Extend Polonius fact generation for (some) move tracking by amandasystems · Pull Request #62800 · rust-lang/rust

amandasystems · 2019-07-19T13:46:05Z

This PR will extend rustc to emit facts used for tracking moves and initialization in Polonius. It is most likely the final part of my master's thesis work.

amandasystems · 2019-07-19T13:47:40Z

r? @nikomatsakis

(for now!)

joelpalmer · 2019-07-29T10:41:09Z

Ping from triage. This PR is in need of review. Should it be re-assigned? @nikomatsakis @cramertj
cc: @AlbinS

amandasystems · 2019-07-29T14:05:02Z

Ping from triage. This PR is in need of review. Should it be re-assigned? @nikomatsakis @cramertj
cc: @AlbinS

@joelpalmer Thanks for asking, but it's still very WIP; the fact generation is currently slightly wrong. I expect to fix it this week, but there are probably more issues than that hiding behind that issue.

However, before this PR can be merged, Polonius needs to be updated from another branch (I will rebase and fixup the commit using my local Polonius to in stead pull the new release).

So in short; I suspect @nikomatsakis might be back from their vacation before I'm done. If not, I'll ask around in the Polonius working group. Please don't reassign for now, thank you!

bors · 2019-08-02T11:26:54Z

☔ The latest upstream changes (presumably #61393) made this pull request unmergeable. Please resolve the merge conflicts.

amandasystems · 2019-08-06T09:56:26Z

This pull request is waiting for rust-lang/polonius#110 to be merged into Polonius.

hdhoang · 2019-08-16T01:43:02Z

Thank you for your update, triage-wg have updated this PR's label accordingly.

bors · 2019-08-19T20:46:29Z

☔ The latest upstream changes (presumably #63579) made this pull request unmergeable. Please resolve the merge conflicts.

- var_starts_path - parent - initialized_at - moved_out_at This also switches to the intended emission of `var_drop_used` fact emission, where that fact is always emitted on a drop-use of a variable, regardless of its initialization status, as Polonius now handles that.

amandasystems · 2019-09-04T07:52:34Z

r? @nikomatsakis

Ready for review, as rust-lang/polonius#110 has landed! 🎉

This should be enough facts to implement move analysis and liveness tracking in Polonius, meaning that an entire new part of the Polonius analysis is finally fully unlocked!

nikomatsakis · 2019-09-04T20:30:59Z

+ all_facts.var_starts_path.extend(move_data.rev_lookup.iter_locals_enumerated().map(|(v, &m)| (v, m)));
+
+ for (idx, move_path) in move_data.move_paths.iter_enumerated() {
+ all_facts.parent.extend(move_path.parents(&move_data.move_paths).iter().map(|&parent| (parent, idx)));


as noted in the polonius PR, ultimately I think we should leave the transitive computation here to polonius. But I guess I don't wnat to do that in this PR.

nikomatsakis · 2019-09-04T20:32:23Z

+ // The initialization happened in (or rather, when arriving at)
+ // the successors, but not in the unwind block.
+ let first_statement = Location { block: successor, statement_index: 0};
+ all_facts.initialized_at.push((init.path, location_table.start_index(first_statement)));


Interesting quirk. I guess this makes sense.

Let's just say I scratched my head for quite some time before I figured out why this happened. The results were completely wrong otherwise. If you are curious about the details, there's a ton of logs on Zulip where I figure this out in the initialization topic.

nikomatsakis · 2019-09-04T20:34:18Z

+ self.path_accessed_at.push((path, self.location_to_index(location)));
+ }
+
+ fn place_to_mpi(&self, place: &Place<'_>) -> Option<MovePathIndex> {


might be nice to have some docs here -- when does this return None?

I guess it is when the place is rooted in a static, though

nikomatsakis · 2019-09-04T20:36:03Z

I'm not going to block on those nits. Doesn't seem worth it.

nikomatsakis · 2019-09-04T20:36:07Z

@bors r+

bors · 2019-09-04T20:36:08Z

📌 Commit 28312b5 has been approved by nikomatsakis

bors · 2019-09-05T00:25:29Z

⌛ Testing commit 28312b5 with merge b6eac35c6f9719fb329d9f6c259eda2c47f6fff8...

bors · 2019-09-05T01:59:07Z

💔 Test failed - checks-azure

rust-highfive · 2019-09-05T01:59:08Z

The job x86_64-gnu-llvm-6.0 of your PR failed (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.

2019-09-05T01:56:52.8631531Z test vec.rs - vec::Vec<T>::with_capacity (line 339) ... ok
2019-09-05T01:56:52.8634946Z 2019-09-05T01:56:52.8638137Z failures:
2019-09-05T01:56:52.8639846Z 2019-09-05T01:56:52.8649490Z ---- prelude/mod.rs - prelude (line 6) stdout ----
2019-09-05T01:56:52.8651457Z error[E0523]: found two different crates with name `alloc` that are not distinguished by differing `-C metadata`. This will result in symbol conflicts between the two.
2019-09-05T01:56:52.8651840Z --> prelude/mod.rs:8:1
2019-09-05T01:56:52.8652019Z 5 | extern crate alloc;
2019-09-05T01:56:52.8652507Z | ^^^^^^^^^^^^^^^^^^^
2019-09-05T01:56:52.8652589Z 2019-09-05T01:56:52.8652656Z error: aborting due to previous error
2019-09-05T01:56:52.8652656Z error: aborting due to previous error
2019-09-05T01:56:52.8652720Z 2019-09-05T01:56:52.8653060Z Couldn't compile the test.
2019-09-05T01:56:52.8653403Z ---- raw_vec.rs - raw_vec::RawVec<T, A>::double (line 261) stdout ----
2019-09-05T01:56:52.8654097Z error[E0523]: found two different crates with name `alloc` that are not distinguished by differing `-C metadata`. This will result in symbol conflicts between the two.
2019-09-05T01:56:52.8654844Z --> raw_vec.rs:263:1
2019-09-05T01:56:52.8655005Z 4 | extern crate alloc;
2019-09-05T01:56:52.8655214Z | ^^^^^^^^^^^^^^^^^^^
2019-09-05T01:56:52.8655270Z 2019-09-05T01:56:52.8655355Z error: aborting due to previous error
2019-09-05T01:56:52.8655355Z error: aborting due to previous error
2019-09-05T01:56:52.8655402Z 2019-09-05T01:56:52.8655705Z Couldn't compile the test.
2019-09-05T01:56:52.8656017Z ---- raw_vec.rs - raw_vec::RawVec<T, A>::reserve (line 466) stdout ----
2019-09-05T01:56:52.8656482Z error[E0523]: found two different crates with name `alloc` that are not distinguished by differing `-C metadata`. This will result in symbol conflicts between the two.
2019-09-05T01:56:52.8656775Z --> raw_vec.rs:468:1
2019-09-05T01:56:52.8656934Z 4 | extern crate alloc;
2019-09-05T01:56:52.8657021Z | ^^^^^^^^^^^^^^^^^^^
2019-09-05T01:56:52.8657078Z 2019-09-05T01:56:52.8657161Z error: aborting due to previous error
---
2019-09-05T01:56:52.8811600Z == clock drift check ==
2019-09-05T01:56:52.8828346Z local time: Thu Sep 5 01:56:52 UTC 2019
2019-09-05T01:56:53.0500684Z network time: Thu, 05 Sep 2019 01:56:53 GMT
2019-09-05T01:56:53.0503670Z == end clock drift check ==
2019-09-05T01:56:53.5693162Z ##[error]Bash exited with code '1'.
2019-09-05T01:56:53.5770296Z ##[section]Starting: Upload CPU usage statistics
2019-09-05T01:56:53.5775860Z ==============================================================================
2019-09-05T01:56:53.5775991Z Task : Bash
2019-09-05T01:56:53.5776060Z Description : Run a Bash script on macOS, Linux, or Windows

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)

Centril · 2019-09-05T07:25:15Z

@bors retry probably spurious

bors · 2019-09-05T08:51:48Z

⌛ Testing commit 28312b5 with merge 9776723...

Extend Polonius fact generation for (some) move tracking This PR will extend rustc to emit facts used for tracking moves and initialization in Polonius. It is most likely the final part of my master's thesis work.

bors · 2019-09-05T12:41:36Z

☀️ Test successful - checks-azure
Approved by: nikomatsakis
Pushing 9776723 to master...

rust-highfive assigned cramertj Jul 19, 2019

This comment has been minimized.
Sign in to view

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Jul 19, 2019

rust-highfive assigned nikomatsakis and unassigned cramertj Jul 19, 2019

This comment has been minimized.
Sign in to view

jonas-schievink 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 Jul 29, 2019

This comment has been minimized.
Sign in to view

hdhoang added S-blocked Status: Blocked on something else such as an RFC or other implementation work. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Aug 16, 2019

This comment has been minimized.
Sign in to view

Albin Stjerna added 7 commits September 4, 2019 09:44

NLL: New Polonius version
2adb246

NLL: Adapt to the new Polonius types
996ba93

Polonius: Rename parent and var_starts_path
9f39e00

Polonius: don't emit var_maybe_initialized_on_exit
6568b08

Polonius: don't emit region_live_at
9cd1a11

Polonius: emit variable access facts
560ef6d

amandasystems changed the title ~~[WIP] Extend Polonius fact generation for (some) move tracking~~Extend Polonius fact generation for (some) move trackingSep 4, 2019

This comment has been minimized.
Sign in to view

Rustfmt the files I touched
28312b5

Dylan-DPC-zz added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-blocked Status: Blocked on something else such as an RFC or other implementation work. labels Sep 4, 2019

nikomatsakis requested changes Sep 4, 2019
View reviewed changes

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 Sep 4, 2019

bors added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Sep 5, 2019

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 Sep 5, 2019

bors added the merged-by-bors This PR was explicitly merged by bors. label Sep 5, 2019

bors merged commit 28312b5 into rust-lang:masterSep 5, 2019

Uh oh!

Uh oh!

Conversation

amandasystems commented Jul 19, 2019

Uh oh!

This comment has been minimized.

amandasystems commented Jul 19, 2019

Uh oh!

This comment has been minimized.

joelpalmer commented Jul 29, 2019

Uh oh!

amandasystems commented Jul 29, 2019

Uh oh!

bors commented Aug 2, 2019

Uh oh!

This comment has been minimized.

amandasystems commented Aug 6, 2019

Uh oh!

This comment has been minimized.

hdhoang commented Aug 16, 2019

Uh oh!

This comment has been minimized.

bors commented Aug 19, 2019

Uh oh!

amandasystems commented Sep 4, 2019

Uh oh!

This comment has been minimized.

nikomatsakisSep 4, 2019

Choose a reason for hiding this comment

Uh oh!

nikomatsakisSep 4, 2019

Choose a reason for hiding this comment

Uh oh!

amandasystemsSep 4, 2019• 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!

nikomatsakisSep 4, 2019

Choose a reason for hiding this comment

Uh oh!

nikomatsakisSep 4, 2019

Choose a reason for hiding this comment

Uh oh!

nikomatsakis commented Sep 4, 2019

Uh oh!

nikomatsakis commented Sep 4, 2019

Uh oh!

bors commented Sep 4, 2019

Uh oh!

bors commented Sep 5, 2019

Uh oh!

bors commented Sep 5, 2019

Uh oh!

rust-highfive commented Sep 5, 2019

Uh oh!

Centril commented Sep 5, 2019

Uh oh!

bors commented Sep 5, 2019

Uh oh!

bors commented Sep 5, 2019

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

10 participants

amandasystemsSep 4, 2019•
edited
Loading