1
0
mirror of https://github.com/golang/go synced 2024-10-01 12:28:37 -06:00
go/oracle/implements.go

330 lines
8.9 KiB
Go
Raw Normal View History

// 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 oracle
import (
"fmt"
"go/ast"
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
"go/token"
"reflect"
"sort"
"strings"
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
oracle: several major improvements Features: More robust: silently ignore type errors in modes that don't need SSA form: describe, referrers, implements, freevars, description. This makes the tool much more robust for everyday queries. Less configuration: don't require a scope argument for all queries. Only queries that do pointer analysis need it. For the rest, the initial position is enough for importQueryPackage to deduce the scope. It now works for queries in GoFiles, TestGoFiles, or XTestGoFiles. (It no longer works for ad-hoc main packages like $GOROOT/src/net/http/triv.go) More complete: "referrers" computes the scope automatically by scanning the import graph of the entire workspace, using gorename's refactor/importgraph package. This requires two passes at loading. Faster: simplified start-up logic avoids unnecessary package loading and SSA construction (a consequence of bad abstraction) in many cases. "callgraph": remove it. Unlike all the other commands it isn't related to the current selection, and we have golang.org/x/tools/cmdcallgraph now. Internals: Drop support for long-running clients (i.e., Pythia), since godoc -analysis supports all the same features except "pointsto", and precomputes all the results so latency is much lower. Get rid of various unhelpful abstractions introduced to support long-running clients. Expand out the set-up logic for each subcommand. This is simpler, easier to read, and gives us more control, at a small cost in duplication---the familiar story of abstractions. Discard PTA warnings. We weren't showing them (nor should we). Split tests into separate directories (so that importgraph works). Change-Id: I55d46b3ab33cdf7ac22436fcc2148fe04c901237 Reviewed-on: https://go-review.googlesource.com/8243 Reviewed-by: David Crawshaw <crawshaw@golang.org>
2015-03-30 09:21:48 -06:00
"golang.org/x/tools/go/loader"
"golang.org/x/tools/go/types"
"golang.org/x/tools/oracle/serial"
)
// Implements displays the "implements" relation as it pertains to the
oracle: several major improvements Features: More robust: silently ignore type errors in modes that don't need SSA form: describe, referrers, implements, freevars, description. This makes the tool much more robust for everyday queries. Less configuration: don't require a scope argument for all queries. Only queries that do pointer analysis need it. For the rest, the initial position is enough for importQueryPackage to deduce the scope. It now works for queries in GoFiles, TestGoFiles, or XTestGoFiles. (It no longer works for ad-hoc main packages like $GOROOT/src/net/http/triv.go) More complete: "referrers" computes the scope automatically by scanning the import graph of the entire workspace, using gorename's refactor/importgraph package. This requires two passes at loading. Faster: simplified start-up logic avoids unnecessary package loading and SSA construction (a consequence of bad abstraction) in many cases. "callgraph": remove it. Unlike all the other commands it isn't related to the current selection, and we have golang.org/x/tools/cmdcallgraph now. Internals: Drop support for long-running clients (i.e., Pythia), since godoc -analysis supports all the same features except "pointsto", and precomputes all the results so latency is much lower. Get rid of various unhelpful abstractions introduced to support long-running clients. Expand out the set-up logic for each subcommand. This is simpler, easier to read, and gives us more control, at a small cost in duplication---the familiar story of abstractions. Discard PTA warnings. We weren't showing them (nor should we). Split tests into separate directories (so that importgraph works). Change-Id: I55d46b3ab33cdf7ac22436fcc2148fe04c901237 Reviewed-on: https://go-review.googlesource.com/8243 Reviewed-by: David Crawshaw <crawshaw@golang.org>
2015-03-30 09:21:48 -06:00
// selected type within a single package.
// If the selection is a method, 'implements' displays
// the corresponding methods of the types that would have been reported
// by an implements query on the receiver type.
//
oracle: several major improvements Features: More robust: silently ignore type errors in modes that don't need SSA form: describe, referrers, implements, freevars, description. This makes the tool much more robust for everyday queries. Less configuration: don't require a scope argument for all queries. Only queries that do pointer analysis need it. For the rest, the initial position is enough for importQueryPackage to deduce the scope. It now works for queries in GoFiles, TestGoFiles, or XTestGoFiles. (It no longer works for ad-hoc main packages like $GOROOT/src/net/http/triv.go) More complete: "referrers" computes the scope automatically by scanning the import graph of the entire workspace, using gorename's refactor/importgraph package. This requires two passes at loading. Faster: simplified start-up logic avoids unnecessary package loading and SSA construction (a consequence of bad abstraction) in many cases. "callgraph": remove it. Unlike all the other commands it isn't related to the current selection, and we have golang.org/x/tools/cmdcallgraph now. Internals: Drop support for long-running clients (i.e., Pythia), since godoc -analysis supports all the same features except "pointsto", and precomputes all the results so latency is much lower. Get rid of various unhelpful abstractions introduced to support long-running clients. Expand out the set-up logic for each subcommand. This is simpler, easier to read, and gives us more control, at a small cost in duplication---the familiar story of abstractions. Discard PTA warnings. We weren't showing them (nor should we). Split tests into separate directories (so that importgraph works). Change-Id: I55d46b3ab33cdf7ac22436fcc2148fe04c901237 Reviewed-on: https://go-review.googlesource.com/8243 Reviewed-by: David Crawshaw <crawshaw@golang.org>
2015-03-30 09:21:48 -06:00
func implements(q *Query) error {
lconf := loader.Config{Build: q.Build}
allowErrors(&lconf)
if err := importQueryPackage(q.Pos, &lconf); err != nil {
return err
}
// Load/parse/type-check the program.
lprog, err := lconf.Load()
if err != nil {
return err
}
q.Fset = lprog.Fset
qpos, err := parseQueryPos(lprog, q.Pos, false)
if err != nil {
return err
}
// Find the selected type.
path, action := findInterestingNode(qpos.info, qpos.path)
var method *types.Func
var T types.Type // selected type (receiver if method != nil)
switch action {
case actionExpr:
// method?
if id, ok := path[0].(*ast.Ident); ok {
if obj, ok := qpos.info.ObjectOf(id).(*types.Func); ok {
recv := obj.Type().(*types.Signature).Recv()
if recv == nil {
oracle: several major improvements Features: More robust: silently ignore type errors in modes that don't need SSA form: describe, referrers, implements, freevars, description. This makes the tool much more robust for everyday queries. Less configuration: don't require a scope argument for all queries. Only queries that do pointer analysis need it. For the rest, the initial position is enough for importQueryPackage to deduce the scope. It now works for queries in GoFiles, TestGoFiles, or XTestGoFiles. (It no longer works for ad-hoc main packages like $GOROOT/src/net/http/triv.go) More complete: "referrers" computes the scope automatically by scanning the import graph of the entire workspace, using gorename's refactor/importgraph package. This requires two passes at loading. Faster: simplified start-up logic avoids unnecessary package loading and SSA construction (a consequence of bad abstraction) in many cases. "callgraph": remove it. Unlike all the other commands it isn't related to the current selection, and we have golang.org/x/tools/cmdcallgraph now. Internals: Drop support for long-running clients (i.e., Pythia), since godoc -analysis supports all the same features except "pointsto", and precomputes all the results so latency is much lower. Get rid of various unhelpful abstractions introduced to support long-running clients. Expand out the set-up logic for each subcommand. This is simpler, easier to read, and gives us more control, at a small cost in duplication---the familiar story of abstractions. Discard PTA warnings. We weren't showing them (nor should we). Split tests into separate directories (so that importgraph works). Change-Id: I55d46b3ab33cdf7ac22436fcc2148fe04c901237 Reviewed-on: https://go-review.googlesource.com/8243 Reviewed-by: David Crawshaw <crawshaw@golang.org>
2015-03-30 09:21:48 -06:00
return fmt.Errorf("this function is not a method")
}
method = obj
T = recv.Type()
}
}
case actionType:
T = qpos.info.TypeOf(path[0].(ast.Expr))
}
if T == nil {
oracle: several major improvements Features: More robust: silently ignore type errors in modes that don't need SSA form: describe, referrers, implements, freevars, description. This makes the tool much more robust for everyday queries. Less configuration: don't require a scope argument for all queries. Only queries that do pointer analysis need it. For the rest, the initial position is enough for importQueryPackage to deduce the scope. It now works for queries in GoFiles, TestGoFiles, or XTestGoFiles. (It no longer works for ad-hoc main packages like $GOROOT/src/net/http/triv.go) More complete: "referrers" computes the scope automatically by scanning the import graph of the entire workspace, using gorename's refactor/importgraph package. This requires two passes at loading. Faster: simplified start-up logic avoids unnecessary package loading and SSA construction (a consequence of bad abstraction) in many cases. "callgraph": remove it. Unlike all the other commands it isn't related to the current selection, and we have golang.org/x/tools/cmdcallgraph now. Internals: Drop support for long-running clients (i.e., Pythia), since godoc -analysis supports all the same features except "pointsto", and precomputes all the results so latency is much lower. Get rid of various unhelpful abstractions introduced to support long-running clients. Expand out the set-up logic for each subcommand. This is simpler, easier to read, and gives us more control, at a small cost in duplication---the familiar story of abstractions. Discard PTA warnings. We weren't showing them (nor should we). Split tests into separate directories (so that importgraph works). Change-Id: I55d46b3ab33cdf7ac22436fcc2148fe04c901237 Reviewed-on: https://go-review.googlesource.com/8243 Reviewed-by: David Crawshaw <crawshaw@golang.org>
2015-03-30 09:21:48 -06:00
return fmt.Errorf("no type or method here")
}
// Find all named types, even local types (which can have
// methods via promotion) and the built-in "error".
//
// TODO(adonovan): include all packages in PTA scope too?
// i.e. don't reduceScope?
//
var allNamed []types.Type
oracle: several major improvements Features: More robust: silently ignore type errors in modes that don't need SSA form: describe, referrers, implements, freevars, description. This makes the tool much more robust for everyday queries. Less configuration: don't require a scope argument for all queries. Only queries that do pointer analysis need it. For the rest, the initial position is enough for importQueryPackage to deduce the scope. It now works for queries in GoFiles, TestGoFiles, or XTestGoFiles. (It no longer works for ad-hoc main packages like $GOROOT/src/net/http/triv.go) More complete: "referrers" computes the scope automatically by scanning the import graph of the entire workspace, using gorename's refactor/importgraph package. This requires two passes at loading. Faster: simplified start-up logic avoids unnecessary package loading and SSA construction (a consequence of bad abstraction) in many cases. "callgraph": remove it. Unlike all the other commands it isn't related to the current selection, and we have golang.org/x/tools/cmdcallgraph now. Internals: Drop support for long-running clients (i.e., Pythia), since godoc -analysis supports all the same features except "pointsto", and precomputes all the results so latency is much lower. Get rid of various unhelpful abstractions introduced to support long-running clients. Expand out the set-up logic for each subcommand. This is simpler, easier to read, and gives us more control, at a small cost in duplication---the familiar story of abstractions. Discard PTA warnings. We weren't showing them (nor should we). Split tests into separate directories (so that importgraph works). Change-Id: I55d46b3ab33cdf7ac22436fcc2148fe04c901237 Reviewed-on: https://go-review.googlesource.com/8243 Reviewed-by: David Crawshaw <crawshaw@golang.org>
2015-03-30 09:21:48 -06:00
for _, info := range lprog.AllPackages {
for _, obj := range info.Defs {
if obj, ok := obj.(*types.TypeName); ok {
allNamed = append(allNamed, obj.Type())
}
}
}
allNamed = append(allNamed, types.Universe.Lookup("error").Type())
var msets types.MethodSetCache
// Test each named type.
var to, from, fromPtr []types.Type
for _, U := range allNamed {
if isInterface(T) {
if msets.MethodSet(T).Len() == 0 {
continue // empty interface
}
if isInterface(U) {
if msets.MethodSet(U).Len() == 0 {
continue // empty interface
}
// T interface, U interface
if !types.Identical(T, U) {
if types.AssignableTo(U, T) {
to = append(to, U)
}
if types.AssignableTo(T, U) {
from = append(from, U)
}
}
} else {
// T interface, U concrete
if types.AssignableTo(U, T) {
to = append(to, U)
} else if pU := types.NewPointer(U); types.AssignableTo(pU, T) {
to = append(to, pU)
}
}
} else if isInterface(U) {
if msets.MethodSet(U).Len() == 0 {
continue // empty interface
}
// T concrete, U interface
if types.AssignableTo(T, U) {
from = append(from, U)
} else if pT := types.NewPointer(T); types.AssignableTo(pT, U) {
fromPtr = append(fromPtr, U)
}
}
}
var pos interface{} = qpos
if nt, ok := deref(T).(*types.Named); ok {
pos = nt.Obj()
}
// Sort types (arbitrarily) to ensure test determinism.
sort.Sort(typesByString(to))
sort.Sort(typesByString(from))
sort.Sort(typesByString(fromPtr))
var toMethod, fromMethod, fromPtrMethod []*types.Selection // contain nils
if method != nil {
for _, t := range to {
toMethod = append(toMethod,
types.NewMethodSet(t).Lookup(method.Pkg(), method.Name()))
}
for _, t := range from {
fromMethod = append(fromMethod,
types.NewMethodSet(t).Lookup(method.Pkg(), method.Name()))
}
for _, t := range fromPtr {
fromPtrMethod = append(fromPtrMethod,
types.NewMethodSet(t).Lookup(method.Pkg(), method.Name()))
}
}
oracle: several major improvements Features: More robust: silently ignore type errors in modes that don't need SSA form: describe, referrers, implements, freevars, description. This makes the tool much more robust for everyday queries. Less configuration: don't require a scope argument for all queries. Only queries that do pointer analysis need it. For the rest, the initial position is enough for importQueryPackage to deduce the scope. It now works for queries in GoFiles, TestGoFiles, or XTestGoFiles. (It no longer works for ad-hoc main packages like $GOROOT/src/net/http/triv.go) More complete: "referrers" computes the scope automatically by scanning the import graph of the entire workspace, using gorename's refactor/importgraph package. This requires two passes at loading. Faster: simplified start-up logic avoids unnecessary package loading and SSA construction (a consequence of bad abstraction) in many cases. "callgraph": remove it. Unlike all the other commands it isn't related to the current selection, and we have golang.org/x/tools/cmdcallgraph now. Internals: Drop support for long-running clients (i.e., Pythia), since godoc -analysis supports all the same features except "pointsto", and precomputes all the results so latency is much lower. Get rid of various unhelpful abstractions introduced to support long-running clients. Expand out the set-up logic for each subcommand. This is simpler, easier to read, and gives us more control, at a small cost in duplication---the familiar story of abstractions. Discard PTA warnings. We weren't showing them (nor should we). Split tests into separate directories (so that importgraph works). Change-Id: I55d46b3ab33cdf7ac22436fcc2148fe04c901237 Reviewed-on: https://go-review.googlesource.com/8243 Reviewed-by: David Crawshaw <crawshaw@golang.org>
2015-03-30 09:21:48 -06:00
q.result = &implementsResult{
qpos, T, pos, to, from, fromPtr, method, toMethod, fromMethod, fromPtrMethod,
}
return nil
}
type implementsResult struct {
oracle: several major improvements Features: More robust: silently ignore type errors in modes that don't need SSA form: describe, referrers, implements, freevars, description. This makes the tool much more robust for everyday queries. Less configuration: don't require a scope argument for all queries. Only queries that do pointer analysis need it. For the rest, the initial position is enough for importQueryPackage to deduce the scope. It now works for queries in GoFiles, TestGoFiles, or XTestGoFiles. (It no longer works for ad-hoc main packages like $GOROOT/src/net/http/triv.go) More complete: "referrers" computes the scope automatically by scanning the import graph of the entire workspace, using gorename's refactor/importgraph package. This requires two passes at loading. Faster: simplified start-up logic avoids unnecessary package loading and SSA construction (a consequence of bad abstraction) in many cases. "callgraph": remove it. Unlike all the other commands it isn't related to the current selection, and we have golang.org/x/tools/cmdcallgraph now. Internals: Drop support for long-running clients (i.e., Pythia), since godoc -analysis supports all the same features except "pointsto", and precomputes all the results so latency is much lower. Get rid of various unhelpful abstractions introduced to support long-running clients. Expand out the set-up logic for each subcommand. This is simpler, easier to read, and gives us more control, at a small cost in duplication---the familiar story of abstractions. Discard PTA warnings. We weren't showing them (nor should we). Split tests into separate directories (so that importgraph works). Change-Id: I55d46b3ab33cdf7ac22436fcc2148fe04c901237 Reviewed-on: https://go-review.googlesource.com/8243 Reviewed-by: David Crawshaw <crawshaw@golang.org>
2015-03-30 09:21:48 -06:00
qpos *queryPos
t types.Type // queried type (not necessarily named)
pos interface{} // pos of t (*types.Name or *QueryPos)
to []types.Type // named or ptr-to-named types assignable to interface T
from []types.Type // named interfaces assignable from T
fromPtr []types.Type // named interfaces assignable only from *T
// if a method was queried:
method *types.Func // queried method
toMethod []*types.Selection // method of type to[i], if any
fromMethod []*types.Selection // method of type from[i], if any
fromPtrMethod []*types.Selection // method of type fromPtrMethod[i], if any
}
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
func (r *implementsResult) display(printf printfFunc) {
relation := "is implemented by"
meth := func(sel *types.Selection) {
if sel != nil {
printf(sel.Obj(), "\t%s method (%s).%s",
oracle: several major improvements Features: More robust: silently ignore type errors in modes that don't need SSA form: describe, referrers, implements, freevars, description. This makes the tool much more robust for everyday queries. Less configuration: don't require a scope argument for all queries. Only queries that do pointer analysis need it. For the rest, the initial position is enough for importQueryPackage to deduce the scope. It now works for queries in GoFiles, TestGoFiles, or XTestGoFiles. (It no longer works for ad-hoc main packages like $GOROOT/src/net/http/triv.go) More complete: "referrers" computes the scope automatically by scanning the import graph of the entire workspace, using gorename's refactor/importgraph package. This requires two passes at loading. Faster: simplified start-up logic avoids unnecessary package loading and SSA construction (a consequence of bad abstraction) in many cases. "callgraph": remove it. Unlike all the other commands it isn't related to the current selection, and we have golang.org/x/tools/cmdcallgraph now. Internals: Drop support for long-running clients (i.e., Pythia), since godoc -analysis supports all the same features except "pointsto", and precomputes all the results so latency is much lower. Get rid of various unhelpful abstractions introduced to support long-running clients. Expand out the set-up logic for each subcommand. This is simpler, easier to read, and gives us more control, at a small cost in duplication---the familiar story of abstractions. Discard PTA warnings. We weren't showing them (nor should we). Split tests into separate directories (so that importgraph works). Change-Id: I55d46b3ab33cdf7ac22436fcc2148fe04c901237 Reviewed-on: https://go-review.googlesource.com/8243 Reviewed-by: David Crawshaw <crawshaw@golang.org>
2015-03-30 09:21:48 -06:00
relation, r.qpos.typeString(sel.Recv()), sel.Obj().Name())
}
}
if isInterface(r.t) {
if types.NewMethodSet(r.t).Len() == 0 { // TODO(adonovan): cache mset
printf(r.pos, "empty interface type %s", r.qpos.typeString(r.t))
return
}
if r.method == nil {
printf(r.pos, "interface type %s", r.qpos.typeString(r.t))
} else {
oracle: several major improvements Features: More robust: silently ignore type errors in modes that don't need SSA form: describe, referrers, implements, freevars, description. This makes the tool much more robust for everyday queries. Less configuration: don't require a scope argument for all queries. Only queries that do pointer analysis need it. For the rest, the initial position is enough for importQueryPackage to deduce the scope. It now works for queries in GoFiles, TestGoFiles, or XTestGoFiles. (It no longer works for ad-hoc main packages like $GOROOT/src/net/http/triv.go) More complete: "referrers" computes the scope automatically by scanning the import graph of the entire workspace, using gorename's refactor/importgraph package. This requires two passes at loading. Faster: simplified start-up logic avoids unnecessary package loading and SSA construction (a consequence of bad abstraction) in many cases. "callgraph": remove it. Unlike all the other commands it isn't related to the current selection, and we have golang.org/x/tools/cmdcallgraph now. Internals: Drop support for long-running clients (i.e., Pythia), since godoc -analysis supports all the same features except "pointsto", and precomputes all the results so latency is much lower. Get rid of various unhelpful abstractions introduced to support long-running clients. Expand out the set-up logic for each subcommand. This is simpler, easier to read, and gives us more control, at a small cost in duplication---the familiar story of abstractions. Discard PTA warnings. We weren't showing them (nor should we). Split tests into separate directories (so that importgraph works). Change-Id: I55d46b3ab33cdf7ac22436fcc2148fe04c901237 Reviewed-on: https://go-review.googlesource.com/8243 Reviewed-by: David Crawshaw <crawshaw@golang.org>
2015-03-30 09:21:48 -06:00
printf(r.method, "abstract method %s", r.qpos.objectString(r.method))
}
// Show concrete types (or methods) first; use two passes.
for i, sub := range r.to {
if !isInterface(sub) {
if r.method == nil {
printf(deref(sub).(*types.Named).Obj(), "\t%s %s type %s",
relation, typeKind(sub), r.qpos.typeString(sub))
} else {
meth(r.toMethod[i])
}
}
}
for i, sub := range r.to {
if isInterface(sub) {
if r.method == nil {
printf(sub.(*types.Named).Obj(), "\t%s %s type %s",
relation, typeKind(sub), r.qpos.typeString(sub))
} else {
meth(r.toMethod[i])
}
}
}
relation = "implements"
for i, super := range r.from {
if r.method == nil {
printf(super.(*types.Named).Obj(), "\t%s %s",
relation, r.qpos.typeString(super))
} else {
meth(r.fromMethod[i])
}
}
} else {
relation = "implements"
if r.from != nil {
if r.method == nil {
printf(r.pos, "%s type %s",
typeKind(r.t), r.qpos.typeString(r.t))
} else {
printf(r.method, "concrete method %s",
oracle: several major improvements Features: More robust: silently ignore type errors in modes that don't need SSA form: describe, referrers, implements, freevars, description. This makes the tool much more robust for everyday queries. Less configuration: don't require a scope argument for all queries. Only queries that do pointer analysis need it. For the rest, the initial position is enough for importQueryPackage to deduce the scope. It now works for queries in GoFiles, TestGoFiles, or XTestGoFiles. (It no longer works for ad-hoc main packages like $GOROOT/src/net/http/triv.go) More complete: "referrers" computes the scope automatically by scanning the import graph of the entire workspace, using gorename's refactor/importgraph package. This requires two passes at loading. Faster: simplified start-up logic avoids unnecessary package loading and SSA construction (a consequence of bad abstraction) in many cases. "callgraph": remove it. Unlike all the other commands it isn't related to the current selection, and we have golang.org/x/tools/cmdcallgraph now. Internals: Drop support for long-running clients (i.e., Pythia), since godoc -analysis supports all the same features except "pointsto", and precomputes all the results so latency is much lower. Get rid of various unhelpful abstractions introduced to support long-running clients. Expand out the set-up logic for each subcommand. This is simpler, easier to read, and gives us more control, at a small cost in duplication---the familiar story of abstractions. Discard PTA warnings. We weren't showing them (nor should we). Split tests into separate directories (so that importgraph works). Change-Id: I55d46b3ab33cdf7ac22436fcc2148fe04c901237 Reviewed-on: https://go-review.googlesource.com/8243 Reviewed-by: David Crawshaw <crawshaw@golang.org>
2015-03-30 09:21:48 -06:00
r.qpos.objectString(r.method))
}
for i, super := range r.from {
if r.method == nil {
printf(super.(*types.Named).Obj(), "\t%s %s",
relation, r.qpos.typeString(super))
} else {
meth(r.fromMethod[i])
}
}
}
if r.fromPtr != nil {
if r.method == nil {
printf(r.pos, "pointer type *%s", r.qpos.typeString(r.t))
} else {
// TODO(adonovan): de-dup (C).f and (*C).f implementing (I).f.
printf(r.method, "concrete method %s",
oracle: several major improvements Features: More robust: silently ignore type errors in modes that don't need SSA form: describe, referrers, implements, freevars, description. This makes the tool much more robust for everyday queries. Less configuration: don't require a scope argument for all queries. Only queries that do pointer analysis need it. For the rest, the initial position is enough for importQueryPackage to deduce the scope. It now works for queries in GoFiles, TestGoFiles, or XTestGoFiles. (It no longer works for ad-hoc main packages like $GOROOT/src/net/http/triv.go) More complete: "referrers" computes the scope automatically by scanning the import graph of the entire workspace, using gorename's refactor/importgraph package. This requires two passes at loading. Faster: simplified start-up logic avoids unnecessary package loading and SSA construction (a consequence of bad abstraction) in many cases. "callgraph": remove it. Unlike all the other commands it isn't related to the current selection, and we have golang.org/x/tools/cmdcallgraph now. Internals: Drop support for long-running clients (i.e., Pythia), since godoc -analysis supports all the same features except "pointsto", and precomputes all the results so latency is much lower. Get rid of various unhelpful abstractions introduced to support long-running clients. Expand out the set-up logic for each subcommand. This is simpler, easier to read, and gives us more control, at a small cost in duplication---the familiar story of abstractions. Discard PTA warnings. We weren't showing them (nor should we). Split tests into separate directories (so that importgraph works). Change-Id: I55d46b3ab33cdf7ac22436fcc2148fe04c901237 Reviewed-on: https://go-review.googlesource.com/8243 Reviewed-by: David Crawshaw <crawshaw@golang.org>
2015-03-30 09:21:48 -06:00
r.qpos.objectString(r.method))
}
for i, psuper := range r.fromPtr {
if r.method == nil {
printf(psuper.(*types.Named).Obj(), "\t%s %s",
relation, r.qpos.typeString(psuper))
} else {
meth(r.fromPtrMethod[i])
}
}
} else if r.from == nil {
printf(r.pos, "%s type %s implements only interface{}",
typeKind(r.t), r.qpos.typeString(r.t))
}
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
}
}
func (r *implementsResult) toSerial(res *serial.Result, fset *token.FileSet) {
res.Implements = &serial.Implements{
T: makeImplementsType(r.t, fset),
AssignableTo: makeImplementsTypes(r.to, fset),
AssignableFrom: makeImplementsTypes(r.from, fset),
AssignableFromPtr: makeImplementsTypes(r.fromPtr, fset),
AssignableToMethod: methodsToSerial(r.qpos.info.Pkg, r.toMethod, fset),
AssignableFromMethod: methodsToSerial(r.qpos.info.Pkg, r.fromMethod, fset),
AssignableFromPtrMethod: methodsToSerial(r.qpos.info.Pkg, r.fromPtrMethod, fset),
}
if r.method != nil {
res.Implements.Method = &serial.DescribeMethod{
oracle: several major improvements Features: More robust: silently ignore type errors in modes that don't need SSA form: describe, referrers, implements, freevars, description. This makes the tool much more robust for everyday queries. Less configuration: don't require a scope argument for all queries. Only queries that do pointer analysis need it. For the rest, the initial position is enough for importQueryPackage to deduce the scope. It now works for queries in GoFiles, TestGoFiles, or XTestGoFiles. (It no longer works for ad-hoc main packages like $GOROOT/src/net/http/triv.go) More complete: "referrers" computes the scope automatically by scanning the import graph of the entire workspace, using gorename's refactor/importgraph package. This requires two passes at loading. Faster: simplified start-up logic avoids unnecessary package loading and SSA construction (a consequence of bad abstraction) in many cases. "callgraph": remove it. Unlike all the other commands it isn't related to the current selection, and we have golang.org/x/tools/cmdcallgraph now. Internals: Drop support for long-running clients (i.e., Pythia), since godoc -analysis supports all the same features except "pointsto", and precomputes all the results so latency is much lower. Get rid of various unhelpful abstractions introduced to support long-running clients. Expand out the set-up logic for each subcommand. This is simpler, easier to read, and gives us more control, at a small cost in duplication---the familiar story of abstractions. Discard PTA warnings. We weren't showing them (nor should we). Split tests into separate directories (so that importgraph works). Change-Id: I55d46b3ab33cdf7ac22436fcc2148fe04c901237 Reviewed-on: https://go-review.googlesource.com/8243 Reviewed-by: David Crawshaw <crawshaw@golang.org>
2015-03-30 09:21:48 -06:00
Name: r.qpos.objectString(r.method),
Pos: fset.Position(r.method.Pos()).String(),
}
}
}
func makeImplementsTypes(tt []types.Type, fset *token.FileSet) []serial.ImplementsType {
var r []serial.ImplementsType
for _, t := range tt {
r = append(r, makeImplementsType(t, fset))
}
return r
}
func makeImplementsType(T types.Type, fset *token.FileSet) serial.ImplementsType {
var pos token.Pos
if nt, ok := deref(T).(*types.Named); ok { // implementsResult.t may be non-named
pos = nt.Obj().Pos()
}
return serial.ImplementsType{
Name: T.String(),
Pos: fset.Position(pos).String(),
Kind: typeKind(T),
}
}
// typeKind returns a string describing the underlying kind of type,
// e.g. "slice", "array", "struct".
func typeKind(T types.Type) string {
s := reflect.TypeOf(T.Underlying()).String()
return strings.ToLower(strings.TrimPrefix(s, "*types."))
}
func isInterface(T types.Type) bool { return types.IsInterface(T) }
type typesByString []types.Type
func (p typesByString) Len() int { return len(p) }
func (p typesByString) Less(i, j int) bool { return p[i].String() < p[j].String() }
func (p typesByString) Swap(i, j int) { p[i], p[j] = p[j], p[i] }