typeck::check::coercion - roll back failed unsizing type vars #44743
Conversation
|
@bors r+ Nice catch! |
|
📌 Commit 3b9e252 has been approved by |
alexcrichton
added
the
S-waiting-on-bors
|
@bors p=1 (perf sensitive) |
|
⌛ Testing commit 3b9e252fb4005c63e5ff0ede06fb5c1c48f1eaa3 with merge ce5534d1854cca6e8f1f8f7bb2a29ba48d8b51ee... |
|
💔 Test failed - status-appveyor |
|
Error emission order of error[E0623]: lifetime mismatch
- --> $DIR/ex3-both-anon-regions-3.rs:12:13
+ --> $DIR/ex3-both-anon-regions-3.rs:12:15
|
11 | fn foo(z: &mut Vec<(&u8,&u8)>, (x, y): (&u8, &u8)) {
- | --- --- these two types are declared with different lifetimes...
+ | --- --- these two types are declared with different lifetimes...
12 | z.push((x,y));
- | ^ ...but data flows into `z` here
+ | ^ ...but data flows into `z` here
error[E0623]: lifetime mismatch
- --> $DIR/ex3-both-anon-regions-3.rs:12:15
+ --> $DIR/ex3-both-anon-regions-3.rs:12:13
|
11 | fn foo(z: &mut Vec<(&u8,&u8)>, (x, y): (&u8, &u8)) {
- | --- --- these two types are declared with different lifetimes...
+ | --- --- these two types are declared with different lifetimes...
12 | z.push((x,y));
- | ^ ...but data flows into `z` here
+ | ^ ...but data flows into `z` here
error: aborting due to 2 previous errors |
|
The order is "flaky" in the sense that it depends on region inference order. Sorting these is non-trivial so I'll just fix up the test. We'll rewrite the region inference code soon enough anyway. |
This wraps unsizing coercions within an additional level of `commit_if_ok`, which rolls back type variables if the unsizing coercion fails. This prevents a large amount of type-variables from accumulating while type-checking a large function, e.g. shaving 2GB off one of the 4GB peaks in rust-lang#36799.
this should produce more error stability
arielb1
force-pushed
the
size-rollback
branch
from
3b9e252
to
b6bce56
Compare
|
@bors r=eddyb |
|
📌 Commit b6bce56 has been approved by |
|
⌛ Testing commit b6bce56 with merge 001aa73342ff270819b11cb869a32c863ee4f209... |
|
💔 Test failed - status-travis |
|
on dist-x86_64-linux:
@bors retry |
typeck::check::coercion - roll back failed unsizing type vars This wraps unsizing coercions within an additional level of `commit_if_ok`, which rolls back type variables if the unsizing coercion fails. This prevents a large amount of type-variables from accumulating while type-checking a large function, e.g. shaving 2GB off one of the 4GB peaks in #36799. This is a performance-sensitive PR so please don't roll it up. r? @eddyb cc @nikomatsakis
|
☀️ Test successful - status-appveyor, status-travis |
encode region::Scope using fewer bytes Now that region::Scope is no longer interned, its size is more important. This PR encodes region::Scope in 8 bytes instead of 12, which should speed up region inference somewhat (perf testing needed) and should improve the margins on #36799 by 64MB (that's not a lot, I did this PR mostly to speed up region inference). This is a perf-sensitive PR. Please don't roll me up. r? @eddyb This is based on #44743 so I could get more accurate measurements on #36799.
|
This improved compile time of the big-array-of-strings benchmark by 6% Nice! |
|
This also decreased memory usage of the same benchmark by 50% |
This wraps unsizing coercions within an additional level of
commit_if_ok, which rolls back type variables if the unsizing coercionfails. This prevents a large amount of type-variables from accumulating
while type-checking a large function, e.g. shaving 2GB off one of the
4GB peaks in #36799.
This is a performance-sensitive PR so please don't roll it up.
r? @eddyb
cc @nikomatsakis