1
0
mirror of https://github.com/golang/go synced 2024-10-01 10:28:31 -06:00
go/oracle/referrers.go

222 lines
5.5 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 (
"bytes"
"fmt"
"go/ast"
"go/token"
"io/ioutil"
"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/types"
"golang.org/x/tools/oracle/serial"
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/refactor/importgraph"
)
// TODO(adonovan): use golang.org/x/tools/refactor/importgraph to choose
// the scope automatically.
// Referrers reports all identifiers that resolve to the same object
// as the queried identifier, within any package in the analysis scope.
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 referrers(q *Query) error {
lconf := loader.Config{Build: q.Build}
allowErrors(&lconf)
if err := importQueryPackage(q.Pos, &lconf); err != nil {
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
var id *ast.Ident
var obj types.Object
var lprog *loader.Program
var pass2 bool
for {
// Load/parse/type-check the program.
var err error
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
}
id, _ = qpos.path[0].(*ast.Ident)
if id == nil {
return fmt.Errorf("no identifier here")
}
obj = qpos.info.ObjectOf(id)
if obj == nil {
// Happens for y in "switch y := x.(type)", but I think that's all.
return fmt.Errorf("no object for identifier")
}
// If the identifier is exported, we must load all packages that
// depend transitively upon the package that defines it.
//
// TODO(adonovan): we should do this for PkgName objects
// too, even though they're lowercase.
//
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
// TODO(adonovan): opt: skip this step if obj.Pkg() is a test or
// main package.
if pass2 || !obj.Exported() {
break
}
// Scan the workspace and build the import graph.
// Ignore broken packages.
_, rev, _ := importgraph.Build(q.Build)
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
// Re-load the larger program.
// Create a new file set so that ...
// External test packages are never imported,
// so they will never appear in the graph.
// (We must reset the Config here, not just reset the Fset field.)
lconf = loader.Config{
Fset: token.NewFileSet(),
Build: q.Build,
}
allowErrors(&lconf)
for path := range rev.Search(obj.Pkg().Path()) {
lconf.ImportWithTests(path)
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
}
pass2 = true
}
// Iterate over all go/types' Uses facts for the entire program.
go.tools/oracle: improvements to command set and performance. Command set: - what: an extremely fast query that parses a single file and returns the AST stack, package name and the set of query modes that apply to the current selection. Intended for GUI tools that need to grey out UI elements. - definition: shows the definition of an identifier. - pointsto: the PTA features of 'describe' have been split out into their own command. - describe: with PTA stripped out, the cost is now bounded by type checking. Performance: - The importer.Config.TypeCheckFuncBodies predicate supports setting the 'IgnoreFuncBodies' typechecker flag on a per-package basis. This means we can load dependencies from source more quickly if we only need exported types. (We avoid gcimport data because it may be absent or stale.) This also means we can run type-based queries on packages that aren't part of the pointer analysis scope. (Yay.) - Modes that require only type analysis of the query package run a "what" query first, and restrict their analysis scope to just that package and its dependencies (sans func bodies), making them much faster. - We call newOracle not oracle.New in Query, so that the 'needs' bitset isn't ignored (oops!). This makes the non-PTA queries faster. Also: - removed vestigial timers junk. - pos.go: existing position utilties split out into own file. Added parsePosFlag utility. - numerous cosmetic tweaks. + very basic tests. To do in follow-ups: - sophisticated editor integration of "what". - better tests. - refactoring of control flow as described in comment. - changes to "implements", "describe" commands. - update design doc + user manual. R=crawshaw, dominik.honnef CC=golang-dev, gri https://golang.org/cl/40630043
2013-12-13 08:04:55 -07:00
var refs []*ast.Ident
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 id2, obj2 := range info.Uses {
if sameObj(obj, obj2) {
go.tools/oracle: improvements to command set and performance. Command set: - what: an extremely fast query that parses a single file and returns the AST stack, package name and the set of query modes that apply to the current selection. Intended for GUI tools that need to grey out UI elements. - definition: shows the definition of an identifier. - pointsto: the PTA features of 'describe' have been split out into their own command. - describe: with PTA stripped out, the cost is now bounded by type checking. Performance: - The importer.Config.TypeCheckFuncBodies predicate supports setting the 'IgnoreFuncBodies' typechecker flag on a per-package basis. This means we can load dependencies from source more quickly if we only need exported types. (We avoid gcimport data because it may be absent or stale.) This also means we can run type-based queries on packages that aren't part of the pointer analysis scope. (Yay.) - Modes that require only type analysis of the query package run a "what" query first, and restrict their analysis scope to just that package and its dependencies (sans func bodies), making them much faster. - We call newOracle not oracle.New in Query, so that the 'needs' bitset isn't ignored (oops!). This makes the non-PTA queries faster. Also: - removed vestigial timers junk. - pos.go: existing position utilties split out into own file. Added parsePosFlag utility. - numerous cosmetic tweaks. + very basic tests. To do in follow-ups: - sophisticated editor integration of "what". - better tests. - refactoring of control flow as described in comment. - changes to "implements", "describe" commands. - update design doc + user manual. R=crawshaw, dominik.honnef CC=golang-dev, gri https://golang.org/cl/40630043
2013-12-13 08:04:55 -07:00
refs = append(refs, id2)
}
}
}
sort.Sort(byNamePos{q.Fset, refs})
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 = &referrersResult{
fset: q.Fset,
go.tools/oracle: improvements to command set and performance. Command set: - what: an extremely fast query that parses a single file and returns the AST stack, package name and the set of query modes that apply to the current selection. Intended for GUI tools that need to grey out UI elements. - definition: shows the definition of an identifier. - pointsto: the PTA features of 'describe' have been split out into their own command. - describe: with PTA stripped out, the cost is now bounded by type checking. Performance: - The importer.Config.TypeCheckFuncBodies predicate supports setting the 'IgnoreFuncBodies' typechecker flag on a per-package basis. This means we can load dependencies from source more quickly if we only need exported types. (We avoid gcimport data because it may be absent or stale.) This also means we can run type-based queries on packages that aren't part of the pointer analysis scope. (Yay.) - Modes that require only type analysis of the query package run a "what" query first, and restrict their analysis scope to just that package and its dependencies (sans func bodies), making them much faster. - We call newOracle not oracle.New in Query, so that the 'needs' bitset isn't ignored (oops!). This makes the non-PTA queries faster. Also: - removed vestigial timers junk. - pos.go: existing position utilties split out into own file. Added parsePosFlag utility. - numerous cosmetic tweaks. + very basic tests. To do in follow-ups: - sophisticated editor integration of "what". - better tests. - refactoring of control flow as described in comment. - changes to "implements", "describe" commands. - update design doc + user manual. R=crawshaw, dominik.honnef CC=golang-dev, gri https://golang.org/cl/40630043
2013-12-13 08:04:55 -07:00
query: id,
obj: obj,
refs: refs,
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
}
// same reports whether x and y are identical, or both are PkgNames
// that import the same Package.
//
func sameObj(x, y types.Object) bool {
if x == y {
return true
}
if x, ok := x.(*types.PkgName); ok {
if y, ok := y.(*types.PkgName); ok {
return x.Imported() == y.Imported()
}
}
return false
}
go.tools/oracle: improvements to command set and performance. Command set: - what: an extremely fast query that parses a single file and returns the AST stack, package name and the set of query modes that apply to the current selection. Intended for GUI tools that need to grey out UI elements. - definition: shows the definition of an identifier. - pointsto: the PTA features of 'describe' have been split out into their own command. - describe: with PTA stripped out, the cost is now bounded by type checking. Performance: - The importer.Config.TypeCheckFuncBodies predicate supports setting the 'IgnoreFuncBodies' typechecker flag on a per-package basis. This means we can load dependencies from source more quickly if we only need exported types. (We avoid gcimport data because it may be absent or stale.) This also means we can run type-based queries on packages that aren't part of the pointer analysis scope. (Yay.) - Modes that require only type analysis of the query package run a "what" query first, and restrict their analysis scope to just that package and its dependencies (sans func bodies), making them much faster. - We call newOracle not oracle.New in Query, so that the 'needs' bitset isn't ignored (oops!). This makes the non-PTA queries faster. Also: - removed vestigial timers junk. - pos.go: existing position utilties split out into own file. Added parsePosFlag utility. - numerous cosmetic tweaks. + very basic tests. To do in follow-ups: - sophisticated editor integration of "what". - better tests. - refactoring of control flow as described in comment. - changes to "implements", "describe" commands. - update design doc + user manual. R=crawshaw, dominik.honnef CC=golang-dev, gri https://golang.org/cl/40630043
2013-12-13 08:04:55 -07:00
// -------- utils --------
// An deterministic ordering for token.Pos that doesn't
// depend on the order in which packages were loaded.
func lessPos(fset *token.FileSet, x, y token.Pos) bool {
fx := fset.File(x)
fy := fset.File(y)
if fx != fy {
return fx.Name() < fy.Name()
}
return x < y
}
go.tools/oracle: improvements to command set and performance. Command set: - what: an extremely fast query that parses a single file and returns the AST stack, package name and the set of query modes that apply to the current selection. Intended for GUI tools that need to grey out UI elements. - definition: shows the definition of an identifier. - pointsto: the PTA features of 'describe' have been split out into their own command. - describe: with PTA stripped out, the cost is now bounded by type checking. Performance: - The importer.Config.TypeCheckFuncBodies predicate supports setting the 'IgnoreFuncBodies' typechecker flag on a per-package basis. This means we can load dependencies from source more quickly if we only need exported types. (We avoid gcimport data because it may be absent or stale.) This also means we can run type-based queries on packages that aren't part of the pointer analysis scope. (Yay.) - Modes that require only type analysis of the query package run a "what" query first, and restrict their analysis scope to just that package and its dependencies (sans func bodies), making them much faster. - We call newOracle not oracle.New in Query, so that the 'needs' bitset isn't ignored (oops!). This makes the non-PTA queries faster. Also: - removed vestigial timers junk. - pos.go: existing position utilties split out into own file. Added parsePosFlag utility. - numerous cosmetic tweaks. + very basic tests. To do in follow-ups: - sophisticated editor integration of "what". - better tests. - refactoring of control flow as described in comment. - changes to "implements", "describe" commands. - update design doc + user manual. R=crawshaw, dominik.honnef CC=golang-dev, gri https://golang.org/cl/40630043
2013-12-13 08:04:55 -07:00
type byNamePos struct {
fset *token.FileSet
ids []*ast.Ident
}
func (p byNamePos) Len() int { return len(p.ids) }
func (p byNamePos) Swap(i, j int) { p.ids[i], p.ids[j] = p.ids[j], p.ids[i] }
func (p byNamePos) Less(i, j int) bool {
return lessPos(p.fset, p.ids[i].NamePos, p.ids[j].NamePos)
}
go.tools/oracle: improvements to command set and performance. Command set: - what: an extremely fast query that parses a single file and returns the AST stack, package name and the set of query modes that apply to the current selection. Intended for GUI tools that need to grey out UI elements. - definition: shows the definition of an identifier. - pointsto: the PTA features of 'describe' have been split out into their own command. - describe: with PTA stripped out, the cost is now bounded by type checking. Performance: - The importer.Config.TypeCheckFuncBodies predicate supports setting the 'IgnoreFuncBodies' typechecker flag on a per-package basis. This means we can load dependencies from source more quickly if we only need exported types. (We avoid gcimport data because it may be absent or stale.) This also means we can run type-based queries on packages that aren't part of the pointer analysis scope. (Yay.) - Modes that require only type analysis of the query package run a "what" query first, and restrict their analysis scope to just that package and its dependencies (sans func bodies), making them much faster. - We call newOracle not oracle.New in Query, so that the 'needs' bitset isn't ignored (oops!). This makes the non-PTA queries faster. Also: - removed vestigial timers junk. - pos.go: existing position utilties split out into own file. Added parsePosFlag utility. - numerous cosmetic tweaks. + very basic tests. To do in follow-ups: - sophisticated editor integration of "what". - better tests. - refactoring of control flow as described in comment. - changes to "implements", "describe" commands. - update design doc + user manual. R=crawshaw, dominik.honnef CC=golang-dev, gri https://golang.org/cl/40630043
2013-12-13 08:04:55 -07:00
type referrersResult 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
fset *token.FileSet
query *ast.Ident // identifier of query
obj types.Object // object it denotes
go.tools/oracle: improvements to command set and performance. Command set: - what: an extremely fast query that parses a single file and returns the AST stack, package name and the set of query modes that apply to the current selection. Intended for GUI tools that need to grey out UI elements. - definition: shows the definition of an identifier. - pointsto: the PTA features of 'describe' have been split out into their own command. - describe: with PTA stripped out, the cost is now bounded by type checking. Performance: - The importer.Config.TypeCheckFuncBodies predicate supports setting the 'IgnoreFuncBodies' typechecker flag on a per-package basis. This means we can load dependencies from source more quickly if we only need exported types. (We avoid gcimport data because it may be absent or stale.) This also means we can run type-based queries on packages that aren't part of the pointer analysis scope. (Yay.) - Modes that require only type analysis of the query package run a "what" query first, and restrict their analysis scope to just that package and its dependencies (sans func bodies), making them much faster. - We call newOracle not oracle.New in Query, so that the 'needs' bitset isn't ignored (oops!). This makes the non-PTA queries faster. Also: - removed vestigial timers junk. - pos.go: existing position utilties split out into own file. Added parsePosFlag utility. - numerous cosmetic tweaks. + very basic tests. To do in follow-ups: - sophisticated editor integration of "what". - better tests. - refactoring of control flow as described in comment. - changes to "implements", "describe" commands. - update design doc + user manual. R=crawshaw, dominik.honnef CC=golang-dev, gri https://golang.org/cl/40630043
2013-12-13 08:04:55 -07:00
refs []*ast.Ident // set of all other references to it
}
func (r *referrersResult) display(printf printfFunc) {
printf(r.obj, "%d references to %s", len(r.refs), r.obj)
// Show referring lines, like grep.
type fileinfo struct {
refs []*ast.Ident
linenums []int // line number of refs[i]
data chan []byte // file contents
}
var fileinfos []*fileinfo
fileinfosByName := make(map[string]*fileinfo)
// First pass: start the file reads concurrently.
for _, ref := range r.refs {
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
posn := r.fset.Position(ref.Pos())
fi := fileinfosByName[posn.Filename]
if fi == nil {
fi = &fileinfo{data: make(chan []byte)}
fileinfosByName[posn.Filename] = fi
fileinfos = append(fileinfos, fi)
// First request for this file:
// start asynchronous read.
go func() {
content, err := ioutil.ReadFile(posn.Filename)
if err != nil {
content = []byte(fmt.Sprintf("error: %v", err))
}
fi.data <- content
}()
}
fi.refs = append(fi.refs, ref)
fi.linenums = append(fi.linenums, posn.Line)
}
// Second pass: print refs in original order.
// One line may have several refs at different columns.
for _, fi := range fileinfos {
content := <-fi.data // wait for I/O completion
lines := bytes.Split(content, []byte("\n"))
for i, ref := range fi.refs {
printf(ref, "%s", lines[fi.linenums[i]-1])
}
}
}
go.tools/oracle: improvements to command set and performance. Command set: - what: an extremely fast query that parses a single file and returns the AST stack, package name and the set of query modes that apply to the current selection. Intended for GUI tools that need to grey out UI elements. - definition: shows the definition of an identifier. - pointsto: the PTA features of 'describe' have been split out into their own command. - describe: with PTA stripped out, the cost is now bounded by type checking. Performance: - The importer.Config.TypeCheckFuncBodies predicate supports setting the 'IgnoreFuncBodies' typechecker flag on a per-package basis. This means we can load dependencies from source more quickly if we only need exported types. (We avoid gcimport data because it may be absent or stale.) This also means we can run type-based queries on packages that aren't part of the pointer analysis scope. (Yay.) - Modes that require only type analysis of the query package run a "what" query first, and restrict their analysis scope to just that package and its dependencies (sans func bodies), making them much faster. - We call newOracle not oracle.New in Query, so that the 'needs' bitset isn't ignored (oops!). This makes the non-PTA queries faster. Also: - removed vestigial timers junk. - pos.go: existing position utilties split out into own file. Added parsePosFlag utility. - numerous cosmetic tweaks. + very basic tests. To do in follow-ups: - sophisticated editor integration of "what". - better tests. - refactoring of control flow as described in comment. - changes to "implements", "describe" commands. - update design doc + user manual. R=crawshaw, dominik.honnef CC=golang-dev, gri https://golang.org/cl/40630043
2013-12-13 08:04:55 -07:00
// TODO(adonovan): encode extent, not just Pos info, in Serial form.
func (r *referrersResult) toSerial(res *serial.Result, fset *token.FileSet) {
referrers := &serial.Referrers{
go.tools/oracle: improvements to command set and performance. Command set: - what: an extremely fast query that parses a single file and returns the AST stack, package name and the set of query modes that apply to the current selection. Intended for GUI tools that need to grey out UI elements. - definition: shows the definition of an identifier. - pointsto: the PTA features of 'describe' have been split out into their own command. - describe: with PTA stripped out, the cost is now bounded by type checking. Performance: - The importer.Config.TypeCheckFuncBodies predicate supports setting the 'IgnoreFuncBodies' typechecker flag on a per-package basis. This means we can load dependencies from source more quickly if we only need exported types. (We avoid gcimport data because it may be absent or stale.) This also means we can run type-based queries on packages that aren't part of the pointer analysis scope. (Yay.) - Modes that require only type analysis of the query package run a "what" query first, and restrict their analysis scope to just that package and its dependencies (sans func bodies), making them much faster. - We call newOracle not oracle.New in Query, so that the 'needs' bitset isn't ignored (oops!). This makes the non-PTA queries faster. Also: - removed vestigial timers junk. - pos.go: existing position utilties split out into own file. Added parsePosFlag utility. - numerous cosmetic tweaks. + very basic tests. To do in follow-ups: - sophisticated editor integration of "what". - better tests. - refactoring of control flow as described in comment. - changes to "implements", "describe" commands. - update design doc + user manual. R=crawshaw, dominik.honnef CC=golang-dev, gri https://golang.org/cl/40630043
2013-12-13 08:04:55 -07:00
Pos: fset.Position(r.query.Pos()).String(),
Desc: r.obj.String(),
}
if pos := r.obj.Pos(); pos != token.NoPos { // Package objects have no Pos()
referrers.ObjPos = fset.Position(pos).String()
}
for _, ref := range r.refs {
go.tools/oracle: improvements to command set and performance. Command set: - what: an extremely fast query that parses a single file and returns the AST stack, package name and the set of query modes that apply to the current selection. Intended for GUI tools that need to grey out UI elements. - definition: shows the definition of an identifier. - pointsto: the PTA features of 'describe' have been split out into their own command. - describe: with PTA stripped out, the cost is now bounded by type checking. Performance: - The importer.Config.TypeCheckFuncBodies predicate supports setting the 'IgnoreFuncBodies' typechecker flag on a per-package basis. This means we can load dependencies from source more quickly if we only need exported types. (We avoid gcimport data because it may be absent or stale.) This also means we can run type-based queries on packages that aren't part of the pointer analysis scope. (Yay.) - Modes that require only type analysis of the query package run a "what" query first, and restrict their analysis scope to just that package and its dependencies (sans func bodies), making them much faster. - We call newOracle not oracle.New in Query, so that the 'needs' bitset isn't ignored (oops!). This makes the non-PTA queries faster. Also: - removed vestigial timers junk. - pos.go: existing position utilties split out into own file. Added parsePosFlag utility. - numerous cosmetic tweaks. + very basic tests. To do in follow-ups: - sophisticated editor integration of "what". - better tests. - refactoring of control flow as described in comment. - changes to "implements", "describe" commands. - update design doc + user manual. R=crawshaw, dominik.honnef CC=golang-dev, gri https://golang.org/cl/40630043
2013-12-13 08:04:55 -07:00
referrers.Refs = append(referrers.Refs, fset.Position(ref.NamePos).String())
}
res.Referrers = referrers
}