1
0
mirror of https://github.com/golang/go synced 2024-11-19 00:44:40 -07:00
go/oracle/serial/serial.go

222 lines
9.6 KiB
Go
Raw Normal View History

go.tools/oracle: add option to output results in JSON syntax. See json.go for interface specification. Example usage: % oracle -format=json -mode=callgraph code.google.com/p/go.tools/cmd/oracle + Tests, based on (small) golden files. Overview: Each <query>Result structure has been "lowered" so that all but the most trivial logic in each display() function has been moved to the main query. Each one now has a toJSON method that populates a json.Result struct. Though the <query>Result structs are similar to the correponding JSON protocol, they're not close enough to be used directly; for example, the former contain richer semantic entities (token.Pos, ast.Expr, ssa.Value, pointer.Pointer, etc) whereas JSON contains only their printed forms using Go basic types. The choices of what levels of abstractions the two sets of structs should have is somewhat arbitrary. We may want richer information in the JSON output in future. Details: - oracle.Main has been split into oracle.Query() and the printing of the oracle.Result. - the display() method no longer needs an *oracle param, only a print function. - callees: sort the result for determinism. - callees: compute the union across all contexts. - callers: sort the results for determinism. - describe(package): fixed a bug in the predicate for method accessibility: an unexported method defined in pkg A may belong to a type defined in package B (via embedding/promotion) and may thus be accessible to A. New accessibleMethods() utility fixes this. - describe(type): filter methods by accessibility. - added tests of 'callgraph'. - pointer: eliminated the 'caller CallGraphNode' parameter from pointer.Context.Call callback since it was redundant w.r.t site.Caller(). - added warning if CGO_ENABLED is unset. R=crawshaw CC=golang-dev https://golang.org/cl/13270045
2013-09-03 13:29:02 -06:00
// Copyright 2013 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Package serial defines the oracle's schema for structured data
// serialization using JSON, XML, etc.
package serial
go.tools/oracle: add option to output results in JSON syntax. See json.go for interface specification. Example usage: % oracle -format=json -mode=callgraph code.google.com/p/go.tools/cmd/oracle + Tests, based on (small) golden files. Overview: Each <query>Result structure has been "lowered" so that all but the most trivial logic in each display() function has been moved to the main query. Each one now has a toJSON method that populates a json.Result struct. Though the <query>Result structs are similar to the correponding JSON protocol, they're not close enough to be used directly; for example, the former contain richer semantic entities (token.Pos, ast.Expr, ssa.Value, pointer.Pointer, etc) whereas JSON contains only their printed forms using Go basic types. The choices of what levels of abstractions the two sets of structs should have is somewhat arbitrary. We may want richer information in the JSON output in future. Details: - oracle.Main has been split into oracle.Query() and the printing of the oracle.Result. - the display() method no longer needs an *oracle param, only a print function. - callees: sort the result for determinism. - callees: compute the union across all contexts. - callers: sort the results for determinism. - describe(package): fixed a bug in the predicate for method accessibility: an unexported method defined in pkg A may belong to a type defined in package B (via embedding/promotion) and may thus be accessible to A. New accessibleMethods() utility fixes this. - describe(type): filter methods by accessibility. - added tests of 'callgraph'. - pointer: eliminated the 'caller CallGraphNode' parameter from pointer.Context.Call callback since it was redundant w.r.t site.Caller(). - added warning if CGO_ENABLED is unset. R=crawshaw CC=golang-dev https://golang.org/cl/13270045
2013-09-03 13:29:02 -06:00
// All 'pos' strings are of the form "file:line:col".
// TODO(adonovan): improve performance by sharing filename strings.
// TODO(adonovan): improve precision by providing the start/end
// interval when available.
//
// TODO(adonovan): consider richer encodings of types, functions,
// methods, etc.
// A Peers is the result of a 'peers' query.
// If Allocs is empty, the selected channel can't point to anything.
type Peers struct {
Pos string `json:"pos"` // location of the selected channel op (<-)
Type string `json:"type"` // type of the selected channel
Allocs []string `json:"allocs,omitempty"` // locations of aliased make(chan) ops
Sends []string `json:"sends,omitempty"` // locations of aliased ch<-x ops
Receives []string `json:"receives,omitempty"` // locations of aliased <-ch ops
}
// A Referrers is the result of a 'referrers' query.
type Referrers struct {
Pos string `json:"pos"` // location of the query reference
ObjPos string `json:"objpos,omitempty"` // location of the definition
Desc string `json:"desc"` // description of the denoted object
Refs []string `json:"refs,omitempty"` // locations of all references
}
go.tools/oracle: add option to output results in JSON syntax. See json.go for interface specification. Example usage: % oracle -format=json -mode=callgraph code.google.com/p/go.tools/cmd/oracle + Tests, based on (small) golden files. Overview: Each <query>Result structure has been "lowered" so that all but the most trivial logic in each display() function has been moved to the main query. Each one now has a toJSON method that populates a json.Result struct. Though the <query>Result structs are similar to the correponding JSON protocol, they're not close enough to be used directly; for example, the former contain richer semantic entities (token.Pos, ast.Expr, ssa.Value, pointer.Pointer, etc) whereas JSON contains only their printed forms using Go basic types. The choices of what levels of abstractions the two sets of structs should have is somewhat arbitrary. We may want richer information in the JSON output in future. Details: - oracle.Main has been split into oracle.Query() and the printing of the oracle.Result. - the display() method no longer needs an *oracle param, only a print function. - callees: sort the result for determinism. - callees: compute the union across all contexts. - callers: sort the results for determinism. - describe(package): fixed a bug in the predicate for method accessibility: an unexported method defined in pkg A may belong to a type defined in package B (via embedding/promotion) and may thus be accessible to A. New accessibleMethods() utility fixes this. - describe(type): filter methods by accessibility. - added tests of 'callgraph'. - pointer: eliminated the 'caller CallGraphNode' parameter from pointer.Context.Call callback since it was redundant w.r.t site.Caller(). - added warning if CGO_ENABLED is unset. R=crawshaw CC=golang-dev https://golang.org/cl/13270045
2013-09-03 13:29:02 -06:00
type CalleesItem struct {
Name string `json:"name"` // full name of called function
Pos string `json:"pos"` // location of called function
}
// A Callees is the result of a 'callees' query.
//
// Callees is nonempty unless the call was a dynamic call on a
// provably nil func or interface value.
type Callees struct {
Pos string `json:"pos"` // location of selected call site
Desc string `json:"desc"` // description of call site
Callees []*CalleesItem `json:"callees,omitempty"` // set of possible call targets
}
// A Caller is one element of the slice returned by a 'callers' query.
// (Callstack also contains a similar slice.)
//
// The root of the callgraph has an unspecified "Caller" string.
type Caller struct {
Pos string `json:"pos,omitempty"` // location of the calling function
Desc string `json:"desc"` // description of call site
Caller string `json:"caller"` // full name of calling function
}
// A CallGraph is one element of the slice returned by a 'callgraph' query.
// The index of each item in the slice is used to identify it in the
// Callers adjacency list.
//
// Multiple nodes may have the same Name due to context-sensitive
// treatment of some functions.
//
// TODO(adonovan): perhaps include edge labels (i.e. callsites).
type CallGraph struct {
Name string `json:"name"` // full name of function
Pos string `json:"pos"` // location of function
Children []int `json:"children,omitempty"` // indices of child nodes in callgraph list
}
// A CallStack is the result of a 'callstack' query.
// It indicates an arbitrary path from the root of the callgraph to
// the query function.
//
// If the Callers slice is empty, the function was unreachable in this
// analysis scope.
type CallStack struct {
Pos string `json:"pos"` // location of the selected function
Target string `json:"target"` // the selected function
Callers []Caller `json:"callers"` // enclosing calls, innermost first.
}
// A FreeVar is one element of the slice returned by a 'freevars'
// query. Each one identifies an expression referencing a local
// identifier defined outside the selected region.
type FreeVar struct {
Pos string `json:"pos"` // location of the identifier's definition
Kind string `json:"kind"` // one of {var,func,type,const,label}
Ref string `json:"ref"` // referring expression (e.g. "x" or "x.y.z")
Type string `json:"type"` // type of the expression
}
// An Implements is one element of the result of an 'implements' query.
// Each one indicates a row in the "implements" relation over
// package-level named types defined by the package containing the
// selection.
type Implements struct {
I string `json:"i"` // full name of the interface type
IPos string `json:"ipos"` // location of its definition
C string `json:"c"` // full name of the concrete type
CPos string `json:"cpos"` // location of its definition
}
// A DescribePTALabel describes a pointer analysis label.
//
// A "label" is an object that may be pointed to by a pointer, map,
// channel, 'func', slice or interface. Labels include:
// - functions
// - globals
// - arrays created by literals (e.g. []byte("foo")) and conversions ([]byte(s))
// - stack- and heap-allocated variables (including composite literals)
// - arrays allocated by append()
// - channels, maps and arrays created by make()
// - and their subelements, e.g. "alloc.y[*].z"
//
type DescribePTALabel struct {
Pos string `json:"pos"` // location of syntax that allocated the object
Desc string `json:"desc"` // description of the label
}
// A DescribePointer describes a single pointer: its type and the
// set of "labels" it points to.
//
type DescribePointer struct {
Type string `json:"type"` // (concrete) type of the pointer
NamePos string `json:"namepos,omitempty"` // location of type defn, if Named
Labels []DescribePTALabel `json:"labels,omitempty"` // pointed-to objects
go.tools/oracle: add option to output results in JSON syntax. See json.go for interface specification. Example usage: % oracle -format=json -mode=callgraph code.google.com/p/go.tools/cmd/oracle + Tests, based on (small) golden files. Overview: Each <query>Result structure has been "lowered" so that all but the most trivial logic in each display() function has been moved to the main query. Each one now has a toJSON method that populates a json.Result struct. Though the <query>Result structs are similar to the correponding JSON protocol, they're not close enough to be used directly; for example, the former contain richer semantic entities (token.Pos, ast.Expr, ssa.Value, pointer.Pointer, etc) whereas JSON contains only their printed forms using Go basic types. The choices of what levels of abstractions the two sets of structs should have is somewhat arbitrary. We may want richer information in the JSON output in future. Details: - oracle.Main has been split into oracle.Query() and the printing of the oracle.Result. - the display() method no longer needs an *oracle param, only a print function. - callees: sort the result for determinism. - callees: compute the union across all contexts. - callers: sort the results for determinism. - describe(package): fixed a bug in the predicate for method accessibility: an unexported method defined in pkg A may belong to a type defined in package B (via embedding/promotion) and may thus be accessible to A. New accessibleMethods() utility fixes this. - describe(type): filter methods by accessibility. - added tests of 'callgraph'. - pointer: eliminated the 'caller CallGraphNode' parameter from pointer.Context.Call callback since it was redundant w.r.t site.Caller(). - added warning if CGO_ENABLED is unset. R=crawshaw CC=golang-dev https://golang.org/cl/13270045
2013-09-03 13:29:02 -06:00
}
// A DescribeValue is the additional result of a 'describe' query
// if the selection indicates a value or expression.
//
// If the described value is an interface, it will have one PTS entry
// describing each concrete type that it may contain. For each
// concrete type that is a pointer, the PTS entry describes the labels
go.tools/pointer: reflection, part 1: maps, and some core features. Core: reflect.TypeOf reflect.ValueOf reflect.Zero reflect.Value.Interface Maps: (reflect.Value).MapIndex (reflect.Value).MapKeys (reflect.Value).SetMapIndex (*reflect.rtype).Elem (*reflect.rtype).Key + tests: pointer/testdata/mapreflect.go. oracle/testdata/src/main/reflection.go. Interface objects (T, V...) have been renamed "tagged objects". Abstraction: we model reflect.Value similar to interface{}---as a pointer that points only to tagged objects---but a reflect.Value may also point to an "indirect tagged object", one in which the payload V is of type *T not T. These are required because reflect.Values can hold lvalues, e.g. when derived via Field() or Elem(), though we won't use them till we get to structs and pointers. Solving: each reflection intrinsic defines a new constraint and resolution rule. Because of the nature of reflection, generalizing across types, the resolution rules dynamically create additional complex constraints during solving, where previously only simple (copy) constraints were created. This requires some solver changes: The work done before the main solver loop (to attach new constraints to the graph) is now done before each iteration, in processNewConstraints. Its loop over constraints is broken into two passes: the first handles base (addr-of) constraints, the second handles simple and complex constraints. constraint.init() has been inlined. The only behaviour that varies across constraints is ptr() Sadly this will pessimize presolver optimisations, when we get there; such is the price of reflection. Objects: reflection intrinsics create objects (i.e. cause memory allocations) with no SSA operation. We will represent them as the cgnode of the instrinsic (e.g. reflect.New), so we extend Labels and node.data to represent objects as a product (not sum) of ssa.Value and cgnode and pull this out into its own type, struct object. This simplifies a number of invariants and saves space. The ntObject flag is now represented by obj!=nil; the other flags are moved into object. cgnodes are now always recorded in objects/Labels for which it is appropriate (all but those for globals, constants and the shared contours for functions). Also: - Prepopulate the flattenMemo cache to consider reflect.Value a fake pointer, not a struct. - Improve accessors and documentation on type Label. - @conctypes assertions renamed @types (since dyn. types needn't be concrete). - add oracle 'describe' test on an interface (missing, an oversight). R=crawshaw CC=golang-dev https://golang.org/cl/13418048
2013-09-16 07:49:10 -06:00
// it may point to. The same is true for reflect.Values, except the
// dynamic types needn't be concrete.
go.tools/oracle: add option to output results in JSON syntax. See json.go for interface specification. Example usage: % oracle -format=json -mode=callgraph code.google.com/p/go.tools/cmd/oracle + Tests, based on (small) golden files. Overview: Each <query>Result structure has been "lowered" so that all but the most trivial logic in each display() function has been moved to the main query. Each one now has a toJSON method that populates a json.Result struct. Though the <query>Result structs are similar to the correponding JSON protocol, they're not close enough to be used directly; for example, the former contain richer semantic entities (token.Pos, ast.Expr, ssa.Value, pointer.Pointer, etc) whereas JSON contains only their printed forms using Go basic types. The choices of what levels of abstractions the two sets of structs should have is somewhat arbitrary. We may want richer information in the JSON output in future. Details: - oracle.Main has been split into oracle.Query() and the printing of the oracle.Result. - the display() method no longer needs an *oracle param, only a print function. - callees: sort the result for determinism. - callees: compute the union across all contexts. - callers: sort the results for determinism. - describe(package): fixed a bug in the predicate for method accessibility: an unexported method defined in pkg A may belong to a type defined in package B (via embedding/promotion) and may thus be accessible to A. New accessibleMethods() utility fixes this. - describe(type): filter methods by accessibility. - added tests of 'callgraph'. - pointer: eliminated the 'caller CallGraphNode' parameter from pointer.Context.Call callback since it was redundant w.r.t site.Caller(). - added warning if CGO_ENABLED is unset. R=crawshaw CC=golang-dev https://golang.org/cl/13270045
2013-09-03 13:29:02 -06:00
//
type DescribeValue struct {
Type string `json:"type"` // type of the expression
Value string `json:"value,omitempty"` // value of the expression, if constant
ObjPos string `json:"objpos,omitempty"` // location of the definition, if an Ident
PTAErr string `json:"ptaerr,omitempty"` // reason pointer analysis wasn't attempted
PTS []*DescribePointer `json:"pts,omitempty"` // points-to set; an interface may have many
}
type DescribeMethod struct {
Name string `json:"name"` // method name, as defined by types.Selection.String()
Pos string `json:"pos"` // location of the method's definition
}
// A DescribeType is the additional result of a 'describe' query
// if the selection indicates a type.
type DescribeType struct {
Type string `json:"type"` // the string form of the type
NamePos string `json:"namepos,omitempty"` // location of definition of type, if named
NameDef string `json:"namedef,omitempty"` // underlying definition of type, if named
Methods []DescribeMethod `json:"methods,omitempty"` // methods of the type
}
type DescribeMember struct {
Name string `json:"name"` // name of member
Type string `json:"type,omitempty"` // type of member (underlying, if 'type')
Value string `json:"value,omitempty"` // value of member (if 'const')
Pos string `json:"pos"` // location of definition of member
Kind string `json:"kind"` // one of {var,const,func,type}
Methods []DescribeMethod `json:"methods,omitempty"` // methods (if member is a type)
}
// A DescribePackage is the additional result of a 'describe' if
// the selection indicates a package.
type DescribePackage struct {
Path string `json:"path"` // import path of the package
Members []*DescribeMember `json:"members,omitempty"` // accessible members of the package
}
// A Describe is the result of a 'describe' query.
// It may contain an element describing the selected semantic entity
// in detail.
type Describe struct {
Desc string `json:"desc"` // description of the selected syntax node
Pos string `json:"pos"` // location of the selected syntax node
Detail string `json:"detail,omitempty"` // one of {package, type, value}, or "".
// At most one of the following fields is populated:
// the one specified by 'detail'.
Package *DescribePackage `json:"package,omitempty"`
Type *DescribeType `json:"type,omitempty"`
Value *DescribeValue `json:"value,omitempty"`
}
type PTAWarning struct {
Pos string `json:"pos"` // location associated with warning
Message string `json:"message"` // warning message
}
// A Result is the common result of any oracle query.
// It contains a query-specific result element.
//
// TODO(adonovan): perhaps include other info such as: analysis scope,
// raw query position, stack of ast nodes, query package, etc.
type Result struct {
Mode string `json:"mode"` // mode of the query
// Exactly one of the following fields is populated:
// the one specified by 'mode'.
Callees *Callees `json:"callees,omitempty"`
Callers []Caller `json:"callers,omitempty"`
Callgraph []CallGraph `json:"callgraph,omitempty"`
Callstack *CallStack `json:"callstack,omitempty"`
Describe *Describe `json:"describe,omitempty"`
Freevars []*FreeVar `json:"freevars,omitempty"`
Implements []*Implements `json:"implements,omitempty"`
Peers *Peers `json:"peers,omitempty"`
Referrers *Referrers `json:"referrers,omitempty"`
go.tools/oracle: add option to output results in JSON syntax. See json.go for interface specification. Example usage: % oracle -format=json -mode=callgraph code.google.com/p/go.tools/cmd/oracle + Tests, based on (small) golden files. Overview: Each <query>Result structure has been "lowered" so that all but the most trivial logic in each display() function has been moved to the main query. Each one now has a toJSON method that populates a json.Result struct. Though the <query>Result structs are similar to the correponding JSON protocol, they're not close enough to be used directly; for example, the former contain richer semantic entities (token.Pos, ast.Expr, ssa.Value, pointer.Pointer, etc) whereas JSON contains only their printed forms using Go basic types. The choices of what levels of abstractions the two sets of structs should have is somewhat arbitrary. We may want richer information in the JSON output in future. Details: - oracle.Main has been split into oracle.Query() and the printing of the oracle.Result. - the display() method no longer needs an *oracle param, only a print function. - callees: sort the result for determinism. - callees: compute the union across all contexts. - callers: sort the results for determinism. - describe(package): fixed a bug in the predicate for method accessibility: an unexported method defined in pkg A may belong to a type defined in package B (via embedding/promotion) and may thus be accessible to A. New accessibleMethods() utility fixes this. - describe(type): filter methods by accessibility. - added tests of 'callgraph'. - pointer: eliminated the 'caller CallGraphNode' parameter from pointer.Context.Call callback since it was redundant w.r.t site.Caller(). - added warning if CGO_ENABLED is unset. R=crawshaw CC=golang-dev https://golang.org/cl/13270045
2013-09-03 13:29:02 -06:00
Warnings []PTAWarning `json:"warnings,omitempty"` // warnings from pointer analysis
}