Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Compute type fragments left over from moves #17439

Closed

Conversation

pnkfelix
Copy link
Member

Compute type fragments left over from moves.

This is part 1 of a pair of PRs that prepare for planned revisions to our drop implementation; this foundation is required for either "static drop semantics" (RFC PR 210) or for the improved dynamic drop semantics that uses an out-of-band bitfield on the stack to track which loan-paths still need to be dropped.

This PR deliberately leaves out the control-flow analysis changes that are actually necessary for the drop changes, which are in part 2. This is more fundamental stuff here.

(The commits are deliberately split up in an effort to ease reading these revisions; I do not plan to squash them.)

…varId`.

(This refactoring looks like a wash in this commit, in terms of
overall code-cleanliness; but it makes things nicer in a future commit
for static drop semantics.)
…ef and by-value captures.

This is necessary to be able to properly backtrack in a loan-path to
the definition site of the original variable (which I found necessary
in later commits).
`LpDowncast` carries the `DefId` of the variant itself.  To support
this, added the enum variant `DefId` to the `cat_downcast` variant in
`mem_categorization::categorization`.
To make this clean, refactored old `LoanPath` enum into a
`LoanPath` struct with a `ty::t` and a newly-added `LoanPathVariant` enum.

This enabled me to get rid of the ugly and fragile `LoanPath::to_type`
method, and I can probably also get rid of other stuff that was
supporting it, maybe.
This is accomplished by:

1. Add `MatchMode` enum to `expr_use_visitor`.

2. Computing the match mode for each pattern via a pre-pass, and then
   passing the mode along when visiting the pattern in
   expr_use_visitor.

3. Adding a `fn matched_pat` callback to expr_use_visitor, which is called on
   all nodes of the pattern (as opposed to `fn consume_pat`, which is only
   invoked for identifiers at the leaves of the pattern), and invoking it
   accordingly.

Of particular interest are the `cat_downcast` instances established
when matching enum variants.
Note that due to PR rust-lang#16053, we only actually deal with `a @ Var(_)`
here, (as opposed to a hypothetical `a @ Var(ref b)`, which cannot
currenty arise, and maybe will never be supported for non-copy data).
… a span.

This will be utilized by the rustc_drop_obligations
code-annotation-via-errors hack to print out the drop obligations that
are removed at the end of scopes.
When you add `rustc_drop_obligations` as an annotation on a function,
it will print all of the drop-obligation manipulations as individual
compiler errors associated with the expressions that caused the effect
on the drop obligation state.

Surround drop-obligations in output with `` quotes to help unintended prefix matches

Prints scope-end messages at end of their scope.
@pnkfelix
Copy link
Member Author

cc @nikomatsakis

@pnkfelix
Copy link
Member Author

r? @nick29581

@nrc nrc self-assigned this Sep 22, 2014
@@ -93,6 +93,10 @@ impl SpanHandler {
self.handler.emit_with_code(Some((&self.cm, sp)), msg, code, Error);
self.handler.bump_err_count();
}
pub fn span_end_err(&self, sp: Span, msg: &str) {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Could you add a comment here - not obvious what this fn does from the name

@nrc
Copy link
Member

nrc commented Sep 22, 2014

I started to review this, but it quickly because apparent that I don't know the borrow checker well enough to do it justice, so going to hand off the review, sorry.

r? @nikomatsakis or @pcwalton

@nrc nrc removed their assignment Sep 22, 2014
// copied or moved depending on `mode`. Note that `matched_pat`
// is called on all variant/structs in the pattern (i.e., the
// interior nodes of the pattern's tree structure) while
// consume_pat is called on the binding identifiers in the pattern
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can you clarify whether BOTH matched_pat and consume_pat are called on leaves, or just one of them?

@pnkfelix
Copy link
Member Author

(temporarily closing pull request; I'm going to cut it down a little to remove more of the dataflow artifacts that can be left for a later PR.)

@pnkfelix pnkfelix closed this Sep 26, 2014
lnicola pushed a commit to lnicola/rust that referenced this pull request Jun 23, 2024
…ykril

Properly prime all crate def maps in parallel_prime_caches
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants