Borrowck ast lifetimes #2771
Borrowck ast lifetimes #2771
Conversation
jdupak
force-pushed
the
borrowck-ast-lifetimes
branch
2 times, most recently
from
5377af2 to
b82aa74
Compare
There was a mismatch whether lifetime 'static is parsed as "static" or "'static". gcc/rust/ChangeLog: * parse/rust-parse-impl.h (Parser::lifetime_from_token): Fix matched pattern. Signed-off-by: Jakub Dupak <[email protected]>
gcc/rust/ChangeLog: * ast/rust-ast.h: Elided lifetime static constructor * ast/rust-type.h: Default lifetime to elided. * parse/rust-parse-impl.h (Parser::parse_lifetime_param): Use elided lifetime. (Parser::parse_lifetime): Use elided lifetime/ (Parser::lifetime_from_token): Use elided lifetime. (Parser::parse_self_param): Use elided lifetime. (Parser::parse_reference_type_inner): Use elided lifetime. Signed-off-by: Jakub Dupak <[email protected]>
gcc/rust/ChangeLog: * parse/rust-parse-impl.h (Parser::parse_generic_param): Lifetime elision control. (Parser::parse_lifetime_where_clause_item): Lifetime elision control. (Parser::parse_type_param_bound): Lifetime elision control. (Parser::parse_lifetime_bounds): Lifetime elision control. (Parser::parse_lifetime): Lifetime elision control. (Parser::parse_path_generic_args): Lifetime elision control. (Parser::parse_self_param): Lifetime elision control. (Parser::parse_break_expr): Lifetime elision control. (Parser::parse_continue_expr): Lifetime elision control. (Parser::parse_reference_type_inner): Lifetime elision control. * parse/rust-parse.h: Lifetime elision control. Signed-off-by: Jakub Dupak <[email protected]>
(probably incomplete propagation) gcc/rust/ChangeLog: * hir/rust-ast-lower-base.cc (ASTLoweringBase::lower_lifetime): Propagate static requirement. * hir/rust-ast-lower-base.h: Propagate static requirement. * hir/rust-ast-lower-implitem.h: Propagate static requirement. * hir/rust-ast-lower-item.cc (ASTLoweringItem::visit): Propagate static requirement. * hir/rust-ast-lower-type.cc (ASTLoweringType::translate): Propagate static requirement. (ASTLoweringType::visit): Propagate static requirement. * hir/rust-ast-lower-type.h: Propagate static requirement. Signed-off-by: Jakub Dupak <[email protected]>
gcc/rust/ChangeLog: * hir/rust-ast-lower-type.cc (ASTLoweringType::visit): For lifetimes. Signed-off-by: Jakub Dupak <[email protected]>
jdupak
force-pushed
the
borrowck-ast-lifetimes
branch
from
b82aa74 to
71f7686
Compare
| if (lifetime_type == AST::Lifetime::WILDCARD && default_to_static_lifetime) | ||
| { | ||
| // If compiling in a static context. | ||
| lifetime_type = AST::Lifetime::STATIC; | ||
| } |
There was a problem hiding this comment.
| if (lifetime_type == AST::Lifetime::WILDCARD && default_to_static_lifetime) | |
| { | |
| // If compiling in a static context. | |
| lifetime_type = AST::Lifetime::STATIC; | |
| } | |
| // If compiling in a static context. | |
| if (lifetime_type == AST::Lifetime::WILDCARD && default_to_static_lifetime) | |
| lifetime_type = AST::Lifetime::STATIC; |
GNU style nit, sorry
| {} | ||
|
|
||
| /** Used when compiling const and static items. */ | ||
| bool default_to_static_lifetime; |
There was a problem hiding this comment.
I don't get how this flag is used or set. If we are lowering an AST::ConstantItem or AST::StaticItem, we know we are doing so and can pass an explicit true flag to ASTLoweringType (or even better, a LifetimeDefault::Static enum variant which is clearer). and we could have a default value to the bool argument to the ASTLoweringType argument?
does that make sense? or is it important that we keep it this way for future pull-requests?
There was a problem hiding this comment.
This flag is used to get the state across the visitor API and propagate further ASTLoweringType::translate calls since the visitor API does not allow more parameters. I think it makes sense. Does it make sense now?
There was a problem hiding this comment.
ah, yeah I see what you mean. so we really can't do without a default parameter then?
There was a problem hiding this comment.
Do you mean the member variable or the parameter? We need the member variable, the parameter is useful because we don't need to change it everywhere.
There was a problem hiding this comment.
I meant the parameter - but I'm fine with keeping it this way, your explanation makes sense. I'll merge this now
Depends on #2770
Fixes #2756