Add cdp support for xpath scrapers (#625)

Co-authored-by: WithoutPants <53250216+WithoutPants@users.noreply.github.com>

vendor/github.com/chromedp/cdproto/inspector/easyjson.go

@@ -0,0 +1,321 @@
+// Code generated by easyjson for marshaling/unmarshaling. DO NOT EDIT.
+
+package inspector
+
+import (
+	json "encoding/json"
+	easyjson "github.com/mailru/easyjson"
+	jlexer "github.com/mailru/easyjson/jlexer"
+	jwriter "github.com/mailru/easyjson/jwriter"
+)
+
+// suppress unused package warning
+var (
+	_ *json.RawMessage
+	_ *jlexer.Lexer
+	_ *jwriter.Writer
+	_ easyjson.Marshaler
+)
+
+func easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector(in *jlexer.Lexer, out *EventTargetReloadedAfterCrash) {
+	isTopLevel := in.IsStart()
+	if in.IsNull() {
+		if isTopLevel {
+			in.Consumed()
+		}
+		in.Skip()
+		return
+	}
+	in.Delim('{')
+	for !in.IsDelim('}') {
+		key := in.UnsafeString()
+		in.WantColon()
+		if in.IsNull() {
+			in.Skip()
+			in.WantComma()
+			continue
+		}
+		switch key {
+		default:
+			in.SkipRecursive()
+		}
+		in.WantComma()
+	}
+	in.Delim('}')
+	if isTopLevel {
+		in.Consumed()
+	}
+}
+func easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector(out *jwriter.Writer, in EventTargetReloadedAfterCrash) {
+	out.RawByte('{')
+	first := true
+	_ = first
+	out.RawByte('}')
+}
+
+// MarshalJSON supports json.Marshaler interface
+func (v EventTargetReloadedAfterCrash) MarshalJSON() ([]byte, error) {
+	w := jwriter.Writer{}
+	easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector(&w, v)
+	return w.Buffer.BuildBytes(), w.Error
+}
+
+// MarshalEasyJSON supports easyjson.Marshaler interface
+func (v EventTargetReloadedAfterCrash) MarshalEasyJSON(w *jwriter.Writer) {
+	easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector(w, v)
+}
+
+// UnmarshalJSON supports json.Unmarshaler interface
+func (v *EventTargetReloadedAfterCrash) UnmarshalJSON(data []byte) error {
+	r := jlexer.Lexer{Data: data}
+	easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector(&r, v)
+	return r.Error()
+}
+
+// UnmarshalEasyJSON supports easyjson.Unmarshaler interface
+func (v *EventTargetReloadedAfterCrash) UnmarshalEasyJSON(l *jlexer.Lexer) {
+	easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector(l, v)
+}
+func easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector1(in *jlexer.Lexer, out *EventTargetCrashed) {
+	isTopLevel := in.IsStart()
+	if in.IsNull() {
+		if isTopLevel {
+			in.Consumed()
+		}
+		in.Skip()
+		return
+	}
+	in.Delim('{')
+	for !in.IsDelim('}') {
+		key := in.UnsafeString()
+		in.WantColon()
+		if in.IsNull() {
+			in.Skip()
+			in.WantComma()
+			continue
+		}
+		switch key {
+		default:
+			in.SkipRecursive()
+		}
+		in.WantComma()
+	}
+	in.Delim('}')
+	if isTopLevel {
+		in.Consumed()
+	}
+}
+func easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector1(out *jwriter.Writer, in EventTargetCrashed) {
+	out.RawByte('{')
+	first := true
+	_ = first
+	out.RawByte('}')
+}
+
+// MarshalJSON supports json.Marshaler interface
+func (v EventTargetCrashed) MarshalJSON() ([]byte, error) {
+	w := jwriter.Writer{}
+	easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector1(&w, v)
+	return w.Buffer.BuildBytes(), w.Error
+}
+
+// MarshalEasyJSON supports easyjson.Marshaler interface
+func (v EventTargetCrashed) MarshalEasyJSON(w *jwriter.Writer) {
+	easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector1(w, v)
+}
+
+// UnmarshalJSON supports json.Unmarshaler interface
+func (v *EventTargetCrashed) UnmarshalJSON(data []byte) error {
+	r := jlexer.Lexer{Data: data}
+	easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector1(&r, v)
+	return r.Error()
+}
+
+// UnmarshalEasyJSON supports easyjson.Unmarshaler interface
+func (v *EventTargetCrashed) UnmarshalEasyJSON(l *jlexer.Lexer) {
+	easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector1(l, v)
+}
+func easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector2(in *jlexer.Lexer, out *EventDetached) {
+	isTopLevel := in.IsStart()
+	if in.IsNull() {
+		if isTopLevel {
+			in.Consumed()
+		}
+		in.Skip()
+		return
+	}
+	in.Delim('{')
+	for !in.IsDelim('}') {
+		key := in.UnsafeString()
+		in.WantColon()
+		if in.IsNull() {
+			in.Skip()
+			in.WantComma()
+			continue
+		}
+		switch key {
+		case "reason":
+			(out.Reason).UnmarshalEasyJSON(in)
+		default:
+			in.SkipRecursive()
+		}
+		in.WantComma()
+	}
+	in.Delim('}')
+	if isTopLevel {
+		in.Consumed()
+	}
+}
+func easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector2(out *jwriter.Writer, in EventDetached) {
+	out.RawByte('{')
+	first := true
+	_ = first
+	{
+		const prefix string = ",\"reason\":"
+		out.RawString(prefix[1:])
+		(in.Reason).MarshalEasyJSON(out)
+	}
+	out.RawByte('}')
+}
+
+// MarshalJSON supports json.Marshaler interface
+func (v EventDetached) MarshalJSON() ([]byte, error) {
+	w := jwriter.Writer{}
+	easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector2(&w, v)
+	return w.Buffer.BuildBytes(), w.Error
+}
+
+// MarshalEasyJSON supports easyjson.Marshaler interface
+func (v EventDetached) MarshalEasyJSON(w *jwriter.Writer) {
+	easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector2(w, v)
+}
+
+// UnmarshalJSON supports json.Unmarshaler interface
+func (v *EventDetached) UnmarshalJSON(data []byte) error {
+	r := jlexer.Lexer{Data: data}
+	easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector2(&r, v)
+	return r.Error()
+}
+
+// UnmarshalEasyJSON supports easyjson.Unmarshaler interface
+func (v *EventDetached) UnmarshalEasyJSON(l *jlexer.Lexer) {
+	easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector2(l, v)
+}
+func easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector3(in *jlexer.Lexer, out *EnableParams) {
+	isTopLevel := in.IsStart()
+	if in.IsNull() {
+		if isTopLevel {
+			in.Consumed()
+		}
+		in.Skip()
+		return
+	}
+	in.Delim('{')
+	for !in.IsDelim('}') {
+		key := in.UnsafeString()
+		in.WantColon()
+		if in.IsNull() {
+			in.Skip()
+			in.WantComma()
+			continue
+		}
+		switch key {
+		default:
+			in.SkipRecursive()
+		}
+		in.WantComma()
+	}
+	in.Delim('}')
+	if isTopLevel {
+		in.Consumed()
+	}
+}
+func easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector3(out *jwriter.Writer, in EnableParams) {
+	out.RawByte('{')
+	first := true
+	_ = first
+	out.RawByte('}')
+}
+
+// MarshalJSON supports json.Marshaler interface
+func (v EnableParams) MarshalJSON() ([]byte, error) {
+	w := jwriter.Writer{}
+	easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector3(&w, v)
+	return w.Buffer.BuildBytes(), w.Error
+}
+
+// MarshalEasyJSON supports easyjson.Marshaler interface
+func (v EnableParams) MarshalEasyJSON(w *jwriter.Writer) {
+	easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector3(w, v)
+}
+
+// UnmarshalJSON supports json.Unmarshaler interface
+func (v *EnableParams) UnmarshalJSON(data []byte) error {
+	r := jlexer.Lexer{Data: data}
+	easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector3(&r, v)
+	return r.Error()
+}
+
+// UnmarshalEasyJSON supports easyjson.Unmarshaler interface
+func (v *EnableParams) UnmarshalEasyJSON(l *jlexer.Lexer) {
+	easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector3(l, v)
+}
+func easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector4(in *jlexer.Lexer, out *DisableParams) {
+	isTopLevel := in.IsStart()
+	if in.IsNull() {
+		if isTopLevel {
+			in.Consumed()
+		}
+		in.Skip()
+		return
+	}
+	in.Delim('{')
+	for !in.IsDelim('}') {
+		key := in.UnsafeString()
+		in.WantColon()
+		if in.IsNull() {
+			in.Skip()
+			in.WantComma()
+			continue
+		}
+		switch key {
+		default:
+			in.SkipRecursive()
+		}
+		in.WantComma()
+	}
+	in.Delim('}')
+	if isTopLevel {
+		in.Consumed()
+	}
+}
+func easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector4(out *jwriter.Writer, in DisableParams) {
+	out.RawByte('{')
+	first := true
+	_ = first
+	out.RawByte('}')
+}
+
+// MarshalJSON supports json.Marshaler interface
+func (v DisableParams) MarshalJSON() ([]byte, error) {
+	w := jwriter.Writer{}
+	easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector4(&w, v)
+	return w.Buffer.BuildBytes(), w.Error
+}
+
+// MarshalEasyJSON supports easyjson.Marshaler interface
+func (v DisableParams) MarshalEasyJSON(w *jwriter.Writer) {
+	easyjsonC5a4559bEncodeGithubComChromedpCdprotoInspector4(w, v)
+}
+
+// UnmarshalJSON supports json.Unmarshaler interface
+func (v *DisableParams) UnmarshalJSON(data []byte) error {
+	r := jlexer.Lexer{Data: data}
+	easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector4(&r, v)
+	return r.Error()
+}
+
+// UnmarshalEasyJSON supports easyjson.Unmarshaler interface
+func (v *DisableParams) UnmarshalEasyJSON(l *jlexer.Lexer) {
+	easyjsonC5a4559bDecodeGithubComChromedpCdprotoInspector4(l, v)
+}

vendor/github.com/chromedp/cdproto/inspector/events.go

@@ -0,0 +1,22 @@
+package inspector
+
+// Code generated by cdproto-gen. DO NOT EDIT.
+
+// EventDetached fired when remote debugging connection is about to be
+// terminated. Contains detach reason.
+//
+// See: https://chromedevtools.github.io/devtools-protocol/tot/Inspector#event-detached
+type EventDetached struct {
+	Reason DetachReason `json:"reason"` // The reason why connection has been terminated.
+}
+
+// EventTargetCrashed fired when debugging target has crashed.
+//
+// See: https://chromedevtools.github.io/devtools-protocol/tot/Inspector#event-targetCrashed
+type EventTargetCrashed struct{}
+
+// EventTargetReloadedAfterCrash fired when debugging target has reloaded
+// after crash.
+//
+// See: https://chromedevtools.github.io/devtools-protocol/tot/Inspector#event-targetReloadedAfterCrash
+type EventTargetReloadedAfterCrash struct{}

vendor/github.com/chromedp/cdproto/inspector/inspector.go

@@ -0,0 +1,49 @@
+// Package inspector provides the Chrome DevTools Protocol
+// commands, types, and events for the Inspector domain.
+//
+// Generated by the cdproto-gen command.
+package inspector
+
+// Code generated by cdproto-gen. DO NOT EDIT.
+
+import (
+	"context"
+
+	"github.com/chromedp/cdproto/cdp"
+)
+
+// DisableParams disables inspector domain notifications.
+type DisableParams struct{}
+
+// Disable disables inspector domain notifications.
+//
+// See: https://chromedevtools.github.io/devtools-protocol/tot/Inspector#method-disable
+func Disable() *DisableParams {
+	return &DisableParams{}
+}
+
+// Do executes Inspector.disable against the provided context.
+func (p *DisableParams) Do(ctx context.Context) (err error) {
+	return cdp.Execute(ctx, CommandDisable, nil, nil)
+}
+
+// EnableParams enables inspector domain notifications.
+type EnableParams struct{}
+
+// Enable enables inspector domain notifications.
+//
+// See: https://chromedevtools.github.io/devtools-protocol/tot/Inspector#method-enable
+func Enable() *EnableParams {
+	return &EnableParams{}
+}
+
+// Do executes Inspector.enable against the provided context.
+func (p *EnableParams) Do(ctx context.Context) (err error) {
+	return cdp.Execute(ctx, CommandEnable, nil, nil)
+}
+
+// Command names.
+const (
+	CommandDisable = "Inspector.disable"
+	CommandEnable  = "Inspector.enable"
+)

vendor/github.com/chromedp/cdproto/inspector/types.go

@@ -0,0 +1,61 @@
+package inspector
+
+// Code generated by cdproto-gen. DO NOT EDIT.
+
+import (
+	"errors"
+
+	"github.com/mailru/easyjson"
+	"github.com/mailru/easyjson/jlexer"
+	"github.com/mailru/easyjson/jwriter"
+)
+
+// DetachReason detach reason.
+//
+// See: ( -- none -- )
+type DetachReason string
+
+// String returns the DetachReason as string value.
+func (t DetachReason) String() string {
+	return string(t)
+}
+
+// DetachReason values.
+const (
+	DetachReasonTargetClosed         DetachReason = "target_closed"
+	DetachReasonCanceledByUser       DetachReason = "canceled_by_user"
+	DetachReasonReplacedWithDevtools DetachReason = "replaced_with_devtools"
+	DetachReasonRenderProcessGone    DetachReason = "Render process gone."
+)
+
+// MarshalEasyJSON satisfies easyjson.Marshaler.
+func (t DetachReason) MarshalEasyJSON(out *jwriter.Writer) {
+	out.String(string(t))
+}
+
+// MarshalJSON satisfies json.Marshaler.
+func (t DetachReason) MarshalJSON() ([]byte, error) {
+	return easyjson.Marshal(t)
+}
+
+// UnmarshalEasyJSON satisfies easyjson.Unmarshaler.
+func (t *DetachReason) UnmarshalEasyJSON(in *jlexer.Lexer) {
+	switch DetachReason(in.String()) {
+	case DetachReasonTargetClosed:
+		*t = DetachReasonTargetClosed
+	case DetachReasonCanceledByUser:
+		*t = DetachReasonCanceledByUser
+	case DetachReasonReplacedWithDevtools:
+		*t = DetachReasonReplacedWithDevtools
+	case DetachReasonRenderProcessGone:
+		*t = DetachReasonRenderProcessGone
+
+	default:
+		in.AddError(errors.New("unknown DetachReason value"))
+	}
+}
+
+// UnmarshalJSON satisfies json.Unmarshaler.
+func (t *DetachReason) UnmarshalJSON(buf []byte) error {
+	return easyjson.Unmarshal(buf, t)
+}