Refactor FeedbackReport, FeedbackUpload to remove chrome/net deps

Refactor FeedbackReport, FeedbackUpload to remove chrome/net deps

Remove use of BrowserContext and BrowserThread from FeedbackReport and
FeedbackUpload. Moves net/ specific functionality into a new class
FeedbackUploadChrome, and moves using the browser context out of
the base classes.

CrOS should be able to directly use feedback_report.cc/h &
feedback_uploader.cc/h, providing a curl-based upload version.

BUG=341554
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=257534

[achaulk, 2014-03-06 20:52:42]

Once this and the other CL goes in I'll move the common files to base/feedback
(or wherever we decide to put it)

[rkc, 2014-03-10 21:55:36]

Adding Zach to do the main review. I'll give it another pass once he's done.

[Zachary Kuznia, 2014-03-10 22:04:51]

https://codereview.chromium.org/186663011/diff/1/chrome/browser/feedback/feed...
File chrome/browser/feedback/feedback_uploader_chrome.cc (right):

https://codereview.chromium.org/186663011/diff/1/chrome/browser/feedback/feed...
chrome/browser/feedback/feedback_uploader_chrome.cc:28: // static
This isn't static

https://codereview.chromium.org/186663011/diff/1/chrome/browser/feedback/feed...
chrome/browser/feedback/feedback_uploader_chrome.cc:32: return base::FilePath();
This seems like bad default behavior.  Is there a reason you don't CHECK() here,
or return an error?

[achaulk, 2014-03-10 22:15:12]

https://codereview.chromium.org/186663011/diff/1/chrome/browser/feedback/feed...
File chrome/browser/feedback/feedback_uploader_chrome.cc (right):

https://codereview.chromium.org/186663011/diff/1/chrome/browser/feedback/feed...
chrome/browser/feedback/feedback_uploader_chrome.cc:28: // static
Ah. Yeah I moved it from a static method and forgot about that

On 2014/03/10 22:04:51, Zachary Kuznia wrote:
> This isn't static

https://codereview.chromium.org/186663011/diff/1/chrome/browser/feedback/feed...
chrome/browser/feedback/feedback_uploader_chrome.cc:32: return base::FilePath();
This is the existing behaviour. Though I think rkc was going to/did add a check
for a null browsercontext, so maybe this doesn't matter anymore

On 2014/03/10 22:04:51, Zachary Kuznia wrote:
> This seems like bad default behavior.  Is there a reason you don't CHECK()
here,
> or return an error?

[rkc, 2014-03-10 22:18:57]

https://codereview.chromium.org/186663011/diff/1/chrome/browser/feedback/feed...
File chrome/browser/feedback/feedback_uploader_chrome.cc (right):

https://codereview.chromium.org/186663011/diff/1/chrome/browser/feedback/feed...
chrome/browser/feedback/feedback_uploader_chrome.cc:32: return base::FilePath();
On 2014/03/10 22:15:13, achaulk wrote:
> This is the existing behaviour. Though I think rkc was going to/did add a
check
> for a null browsercontext, so maybe this doesn't matter anymore
> 
> On 2014/03/10 22:04:51, Zachary Kuznia wrote:
> > This seems like bad default behavior.  Is there a reason you don't CHECK()
> here,
> > or return an error?
> 

In the original code, the reports_path was checked for it being empty, in which
case it the upload would be dropped. I haven't seen in this CL how it is being
handled.

The context being null is quite possible and an acceptable situation in Chrome,
particularly near shutdown, hence the context being NULL is not really an error.

[Zachary Kuznia, 2014-03-12 19:08:10]

https://codereview.chromium.org/186663011/diff/1/chrome/browser/feedback/feed...
File chrome/browser/feedback/feedback_uploader_chrome.cc (right):

https://codereview.chromium.org/186663011/diff/1/chrome/browser/feedback/feed...
chrome/browser/feedback/feedback_uploader_chrome.cc:32: return base::FilePath();
On 2014/03/10 22:18:57, Rahul Chaturvedi wrote:
> On 2014/03/10 22:15:13, achaulk wrote:
> > This is the existing behaviour. Though I think rkc was going to/did add a
> check
> > for a null browsercontext, so maybe this doesn't matter anymore
> > 
> > On 2014/03/10 22:04:51, Zachary Kuznia wrote:
> > > This seems like bad default behavior.  Is there a reason you don't CHECK()
> > here,
> > > or return an error?
> > 
> 
> In the original code, the reports_path was checked for it being empty, in
which
> case it the upload would be dropped. I haven't seen in this CL how it is being
> handled.
> 
> The context being null is quite possible and an acceptable situation in
Chrome,
> particularly near shutdown, hence the context being NULL is not really an
error.

In the function that calls this, the is a CHECK(context_).  Since it's not
acceptable there, you should add a CHECK() here as well so that they don't
diverge in a refactor.

[achaulk, 2014-03-12 20:17:56]

https://codereview.chromium.org/186663011/diff/1/chrome/browser/feedback/feed...
File chrome/browser/feedback/feedback_uploader_chrome.cc (right):

https://codereview.chromium.org/186663011/diff/1/chrome/browser/feedback/feed...
chrome/browser/feedback/feedback_uploader_chrome.cc:32: return base::FilePath();
I'll just inline it into the constructor which is doing the check anyway

On 2014/03/12 19:08:10, Zachary Kuznia wrote:
> On 2014/03/10 22:18:57, Rahul Chaturvedi wrote:
> > On 2014/03/10 22:15:13, achaulk wrote:
> > > This is the existing behaviour. Though I think rkc was going to/did add a
> > check
> > > for a null browsercontext, so maybe this doesn't matter anymore
> > > 
> > > On 2014/03/10 22:04:51, Zachary Kuznia wrote:
> > > > This seems like bad default behavior.  Is there a reason you don't
CHECK()
> > > here,
> > > > or return an error?
> > > 
> > 
> > In the original code, the reports_path was checked for it being empty, in
> which
> > case it the upload would be dropped. I haven't seen in this CL how it is
being
> > handled.
> > 
> > The context being null is quite possible and an acceptable situation in
> Chrome,
> > particularly near shutdown, hence the context being NULL is not really an
> error.
> 
> In the function that calls this, the is a CHECK(context_).  Since it's not
> acceptable there, you should add a CHECK() here as well so that they don't
> diverge in a refactor.

[Zachary Kuznia, 2014-03-12 20:21:08]

https://codereview.chromium.org/186663011/diff/1/chrome/browser/feedback/feed...
File chrome/browser/feedback/feedback_uploader_chrome.cc (right):

https://codereview.chromium.org/186663011/diff/1/chrome/browser/feedback/feed...
chrome/browser/feedback/feedback_uploader_chrome.cc:32: return base::FilePath();
On 2014/03/12 20:17:57, achaulk wrote:
> I'll just inline it into the constructor which is doing the check anyway
> 
> On 2014/03/12 19:08:10, Zachary Kuznia wrote:
> > On 2014/03/10 22:18:57, Rahul Chaturvedi wrote:
> > > On 2014/03/10 22:15:13, achaulk wrote:
> > > > This is the existing behaviour. Though I think rkc was going to/did add
a
> > > check
> > > > for a null browsercontext, so maybe this doesn't matter anymore
> > > > 
> > > > On 2014/03/10 22:04:51, Zachary Kuznia wrote:
> > > > > This seems like bad default behavior.  Is there a reason you don't
> CHECK()
> > > > here,
> > > > > or return an error?
> > > > 
> > > 
> > > In the original code, the reports_path was checked for it being empty, in
> > which
> > > case it the upload would be dropped. I haven't seen in this CL how it is
> being
> > > handled.
> > > 
> > > The context being null is quite possible and an acceptable situation in
> > Chrome,
> > > particularly near shutdown, hence the context being NULL is not really an
> > error.
> > 
> > In the function that calls this, the is a CHECK(context_).  Since it's not
> > acceptable there, you should add a CHECK() here as well so that they don't
> > diverge in a refactor.
> 

Sounds good.

[achaulk, 2014-03-14 18:26:17]

ping

[rkc, 2014-03-14 21:35:54]

lgtm

I think Zach is OOO.
LGTM

[achaulk, 2014-03-14 21:41:19]

The CQ bit was checked by [email protected]

[commit-bot: I haz the power, 2014-03-14 21:42:06]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/186663011/20001

[commit-bot: I haz the power, 2014-03-14 21:42:27]

The CQ bit was unchecked by [email protected]

[commit-bot: I haz the power, 2014-03-14 21:42:28]

Failed to apply patch for chrome/browser/feedback/feedback_uploader.h:
While running patch -p1 --forward --force --no-backup-if-mismatch;
  patching file chrome/browser/feedback/feedback_uploader.h
  Hunk #1 FAILED at 9.
  1 out of 3 hunks FAILED -- saving rejects to file
chrome/browser/feedback/feedback_uploader.h.rej

Patch:       chrome/browser/feedback/feedback_uploader.h
Index: chrome/browser/feedback/feedback_uploader.h
diff --git a/chrome/browser/feedback/feedback_uploader.h
b/chrome/browser/feedback/feedback_uploader.h
index
aa91613d1144ae3ec923e9a94431e7df62840e3c..2bef774c0714901090f255501ed9c4610066e961
100644
--- a/chrome/browser/feedback/feedback_uploader.h
+++ b/chrome/browser/feedback/feedback_uploader.h
@@ -9,41 +9,40 @@
 #include <string>
 
 #include "base/basictypes.h"
+#include "base/file_util.h"
 #include "base/memory/scoped_ptr.h"
 #include "base/memory/weak_ptr.h"
 #include "base/time/time.h"
 #include "base/timer/timer.h"
-#include "chrome/browser/feedback/feedback_uploader_delegate.h"
-#include
"components/browser_context_keyed_service/browser_context_keyed_service.h"
-
-namespace content {
-class BrowserContext;
-}
 
 namespace feedback {
 
+typedef base::Callback<void(const std::string&)> ReportDataCallback;
+
 class FeedbackReport;
 
 // FeedbackUploader is used to add a feedback report to the queue of reports
 // being uploaded. In case uploading a report fails, it is written to disk and
 // tried again when it's turn comes up next in the queue.
-class FeedbackUploader : public BrowserContextKeyedService,
-                         public base::SupportsWeakPtr<FeedbackUploader> {
+class FeedbackUploader : public base::SupportsWeakPtr<FeedbackUploader> {
  public:
-  explicit FeedbackUploader(content::BrowserContext* context);
+  explicit FeedbackUploader(const base::FilePath& path,
+                            base::SequencedWorkerPool* pool);
   virtual ~FeedbackUploader();
 
   // Queues a report for uploading.
   void QueueReport(const std::string& data);
 
- private:
+  base::FilePath GetFeedbackReportsPath() { return report_path_; }
+
+ protected:
   friend class FeedbackUploaderTest;
   struct ReportsUploadTimeComparator {
     bool operator()(FeedbackReport* a, FeedbackReport* b) const;
   };
 
   // Dispatches the report to be uploaded.
-  void DispatchReport(const std::string& data);
+  virtual void DispatchReport(const std::string& data) = 0;
 
   // Update our timer for uploading the next report.
   void UpdateUploadTimer();
@@ -51,11 +50,12 @@ class FeedbackUploader : public BrowserContextKeyedService,
   // Requeue this report with a delay.
   void RetryReport(const std::string& data);
 
+  void QueueReportWithDelay(const std::string& data, base::TimeDelta delay);
+
   void setup_for_test(const ReportDataCallback& dispatch_callback,
                       const base::TimeDelta& retry_delay);
 
-  // Browser context this uploader was created for.
-  content::BrowserContext* context_;
+  base::FilePath report_path_;
   // Timer to upload the next report at.
   base::OneShotTimer<FeedbackUploader> upload_timer_;
   // Priority queue of reports prioritized by the time the report is supposed
@@ -66,6 +66,8 @@ class FeedbackUploader : public BrowserContextKeyedService,
 
   ReportDataCallback dispatch_callback_;
   base::TimeDelta retry_delay_;
+  std::string url_;
+  base::SequencedWorkerPool* pool_;
 
   DISALLOW_COPY_AND_ASSIGN(FeedbackUploader);
 };

[achaulk, 2014-03-17 18:08:52]

The CQ bit was checked by [email protected]

[commit-bot: I haz the power, 2014-03-17 18:10:35]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/186663011/40001

[commit-bot: I haz the power, 2014-03-17 18:12:53]

The CQ bit was unchecked by [email protected]

[commit-bot: I haz the power, 2014-03-17 18:12:53]

Try jobs failed on following builders:
  tryserver.chromium on linux_chromium_chromeos_clang_dbg
  tryserver.chromium on linux_chromium_chromeos_rel
  tryserver.chromium on linux_chromium_clang_dbg
  tryserver.chromium on linux_chromium_rel
  tryserver.chromium on mac_chromium_compile_dbg
  tryserver.chromium on mac_chromium_rel
  tryserver.chromium on win_chromium_compile_dbg

[achaulk, 2014-03-17 20:58:40]

The CQ bit was checked by [email protected]

[commit-bot: I haz the power, 2014-03-17 20:59:46]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/186663011/40001

[commit-bot: I haz the power, 2014-03-18 00:00:08]

Change committed as 257534