Age | Commit message (Collapse) | Author |
|
local types
Rather than using a local types' position information, use the type
name's scope numbers to uniquely identify the type from others with
the same name.
We use scope numbers rather than indices (with number-1 == index)
to preserve the invariant that the zero value for a scope is a ready
to use empty scope.
Using scope numbers turned out to be fairly simple after all and
provides a reasonably stable identification which will make debugging
simpler. A scope number series may be a bit longer than a unique ID for
each type name but local types should be reasonably rare.
Also did a minor cleanup in universe.go to ensure Named.orig is set up
correctly (there's still an open TODO but with a work-around).
Change-Id: I73935fa9bd960809fd5c95fe8b8a535c313cfc8f
Reviewed-on: https://go-review.googlesource.com/c/go/+/333192
Trust: Robert Griesemer <gri@golang.org>
Run-TryBot: Robert Griesemer <gri@golang.org>
Reviewed-by: Robert Findley <rfindley@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
|
|
This CL adds three new functions to the types2 API to support lazy
import resolution:
1. A new Scope.InsertLazy method to allow recording that Objects exist
in a particular Scope (in particular, package scopes) without having
to yet fully construct those objects. Instead, types2 will call the
provided `resolve` function if/when the object is actually needed.
2. Similarly, a new NewTypeNameLazy function to create TypeName
objects without yet instantiating their underlying Named
instance.
3. Finally, an InstantiateLazy method, that allows creating type
instances without requiring any of the types to be expanded right
away. Importantly, this requires providing a types2.Checker argument
to handle recursive types correctly.
The APIs as-is are a bit clumsy (esp. NewTypeNameLazy), but seem to
work well for cmd/compile's needs. In particular, they simplify some
of the complexities of handling recursive type definitions within the
importer.
Also, the current prototype is a bit fragile. It uses sync.Once to
manage concurrent lazy resolution, which is frustrating to debug in
the presence of reentrancy issues. It also means the importer needs to
deal with concurrency as well. These aren't issues for types2 though
as cmd/compile only walks the type-checked AST sequentially.
Finally, it looks like some of the details of lazy type names are
similar to the lazy "instance" stuff used for generics, so maybe
there's opportunity for unifying them under a more general (but still
internal) lazy type mechanism.
I had originally intended for this CL to also update the types2
importer, but (1) it doesn't have access to the types2.Checker
instance needed to call InstantiateLazy, and (2) it creates a new
TypeName/TypeParam at each use rather than reusing them, which
evidently works with types2.Instantiate but not
types2.(*Checker).instantiate (i.e., InstantiateLazy). I spent a while
trying to fix these issues, but kept running into more subtle
issues. Instead, I've included my WIP "unified IR" CL as a followup CL
that demonstrates these Lazy methods (see noder/reader2.go).
Updates #46449.
Change-Id: I4d1e8e649f6325a11790d25fd90c39fa07c8d41d
Reviewed-on: https://go-review.googlesource.com/c/go/+/323569
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Trust: Matthew Dempsky <mdempsky@google.com>
Trust: Robert Griesemer <gri@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
|
|
Make position comparison generally available.
Change-Id: I94b6f658fa19a15b30574dbb2181879115c131a8
Reviewed-on: https://go-review.googlesource.com/c/go/+/307215
Trust: Robert Griesemer <gri@golang.org>
Reviewed-by: Robert Findley <rfindley@google.com>
|
|
This file has a few changes compared to the go/types version:
1) syntax.Pos is used instead of token.Pos.
2) The cmpPos helper function (defined elsewhere) is used to
compare positions (syntax.Pos positions cannot be compared
directly with <=).
3) A new method Scope.Squash was added (primary difference).
f=scope.go; diff $f ../../../../go/types/$f
7c7
< package types2
---
> package types
11d10
< "cmd/compile/internal/syntax"
12a12
> "go/token"
26c26
< pos, end syntax.Pos // scope extent; may be invalid
---
> pos, end token.Pos // scope extent; may be invalid
33c33
< func NewScope(parent *Scope, pos, end syntax.Pos, comment string) *Scope {
---
> func NewScope(parent *Scope, pos, end token.Pos, comment string) *Scope {
82c82
< func (s *Scope) LookupParent(name string, pos syntax.Pos) (*Scope, Object) {
---
> func (s *Scope) LookupParent(name string, pos token.Pos) (*Scope, Object) {
84c84
< if obj := s.elems[name]; obj != nil && (!pos.IsKnown() || cmpPos(obj.scopePos(), pos) <= 0) {
---
> if obj := s.elems[name]; obj != nil && (!pos.IsValid() || obj.scopePos() <= pos) {
111,144d110
< // Squash merges s with its parent scope p by adding all
< // objects of s to p, adding all children of s to the
< // children of p, and removing s from p's children.
< // The function f is called for each object obj in s which
< // has an object alt in p. s should be discarded after
< // having been squashed.
< func (s *Scope) Squash(err func(obj, alt Object)) {
< p := s.parent
< assert(p != nil)
< for _, obj := range s.elems {
< obj.setParent(nil)
< if alt := p.Insert(obj); alt != nil {
< err(obj, alt)
< }
< }
<
< j := -1 // index of s in p.children
< for i, ch := range p.children {
< if ch == s {
< j = i
< break
< }
< }
< assert(j >= 0)
< k := len(p.children) - 1
< p.children[j] = p.children[k]
< p.children = p.children[:k]
<
< p.children = append(p.children, s.children...)
<
< s.children = nil
< s.elems = nil
< }
<
149,150c115,116
< func (s *Scope) Pos() syntax.Pos { return s.pos }
< func (s *Scope) End() syntax.Pos { return s.end }
---
> func (s *Scope) Pos() token.Pos { return s.pos }
> func (s *Scope) End() token.Pos { return s.end }
155,156c121,122
< func (s *Scope) Contains(pos syntax.Pos) bool {
< return cmpPos(s.pos, pos) <= 0 && cmpPos(pos, s.end) < 0
---
> func (s *Scope) Contains(pos token.Pos) bool {
> return s.pos <= pos && pos < s.end
164c130
< func (s *Scope) Innermost(pos syntax.Pos) *Scope {
---
> func (s *Scope) Innermost(pos token.Pos) *Scope {
Change-Id: If6c459f45dae8980ffb3a902a46b1700e9b55dc7
Reviewed-on: https://go-review.googlesource.com/c/go/+/265700
Trust: Robert Griesemer <gri@golang.org>
Reviewed-by: Robert Findley <rfindley@google.com>
|
|
types2 and importer
This is a copy of the importer and types2 (unreviewed) prototype version
excluding the testdata directory containing tests (see below). Each file
is marked with the comment
// UNREVIEWED
on the first line. The plan is to check in this code wholesale (it runs and
passes all tests) and then review the code file-by-file via subsequent CLs
and remove the "// UNREVIEWED" comments as we review the files.
Since most tests are unchanged from the original go/types, the next CL will
commit those tests as they don't need to be reviewed again. (Eventually we
may want to factor them out and share them from a single place, e.g. the
test directory.)
The existing file fmtmap_test.go was updated.
Change-Id: I9bd0ad1a7e7188b501423483a44d18e623c0fe71
Reviewed-on: https://go-review.googlesource.com/c/go/+/263624
Trust: Robert Griesemer <gri@golang.org>
Trust: Keith Randall <khr@golang.org>
Run-TryBot: Robert Griesemer <gri@golang.org>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
|