[ConstraintSystem] Revert new disjunction favoring algorithm #79128
Conversation
xedin
requested review from
a team,
hborla,
slavapestov,
ahoppen,
bnbarham,
hamishknight and
rintaro
as code owners
|
@swift-ci please test |
|
@swift-ci please test source compatibility |
xedin
force-pushed
the
revert-solver-perf
branch
from
1a8a67e to
276be15
Compare
|
@swift-ci please test |
|
@swift-ci please test source compatibility |
|
@swift-ci please test |
|
@swift-ci please test compiler performance |
|
@swift-ci please test source compatibility |
…ling with unapplied disjunctions" This reverts commit 72340f3.
… whether disjunction is supported" This reverts commit 6bc23b5.
…pported" This reverts commit 471ee21.
…s that come from implicit imports" This reverts commit aa4a2b9.
…nitializer is used" This reverts commit 3cc76ea.
… for operators" This reverts commit e3987be.
…ead of whole disjunction" This reverts commit 6c82892.
…ed literal array handling" This reverts commit cfd34e5.
…array literals" This reverts commit 8a304f8.
This reverts commit 0737542.
…operators" This reverts commit bc3a15f.
This reverts commit 860ae08.
|
What is the reason for reverting? |
|
It breaks source compatibility in a few situations. I'm planning to re-land this once I figure out how to fix it. |
There was a problem hiding this comment.
Could we keep this regression test? This behavior should be preserved regardless of algorithm, so I don't want to lose it.
There was a problem hiding this comment.
I'll bring all of the test-cases back, yeah.
The constraint solver performance changes are reverted via swiftlang/swift#79128 which means that Doggie tests are going to fail again.
xedin
force-pushed
the
revert-solver-perf
branch
from
4c6dd36 to
875a23a
Compare
|
@swift-ci please test |
xedin
force-pushed
the
revert-solver-perf
branch
from
875a23a to
41fd4de
Compare
|
@swift-ci please test |
|
Let's give this another chance. |
|
I attempted to fix the problems in #79309 but this is still running into issues because of how other things like conversion restrictions are used in the solver at the moment. |
|
@swift-ci please test Windows platform |
|
Doggie failure in Debug mode is expected since this is something that changes in question resolved. |
|
@swift-ci please test Windows platform |
1 similar comment
|
@swift-ci please test Windows platform |
…wiftlang#79128)" This reverts commit 725bd91.
…wiftlang#79128)" This reverts commit 725bd91.
…wiftlang#79128)" This reverts commit 725bd91.
…wiftlang#79128)" This reverts commit 725bd91.
…wiftlang#79128)" This reverts commit 725bd91.
…wiftlang#79128)" This reverts commit 725bd91.
…wiftlang#79128)" This reverts commit 725bd91.
…wiftlang#79128)" This reverts commit 725bd91.
…wiftlang#79128)" This reverts commit 725bd91.
No description provided.