rs

    /// Mark an argument local (which must be a tuple) as getting passed as
    /// its individual components at the LLVM level.
    ///
    /// This is used for the "rust-call" ABI.
    pub spread_arg: Option<Local>,

+    /// Names and capture modes of all the closure upvars, assuming
+    /// the first argument is either the closure or a reference to it.
+    pub upvar_decls: Vec<UpvarDecl>,


btw, I think that if there are fields that are "debuginfo" that you feel other people should not rely on, we should comment that =)

(Like this one, judging from previous commit)

nikomatsakis · 2018-11-29T18:32:19Z

src/librustc_mir/interpret/validity.rs

-                    PathElem::ClosureVar(Symbol::intern(&field.to_string()))
+                let mut name = None;
+                if let Some(node_id) = self.ecx.tcx.hir.as_local_node_id(def_id) {
+                     self.ecx.tcx.with_freevars(node_id, |freevars| {


I wonder if this has any overlap with the borrow check code I was noting earlier? (i.e., if some common helper could make both of these more concise)

nikomatsakis · 2018-11-29T18:33:04Z

src/librustc/mir/mod.rs

+    // debuginfo generation, and will be removed at some point.
+    // Do **NOT** use it for anything else, upvar information should not be
+    // in the MIR, please rely on local crate HIR or other side-channels.
+    pub __upvar_debuginfo_codegen_only_do_not_use: Vec<UpvarDebuginfo>,


lol ok I guess that addresses my earlier concern

I would like this to be part of the doc comment please though, why use only //?

I got used to only using // for NOTE/FIXME/HACK, but I'll go change it.

nikomatsakis · 2018-11-29T18:38:16Z

src/librustc_mir/borrow_check/mod.rs

@@ -2208,3 +2277,15 @@ impl ContextKind {
        }
    }
 }
+
+#[derive(Debug)]
+crate struct Upvar {


Yeah I think I would prefer to see closure_upvars(def_id, n) -> Upvar that returns this information "nicely collected".... (or closure_upvars(def_id) -> &[Upvar]).

But I won't hold up the PR on it.

r=me, depending on how you feel @eddyb

bors · 2018-12-02T20:52:29Z

☔ The latest upstream changes (presumably #56198) made this pull request unmergeable. Please resolve the merge conflicts.

XAMPPRocky · 2018-12-10T08:03:44Z

Triage; @eddyb Hello, have you been able to get back to this PR?

XAMPPRocky · 2019-01-11T13:20:33Z

Triage; @eddyb Hello again, have you been able to get back to this PR?

TimNN · 2019-01-22T18:40:34Z

Ping from triage @eddyb: What is the status of this PR?

Dylan-DPC-zz · 2019-02-04T13:03:25Z

ping from triage @eddyb Unfortunately we haven't heard from you on this in a while, so I'm closing the PR to keep things tidy. Don't worry though, if you'll have time again in the future please reopen this PR, we'll be happy to review it again!

eddyb · 2019-02-14T11:44:48Z

I've started rebasing this, I'll try to get it merged next week.

eddyb · 2019-02-15T12:16:16Z

I think I want to get rid of rustc_mir::borrowck::Upvar, I might be able to reuse some of the new information that's in the TypeckTables now.

bors · 2019-02-22T09:34:47Z

☔ The latest upstream changes (presumably #56113) made this pull request unmergeable. Please resolve the merge conflicts.

Dylan-DPC-zz · 2019-04-01T10:40:47Z

ping from triage @eddyb any updates? you have conflicts to resolve

eddyb · 2019-04-17T10:23:08Z

Regarding #56278 (comment), the TypeckTables field in question is from issue #56905 / PR #56906, i.e.:

rust/src/librustc_mir/build/mod.rs

Lines 628 to 640 in e4e032a

    
           // In analyze_closure() in upvar.rs we gathered a list of upvars used by a 
        
           // closure and we stored in a map called upvar_list in TypeckTables indexed 
        
           // with the closure's DefId. Here, we run through that vec of UpvarIds for 
        
           // the given closure and use the necessary information to create UpvarDecl. 
        
           let upvar_decls: Vec<_> = hir_tables 
        
               .upvar_list 
        
               .get(&fn_def_id) 
        
               .into_iter() 
        
               .flatten() 
        
               .map(|upvar_id| { 
        
                   let var_hir_id = upvar_id.var_path.hir_id; 
        
                   let var_node_id = tcx_hir.hir_to_node_id(var_hir_id); 
        
                   let capture = hir_tables.upvar_capture(*upvar_id);

…info_upvar_ops_sequence.

…R borrowck.

eddyb · 2019-04-23T20:37:04Z

@bors r=nikomatsakis

bors · 2019-04-23T20:37:06Z

📌 Commit c3ca9a3 has been approved by nikomatsakis

matthewjasper · 2019-04-23T20:15:46Z

src/librustc_mir/borrow_check/nll/explain_borrow/mod.rs

-                                mir.local_decls[*borrowed_local].visibility_scope;
-
-                            if mir.is_sub_scope(borrowed_local_scope, dropped_local_scope)
+                             if mir.local_decls[*borrowed_local].name.is_some()


Why isn't this checking is_user_variable?

Oh, looks like is_user_variable is newer than this PR originally, and is actually part of the set of things I meant to review at some point.

@nikomatsakis

…tsakis Future-proof MIR for dedicated debuginfo. This is rust-lang#56231 without the last commit (the one that actually moves to `VarDebuginfo`). Nothing should be broken, but it should no longer depend on debuginfo for anything else. r? @nikomatsakis

bors · 2019-04-24T03:13:56Z

⌛ Testing commit c3ca9a3 with merge d101749a5366c81e052027e957d90b7430566438...

@nikomatsakis

…tsakis Future-proof MIR for dedicated debuginfo. This is rust-lang#56231 without the last commit (the one that actually moves to `VarDebuginfo`). Nothing should be broken, but it should no longer depend on debuginfo for anything else. r? @nikomatsakis

Centril · 2019-04-24T03:16:43Z

@bors retry

@ghost

Rollup of 5 pull requests Successful merges: - #56278 (Future-proof MIR for dedicated debuginfo.) - #59739 (Stabilize futures_api) - #59822 (Fix dark css rule) - #60186 (Temporarily accept [i|u][32|size] suffixes on a tuple index and warn) - #60190 (Don't generate unnecessary rmeta files.) Failed merges: r? @ghost

rust-highfive assigned nikomatsakis Nov 27, 2018

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Nov 27, 2018

eddyb force-pushed the mir-debuginfo-proof branch from b4c1d38 to 58ff04b Compare November 29, 2018 08:25

nikomatsakis approved these changes Nov 29, 2018

View reviewed changes

nikomatsakis reviewed Nov 29, 2018

View reviewed changes

nikomatsakis added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Nov 30, 2018

eddyb mentioned this pull request Dec 10, 2018

rustc: move debug info from LocalDecl and UpvarDecl into a dedicated VarDebugInfo. #56231

Merged

TimNN added A-allocators Area: Custom and system allocators and removed A-allocators Area: Custom and system allocators labels Jan 29, 2019

Dylan-DPC-zz closed this Feb 4, 2019

Dylan-DPC-zz added S-inactive Status: Inactive and waiting on the author. This is often applied to closed PRs. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Feb 4, 2019

eddyb mentioned this pull request Feb 14, 2019

Only insert nodes which changes lint levels in the LintLevelMap #58176

Merged

eddyb reopened this Feb 14, 2019

eddyb force-pushed the mir-debuginfo-proof branch from 58ff04b to 2b546c4 Compare February 15, 2019 12:15

bors added the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Feb 22, 2019

Dylan-DPC-zz removed the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Mar 11, 2019

TimNN added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-inactive Status: Inactive and waiting on the author. This is often applied to closed PRs. labels Mar 19, 2019

eddyb force-pushed the mir-debuginfo-proof branch from 2b546c4 to b6055ef Compare April 23, 2019 08:11

eddyb mentioned this pull request Apr 23, 2019

Add match arm scopes and other scope fixes #60174

Merged

eddyb added 9 commits April 23, 2019 17:27

rustc_mir: don't rely on mir::UpvarDecl in the MIR borrowck.

61fcbfc

rustc: don't track var_hir_id or mutability in mir::UpvarDecl.

82bd719

rustc_mir: don't use upvar_decls in the generator state transform.

c1ec45f

rustc_mir: don't use upvar_decls in the miri validity checks.

2384e98

rustc_codegen_ssa: rename debuginfo_upvar_decls_ops_sequence to debug…

e6ec968

…info_upvar_ops_sequence.

rustc: dissuade compiler developers from misusing upvar debuginfo.

1525dc2

rustc_mir: don't rely on mir::LocalDecl visibility_scopes in the MI…

d20da13

…R borrowck.

rustc_mir: pretty-print all locals into their respective scopes.

9260305

rustc_mir: create the let and "remainder" scopes in source order.

c3ca9a3

eddyb force-pushed the mir-debuginfo-proof branch from b6055ef to c3ca9a3 Compare April 23, 2019 20:35

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-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Apr 23, 2019

matthewjasper reviewed Apr 23, 2019

View reviewed changes

Centril mentioned this pull request Apr 24, 2019

Rollup of 6 pull requests #60221

Closed

Centril mentioned this pull request Apr 24, 2019

Rollup of 5 pull requests #60224

Merged

bors merged commit c3ca9a3 into rust-lang:master Apr 24, 2019

eddyb deleted the mir-debuginfo-proof branch April 24, 2019 08:23

Mark-Simulacrum mentioned this pull request Apr 26, 2019

ICE while bootstrapping (./x.py build) #2 #60228

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Future-proof MIR for dedicated debuginfo. #56278

Future-proof MIR for dedicated debuginfo. #56278

eddyb commented Nov 27, 2018

nikomatsakis left a comment

nikomatsakis Nov 29, 2018

nikomatsakis Nov 29, 2018

nikomatsakis Nov 29, 2018

nikomatsakis Nov 29, 2018

nikomatsakis Nov 29, 2018

nikomatsakis Nov 29, 2018

nikomatsakis Nov 29, 2018

eddyb Nov 30, 2018

nikomatsakis Nov 29, 2018

bors commented Dec 2, 2018

XAMPPRocky commented Dec 10, 2018

XAMPPRocky commented Jan 11, 2019

TimNN commented Jan 22, 2019

Dylan-DPC-zz commented Feb 4, 2019

eddyb commented Feb 14, 2019

eddyb commented Feb 15, 2019

bors commented Feb 22, 2019

Dylan-DPC-zz commented Apr 1, 2019

eddyb commented Apr 17, 2019

eddyb commented Apr 23, 2019

bors commented Apr 23, 2019

matthewjasper Apr 23, 2019

eddyb Apr 24, 2019

bors commented Apr 24, 2019

Centril commented Apr 24, 2019

Future-proof MIR for dedicated debuginfo. #56278

Future-proof MIR for dedicated debuginfo. #56278

Conversation

eddyb commented Nov 27, 2018

nikomatsakis left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

bors commented Dec 2, 2018

XAMPPRocky commented Dec 10, 2018

XAMPPRocky commented Jan 11, 2019

TimNN commented Jan 22, 2019

Dylan-DPC-zz commented Feb 4, 2019

eddyb commented Feb 14, 2019

eddyb commented Feb 15, 2019

bors commented Feb 22, 2019

Dylan-DPC-zz commented Apr 1, 2019

eddyb commented Apr 17, 2019

eddyb commented Apr 23, 2019

bors commented Apr 23, 2019

Choose a reason for hiding this comment

Choose a reason for hiding this comment

bors commented Apr 24, 2019

Centril commented Apr 24, 2019