You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Now that we have a type hierarchy (from bug 22464), we don't need DW_TAG_expression to discriminate DWARF expressions. We can use Metadata::getMetadataID() instead.
Change MDExpression::getTag() to return 0u and fix any fallout.
Remove dwarf::DW_TAG_expression.
Change MDExpression to inherit directly from MDNode instead of DebugNode.
Investigate having it inherit directly from Metadata -- expressions don't need/use metadata operands, they're not referenced from NamedMDNode or !attachments, and they have no need for distinct/temporary semantics.
The text was updated successfully, but these errors were encountered:
There aren't very many MDExpressions -- I haven't seen MDExpression in any memory profile -- so step #4 doesn't seem urgent. Since r232550 fixed the real problem, leaving the PR around is just noise at this point. Closing.
I started implementing step #4 of making DIExpression not be an MDNode in https://reviews.llvm.org/D37281, but I'm going to abandon it. It doesn't seem worth it.
Extended Description
Now that we have a type hierarchy (from bug 22464), we don't need
DW_TAG_expression
to discriminate DWARF expressions. We can useMetadata::getMetadataID()
instead.MDExpression::getTag()
to return0u
and fix any fallout.dwarf::DW_TAG_expression
.MDExpression
to inherit directly fromMDNode
instead ofDebugNode
.Metadata
-- expressions don't need/use metadata operands, they're not referenced fromNamedMDNode
or!attachments
, and they have no need for distinct/temporary semantics.The text was updated successfully, but these errors were encountered: