1110036 - "Save As" stops working from "Page Info" > "Media" pane. Progress of downloading is 0% forever

Status: RESOLVED FIXED

(Firefox :: Page Info Window, defect)

Description

[Alice0775 White]

Build Identifier:
https://hg.mozilla.org/mozilla-central/rev/5b01216f97f8
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:37.0) Gecko/20100101 Firefox/37.0 ID:20141210030207
https://hg.mozilla.org/releases/mozilla-aurora/rev/099b946a520f
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:36.0) Gecko/20100101 Firefox/36.0 ID:20141210004006



[Tracking Requested - why for this release]: Broken a "Page info" feature due to regression

The problem happens Aurora36.0a2 and Nightly37.0a1.

Steps To Reproduce:
1. Start Browser with newly created profile
2. Open any page (e.g. https://www.mozilla.org/en-US/firefox/36.0a2/firstrun/ or any page )
3. Open Page Info and switch to "Media" pane
4. Attempt to save an Image(e.g. https://mozorg.cdn.mozilla.net/media/img/tabzilla/tab.png )

Actual Results:
Although the browser tries to download, progress of downloading is 0% forever.

Expected Results:
Download should be performed properly.

Comment 1

[Alice0775 White]

Regression window(m-i)
Good:
https://hg.mozilla.org/integration/mozilla-inbound/rev/2e541f48b3c8
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:36.0) Gecko/20100101 Firefox/36.0 ID:20141118053526
Bad:
https://hg.mozilla.org/integration/mozilla-inbound/rev/f268d422e9fb
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:36.0) Gecko/20100101 Firefox/36.0 ID:20141118060426
Pushlog:
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=2e541f48b3c8&tochange=f268d422e9fb

Regressed by Bug 704320

Comment 2

[Alice0775 White]

In local build,
Last Good: d76316a3c1dd
First Bad: 4a7db36c3c75

Triggered by: 	4a7db36c3c75	Sid Stamm — Bug 704320 - Add referrer policy support for nsIWebBrowserPersist and its consumers. (r=jst)

Comment 3

[Sid Stamm [:geekboy or :sstamm]]

Hm, I was hoping that the download manager would just use the default referrer policy, but apparently it's halting the loads?

Does this cause any assertions or warnings on the console in a debug build?  (Some output would help me debug).

Also, see: bug 1073187 ... the referrer policy is not yet being used by Downloads.jsm (but the requests should still work)

Comment 4

[Alice0775 White]

(In reply to Sid Stamm [:geekboy or :sstamm] from comment #3)
> Hm, I was hoping that the download manager would just use the default
> referrer policy, but apparently it's halting the loads?

Sorry, I do not understand what you request.

> 
> Does this cause any assertions or warnings on the console in a debug build? 
> (Some output would help me debug).
> 

No error.

Comment 5

[Alice0775 White]

(In reply to Alice0775 White from comment #4)
> (In reply to Sid Stamm [:geekboy or :sstamm] from comment #3)
> > Hm, I was hoping that the download manager would just use the default
> > referrer policy, but apparently it's halting the loads?
> 
> Sorry, I do not understand what you request.
> 
> > 
> > Does this cause any assertions or warnings on the console in a debug build? 
> > (Some output would help me debug).
> > 
> 
> No error.

Oops, There are two errors in the Browser Console:

NS_ERROR_IN_PROGRESS: Component returned failure code: 0x804b000f (NS_ERROR_IN_PROGRESS) [nsICacheEntry.dataSize] pageInfo.js:607:0

Ci is not defined contentAreaUtils.js:424:0

Comment 6

[Sid Stamm [:geekboy or :sstamm]]

(In reply to Alice0775 White from comment #4)
> Sorry, I do not understand what you request.

Apologies, that was not at all clear.  My assumption in implementing the referrer policy stuff was that any downloads would use the standard/default referrer policy (and not break).  I was asking for confirmation that the behavior is: Firefox does not send the request to download/save the file.  It simply does not make the HTTP request when it should, correct.  

> > Does this cause any assertions or warnings on the console in a debug build? 
> > (Some output would help me debug).
> > 
> 
> No error.

Specifically, I was curious about command-line errors (like assertions or warnings) when running it in a debug build.  Sometimes that is more helpful than browser console warnings when debugging network-level stuff.

I see this when I run the STR in a debug build (dumping here so when I get some time to fully dig into this I have an idea of where to start):

[Parent 27900] WARNING: NS_ENSURE_TRUE(aURI) failed: file /home/sstamm/src/mozilla-central/netwerk/dns/nsEffectiveTLDService.cpp, line 158
[Parent 27900] WARNING: NS_ENSURE_SUCCESS(rv, rv) failed with result 0x80070057: file /home/sstamm/src/mozilla-central/dom/base/ThirdPartyUtil.cpp, line 353
spec=/grain.png
[Parent 27900] WARNING: malformed url: no scheme: file /home/sstamm/src/mozilla-central/netwerk/base/src/nsStandardURL.cpp, line 794
--DOMWINDOW == 11 (0x7f4543f1c400) [pid = 27900] [serial = 12] [outer = 0x7f4544307c00] [url = about:blank]
--DOMWINDOW == 10 (0x7f4544307c00) [pid = 27900] [serial = 8] [outer = (nil)] [url = about:blank]
WARNING: content window passed to PrivateBrowsingUtils.isWindowPrivate. Use isContentWindowPrivate instead (but only for frame scripts).
pbu_isWindowPrivate@resource://gre/modules/PrivateBrowsingUtils.jsm:25:14
internalPersist@chrome://global/content/contentAreaUtils.js:381:19
continueSave@chrome://global/content/contentAreaUtils.js:328:5
internalSave/<@chrome://global/content/contentAreaUtils.js:301:7
Handler.prototype.process@resource://gre/modules/Promise.jsm -> resource://gre/modules/Promise-backend.js:870:23
this.PromiseWalker.walkerLoop@resource://gre/modules/Promise.jsm -> resource://gre/modules/Promise-backend.js:749:7
this.PromiseWalker.scheduleWalkerLoop/<@resource://gre/modules/Promise.jsm -> resource://gre/modules/Promise-backend.js:691:37
*************************
A coding exception was thrown in a Promise resolution callback.
See https://developer.mozilla.org/Mozilla/JavaScript_code_modules/Promise.jsm/Promise

Full message: ReferenceError: Ci is not defined
Full stack: internalPersist@chrome://global/content/contentAreaUtils.js:424:1
continueSave@chrome://global/content/contentAreaUtils.js:328:5
internalSave/<@chrome://global/content/contentAreaUtils.js:301:7
Handler.prototype.process@resource://gre/modules/Promise.jsm -> resource://gre/modules/Promise-backend.js:870:23
this.PromiseWalker.walkerLoop@resource://gre/modules/Promise.jsm -> resource://gre/modules/Promise-backend.js:749:7
this.PromiseWalker.scheduleWalkerLoop/<@resource://gre/modules/Promise.jsm -> resource://gre/modules/Promise-backend.js:691:37

*************************
[Child 27957] WARNING: NS_ENSURE_TRUE(node) failed: file /home/sstamm/src/mozilla-central/layout/base/nsDocumentViewer.cpp, line 3524
[Child 27957] WARNING: NS_ENSURE_TRUE(node) failed: file /home/sstamm/src/mozilla-central/layout/base/nsDocumentViewer.cpp, line 3524
[Child 27957] WARNING: NS_ENSURE_TRUE(node) failed: file /home/sstamm/src/mozilla-central/layout/base/nsDocumentViewer.cpp, line 3524
JavaScript strict warning: resource://gre/modules/DownloadUtils.jsm, line 79: SyntaxError: applying the 'delete' operator to an unqualified name is deprecated
JavaScript warning: resource://gre/modules/DownloadUtils.jsm, line 273: JavaScript 1.7's let blocks are deprecated
[Parent 27900] WARNING: No docshells for remote frames!: file /home/sstamm/src/mozilla-central/dom/base/nsFrameLoader.cpp, line 514
[Parent 27900] WARNING: No docshells for remote frames!: file /home/sstamm/src/mozilla-central/dom/base/nsFrameLoader.cpp, line 514
[Child 27957] WARNING: NS_ENSURE_TRUE(uri) failed: file /home/sstamm/src/mozilla-central/dom/ipc/ContentChild.cpp, line 1951
[Parent 27900] WARNING: We should have hit the document element...: file /home/sstamm/src/mozilla-central/layout/xul/BoxObject.cpp, line 183

Comment 7

[Philip Chee]

> Ci is not defined contentAreaUtils.js:424:0
I'm getting this as well.
https://hg.mozilla.org/mozilla-central/rev/4a7db36c3c75#l15.12
This introduced an unnecessary dependency on Ci. which may or may not be defined in the context of the page info window.

Comment 8

[Philip Chee]

Attachment: Patch v1.0 replace "Ci." shortcut with Components.interfaces.

Simple search and replace.

Comment 9

[Sid Stamm [:geekboy or :sstamm]]

Comment on attachment 8536203 [details] [diff] [review]
Patch v1.0 replace "Ci." shortcut with Components.interfaces.

Review of attachment 8536203 [details] [diff] [review]:
-----------------------------------------------------------------

Yep, sorry for the delay, that looks great and seems to fix the problem for me.

Comment 10

[Philip Chee]

http://hg.mozilla.org/integration/mozilla-inbound/rev/cd44fbb9f362

Comment 11

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/cd44fbb9f362

Comment 12

[Sylvestre Ledru [:Sylvestre]]

Philip, could you fill an uplift request to aurora? Thanks

Comment 13

[Philip Chee]

Comment on attachment 8536203 [details] [diff] [review]
Patch v1.0 replace "Ci." shortcut with Components.interfaces.

Approval Request Comment
[Feature/regressing bug #]: Bug 704320 - Implement <meta name="referrer">
[User impact if declined]: Users unable to download files from the Page Info window.
[Describe test coverage new/current, TBPL]: did not cause any tests to fail on -inbound or on -central
[Risks and why]: no risk. Bustage/Typo fix.
[String/UUID change made/needed]: None.

Comment 14

[Sylvestre Ledru [:Sylvestre]]

Comment on attachment 8536203 [details] [diff] [review]
Patch v1.0 replace "Ci." shortcut with Components.interfaces.

Thanks

Comment 15

[Philip Chee]

https://hg.mozilla.org/releases/mozilla-aurora/rev/10e3c131bbaf