mirrors/electron
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages 1 Wiki Activity Actions 1

Implement GetLoadTimingInfo in JsAsker class.

Browse source 
This is required to avoid a crash in blink when implementing a custom protocol
handler that deals with redirects.
This commit is contained in:
Thiago de Arruda 2017-11-03 14:53:08 -03:00
parent 6b2ddc4814
commit 6b666b69c0
1 changed files with 13 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

13
atom/browser/net/js_asker.h
View file

@ -71,6 +71,7 @@ class JsAsker : public RequestJob {
void Start() override { void Start() override {
std::unique_ptr<base::DictionaryValue> request_details( std::unique_ptr<base::DictionaryValue> request_details(
new base::DictionaryValue); new base::DictionaryValue);
request_start_time_ = base::TimeTicks::Now();
FillRequestDetails(request_details.get(), RequestJob::request()); FillRequestDetails(request_details.get(), RequestJob::request());
content::BrowserThread::PostTask( content::BrowserThread::PostTask(
content::BrowserThread::UI, FROM_HERE, content::BrowserThread::UI, FROM_HERE,
@ -86,6 +87,15 @@ class JsAsker : public RequestJob {
int GetResponseCode() const override { return net::HTTP_OK; } int GetResponseCode() const override { return net::HTTP_OK; }
// NOTE: We have to implement this method or risk a crash in blink for
// redirects!
void GetLoadTimingInfo(net::LoadTimingInfo* load_timing_info) const override {
load_timing_info->send_start = request_start_time_;
load_timing_info->send_end = request_start_time_;
load_timing_info->request_start = request_start_time_;
load_timing_info->receive_headers_end = response_start_time_;
}
void GetResponseInfo(net::HttpResponseInfo* info) override { void GetResponseInfo(net::HttpResponseInfo* info) override {
info->headers = new net::HttpResponseHeaders(""); info->headers = new net::HttpResponseHeaders("");
} }
@ -93,6 +103,7 @@ class JsAsker : public RequestJob {
// Called when the JS handler has sent the response, we need to decide whether // Called when the JS handler has sent the response, we need to decide whether
// to start, or fail the job. // to start, or fail the job.
void OnResponse(bool success, std::unique_ptr<base::Value> value) { void OnResponse(bool success, std::unique_ptr<base::Value> value) {
response_start_time_ = base::TimeTicks::Now();
int error = net::ERR_NOT_IMPLEMENTED; int error = net::ERR_NOT_IMPLEMENTED;
if (success && value && !internal::IsErrorOptions(value.get(), &error)) { if (success && value && !internal::IsErrorOptions(value.get(), &error)) {
StartAsync(std::move(value)); StartAsync(std::move(value));
@ -105,6 +116,8 @@ class JsAsker : public RequestJob {
v8::Isolate* isolate_; v8::Isolate* isolate_;
net::URLRequestContextGetter* request_context_getter_; net::URLRequestContextGetter* request_context_getter_;
JavaScriptHandler handler_; JavaScriptHandler handler_;
base::TimeTicks request_start_time_;
base::TimeTicks response_start_time_;
base::WeakPtrFactory<JsAsker> weak_factory_; base::WeakPtrFactory<JsAsker> weak_factory_;