1
0
mirror of https://github.com/golang/go synced 2024-11-19 05:54:44 -07:00
go/oracle/callees.go

254 lines
6.4 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"
"sort"
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/pointer"
"golang.org/x/tools/go/ssa"
"golang.org/x/tools/go/ssa/ssautil"
"golang.org/x/tools/go/types"
"golang.org/x/tools/oracle/serial"
)
// Callees reports the possible callees of the function call site
// identified by the specified source location.
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 callees(q *Query) error {
lconf := loader.Config{Build: q.Build}
if err := setPTAScope(&lconf, q.Scope); err != 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 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, true) // needs exact pos
if err != nil {
return err
}
// Determine the enclosing call for the specified position.
var e *ast.CallExpr
for _, n := range qpos.path {
if e, _ = n.(*ast.CallExpr); e != nil {
break
}
}
if e == 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("there is no function call here")
}
// TODO(adonovan): issue an error if the call is "too far
// away" from the current selection, as this most likely is
// not what the user intended.
// Reject type conversions.
if qpos.info.Types[e.Fun].IsType() {
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 is a type conversion, not a function call")
}
// Deal with obviously static calls before constructing SSA form.
// Some static calls may yet require SSA construction,
// e.g. f := func(){}; f().
switch funexpr := unparen(e.Fun).(type) {
case *ast.Ident:
switch obj := qpos.info.Uses[funexpr].(type) {
case *types.Builtin:
// Reject calls to built-ins.
return fmt.Errorf("this is a call to the built-in '%s' operator", obj.Name())
case *types.Func:
// This is a static function call
q.result = &calleesTypesResult{
site: e,
callee: obj,
}
return nil
}
case *ast.SelectorExpr:
sel := qpos.info.Selections[funexpr]
if sel == nil {
// qualified identifier.
// May refer to top level function variable
// or to top level function.
callee := qpos.info.Uses[funexpr.Sel]
if obj, ok := callee.(*types.Func); ok {
q.result = &calleesTypesResult{
site: e,
callee: obj,
}
return nil
}
} else if sel.Kind() == types.MethodVal {
recvtype := sel.Recv()
if !types.IsInterface(recvtype) {
// static method call
q.result = &calleesTypesResult{
site: e,
callee: sel.Obj().(*types.Func),
}
return nil
}
}
}
prog := ssautil.CreateProgram(lprog, ssa.GlobalDebug)
ptaConfig, err := setupPTA(prog, lprog, q.PTALog, q.Reflection)
if err != nil {
return err
}
pkg := prog.Package(qpos.info.Pkg)
if pkg == nil {
return fmt.Errorf("no SSA package")
}
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
// Defer SSA construction till after errors are reported.
prog.Build()
// Ascertain calling function and call site.
callerFn := ssa.EnclosingFunction(pkg, qpos.path)
if callerFn == 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 SSA function built for this location (dead code?)")
}
// Find the call site.
site, err := findCallSite(callerFn, e)
if err != 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 err
}
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
funcs, err := findCallees(ptaConfig, site)
if err != 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 err
}
q.result = &calleesSSAResult{
site: site,
funcs: funcs,
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 nil
}
func findCallSite(fn *ssa.Function, call *ast.CallExpr) (ssa.CallInstruction, error) {
instr, _ := fn.ValueForExpr(call)
callInstr, _ := instr.(ssa.CallInstruction)
if instr == nil {
return nil, fmt.Errorf("this call site is unreachable in this analysis")
}
return callInstr, 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
func findCallees(conf *pointer.Config, site ssa.CallInstruction) ([]*ssa.Function, error) {
// Avoid running the pointer analysis for static calls.
if callee := site.Common().StaticCallee(); callee != nil {
switch callee.String() {
case "runtime.SetFinalizer", "(reflect.Value).Call":
// The PTA treats calls to these intrinsics as dynamic.
// TODO(adonovan): avoid reliance on PTA internals.
default:
return []*ssa.Function{callee}, nil // singleton
}
}
// Dynamic call: use pointer analysis.
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
conf.BuildCallGraph = true
cg := ptrAnalysis(conf).CallGraph
cg.DeleteSyntheticNodes()
// Find all call edges from the site.
n := cg.Nodes[site.Parent()]
if n == nil {
return nil, fmt.Errorf("this call site is unreachable in this analysis")
}
calleesMap := make(map[*ssa.Function]bool)
for _, edge := range n.Out {
if edge.Site == site {
calleesMap[edge.Callee.Func] = true
}
}
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
// De-duplicate and sort.
funcs := make([]*ssa.Function, 0, len(calleesMap))
for f := range calleesMap {
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
funcs = append(funcs, f)
}
sort.Sort(byFuncPos(funcs))
return funcs, nil
}
type calleesSSAResult struct {
site ssa.CallInstruction
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
funcs []*ssa.Function
}
type calleesTypesResult struct {
site *ast.CallExpr
callee *types.Func
}
func (r *calleesSSAResult) display(printf printfFunc) {
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
if len(r.funcs) == 0 {
// dynamic call on a provably nil func/interface
printf(r.site, "%s on nil value", r.site.Common().Description())
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
} else {
printf(r.site, "this %s dispatches to:", r.site.Common().Description())
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
for _, callee := range r.funcs {
printf(callee, "\t%s", callee)
}
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 *calleesSSAResult) toSerial(res *serial.Result, fset *token.FileSet) {
j := &serial.Callees{
Pos: fset.Position(r.site.Pos()).String(),
Desc: r.site.Common().Description(),
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
}
for _, callee := range r.funcs {
j.Callees = append(j.Callees, &serial.CalleesItem{
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
Name: callee.String(),
Pos: fset.Position(callee.Pos()).String(),
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.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
res.Callees = j
}
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 *calleesTypesResult) display(printf printfFunc) {
printf(r.site, "this static function call dispatches to:")
printf(r.callee, "\t%s", r.callee.FullName())
}
func (r *calleesTypesResult) toSerial(res *serial.Result, fset *token.FileSet) {
j := &serial.Callees{
Pos: fset.Position(r.site.Pos()).String(),
Desc: "static function call",
}
j.Callees = []*serial.CalleesItem{
&serial.CalleesItem{
Name: r.callee.FullName(),
Pos: fset.Position(r.callee.Pos()).String(),
},
}
res.Callees = j
}
// NB: byFuncPos is not deterministic across packages since it depends on load order.
// Use lessPos if the tests need it.
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 byFuncPos []*ssa.Function
func (a byFuncPos) Len() int { return len(a) }
func (a byFuncPos) Less(i, j int) bool { return a[i].Pos() < a[j].Pos() }
func (a byFuncPos) Swap(i, j int) { a[i], a[j] = a[j], a[i] }