1325876 - Show internal error page for HTTP responses with error codes (4xx, 5xx) and "Content-Length: 0" instead of blank page

Status: RESOLVED FIXED

(Core :: Networking, defect, P2)

Description

[Brad Lassey [:blassey] (use needinfo?)]

http://www.theworldin.com/article/12597/martin-luther-church-splitter?fsrc=scn/tw/wi/bl/ed/ didn't load

http://www.theworldin.com/article/12597/martin-luther-church-splitter does

Comment 1

[Wesley Huang [:wesley_huang] (EPM) (NI me)]

Using Chrome to open [1] gets HTTP ERROR 503 error. 

[1] http://www.theworldin.com/article/12597/martin-luther-church-splitter?fsrc=scn/tw/wi/bl/ed/

Comment 2

[Kevin Brosnan [Ex-Mozilla]]

Not a recent regression.

Comment 3

[Wesley Huang [:wesley_huang] (EPM) (NI me)]

(In reply to Wesley Huang [:wesley_huang] (EPM) (NI me) from comment #1)
> Using Chrome to open [1] gets HTTP ERROR 503 error. 
> 
> [1]
> http://www.theworldin.com/article/12597/martin-luther-church-
> splitter?fsrc=scn/tw/wi/bl/ed/

Server replies differently now. Now both chrome and firefox can load page thru the URL.

Comment 4

[Nevin Chen(Not active on Bugzilla)]

The website in comment 1 is normal now.
Julian gave me this website to test all the http status code: http://ozuma.sakura.ne.jp/httpstatus/
I've checked Chrome could display error message but we only show a blank page.

Comment 5

[James Willcox (:snorp) ([e-mail]) (he/him)]

I get this on desktop nightly too, so I think this might be intentional? Patrick, do we want show our internal load error pages if Content-length=0?

Comment 6

[Patrick McManus [:mcmanus]]

I think for the case of 0 bytes of content the front end should indeed show our local error page.. but if there is content then it probably makes sense to render it.

Comment 7

[Wesley Huang [:wesley_huang] (EPM) (NI me)]

So I guess the issue is not fennec specific. I'm changing the product/component.

Comment 8

[Sorina Florean [:sflorean]]

Hello and sorry for delay here. 

Tested with Huawei MediaPad M2 (Android 5.1.1) on latest Nightly (54.0a1 from 2017-01-29) and with page http://ozuma.sakura.ne.jp/httpstatus/ from comment 4.
Here are the results:
- http://ozuma.sakura.ne.jp/httpstatus/503 shows a blank page (same for 101, 200, 201, 202, 203, 206, 207, 208, 304, 305, 306, 400, 401 and the rest of the list)
- on Chrome I got the following message that page isn't working (http://ozuma.sakura.ne.jp/httpstatus/)
- the first link from description, is working in latest Nightly

Comment 9

[Julien Cristau [:jcristau]]

Too late for firefox 52, mass-wontfix.

Comment 10

[Manuel Bucher [:manuel]]

Using this bug to track showing error pages on http error responses with "Content-Length: 0", because it is the oldest I could find. Is there any specification which says that the browser should or may show an internal error page for http errors? In any case I think it would be less confusing for users instead of seeing an empty page.

Comment 15

[Sean Kim [:sekim]]

Was able to reproduce this issue with:

http.server import HTTPServer, BaseHTTPRequestHandler

class TestHandler(BaseHTTPRequestHandler):
    def do_GET(self):
        self.send_response(500)
        self.send_header("Content-Length", 0)
        self.end_headers()

if __name__ == "__main__":
    server_address = ('', 8000)
    httpd = HTTPServer(server_address, TestHandler)
    httpd.serve_forever()

Marking as P2 since it is relatively easy to fix this problem.

Comment 16

[Sean Kim [:sekim]]

Attachment: Bug 1325876 - Add a pref for showing internal error page for HTTP responses with error codes (4xx, 5xx) and Content-Length: 0 instead of blank page r=manuel,#necko

Comment 17

[Masatoshi Kimura [:emk]]

I found this change breaks session history when navigating to a page with HTTP status 500 and Content-Length: 0.

STR:

1. Open https://emk.name/test/500download.cgi in a new tab with the patched build.
2. Click the "500 zero byte response" link.

AR:
Back button will be disabled.

ER:
Back button should be enabled.

This bug already happens with downloads with HTTP status 500. (Try "500 download" link above).

Comment 18

[Sean Kim [:sekim]]

Attachment: back_button.png

Comment 19

[Sean Kim [:sekim]]

The updated patch now should have back button enabled unless we do something like ./mach run https://emk.name/test/500download.cgi

Comment 20

[Masatoshi Kimura [:emk]]

The latest patch didn't fix the problem at all.

unless we do something like ./mach run https://emk.name/test/500download.cgi

Why did you exclude it arbitrarily? It is the exact case I want to fix. Did you notice a phrase "in a new tab" in STR?

Moreover, even if we exclude that case, the patch still breaks the session history.
STR:

1. Open https://bugzilla.mozilla.org/show_bug.cgi?id=1325876 (this bug).
2. Click a link in this bug to navaigate to https://emk.name/test/500download.cgi.
3. Click the "500 zero byte response" link.
4. Click the back button.

AR:
Back to https://bugzilla.mozilla.org/show_bug.cgi?id=1325876.

ER:
Back to https://emk.name/test/500download.cgi.

I showed a new tab case just because the bug was visually obvious. The underlying problem is present regardless of a new tab is opened or not.

Comment 21

[Sean Kim [:sekim]]

Currently investigating the second STR.
I did further investigations and the first STR is an issue that persists on other errors (e.g. NS_ERROR_NET_ERROR_RESPONSE)

This was reproducible on other error pages with:

1. Run this Python script.

from http.server import HTTPServer, BaseHTTPRequestHandler

class TestHandler(BaseHTTPRequestHandler):
    def do_GET(self):
        if self.path == '/test':
            self.send_response(500)
            self.send_header("Content-Length", "0")
            self.end_headers()
        else:
            self.send_response(200)
            self.send_header("Content-type", "text/html")
            self.end_headers()
            html_content = """
            <html>
            <head>
                <title>Test Page</title>
            </head>
            <body>
                <a href="/test">HTTP 500 and unusual content-type</a>
            </body>
            </html>
            """
            self.wfile.write(html_content.encode('utf-8'))

if __name__ == "__main__":
    server_address = ('', 8000)
    httpd = HTTPServer(server_address, TestHandler)
    httpd.serve_forever()

2. Build this patch and run ./mach run localhost:8000

3. Navigate to "HTTP 500 and unusual content-type"

Expected: Back button is enabled.
Actual: Back button is not enabled.

I think we should create a separate bug to handle session history.

Comment 22

[Sean Kim [:sekim]]

Attachment: Bug 1325876 - Add a test for internal error page with error codes (4xx, 5xx) and Content-Length: 0 r=#necko

Comment 23

[Sean Kim [:sekim]]

Attachment: Bug 1325876 - Expose responseStatus and responseStatusText for about:neterror r=#necko,manuel

Comment 24

[Sean Kim [:sekim]]

Attachment: Bug 1325876 - Modify tests to verify exposed responseStatus and responseStatusText r=#necko

Comment 26

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/b5a8d01c0651
Show internal error page for HTTP responses with error codes (4xx, 5xx) and Content-Length: 0 instead of blank page r=manuel,necko-reviewers,fluent-reviewers,jesup,bolsson,valentin,kershaw
https://hg.mozilla.org/integration/autoland/rev/b3c3a5d3c6b9
Expose responseStatus and responseStatusText for about:neterror r=necko-reviewers,fluent-reviewers,bolsson,kershaw,webidl,saschanaz
https://hg.mozilla.org/integration/autoland/rev/4b1363236bcd
Add a test for internal error page with error codes (4xx, 5xx) and Content-Length: 0 r=necko-reviewers,jesup,kershaw

Comment 27

[Pulsebot]

Backout by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/49ba5bf01339
Backed out 3 changesets for causing multiple failures. CLOSED TREE

Comment 28

[agoloman]

Backed out for causing multiple failures.

• Backout link
• Push with failures all failures

————————————————————————————————————

• Push with failures wpt failures
• Failure Log

————————————————————————————————————

• Push with failures mochitest failures @test_http_background_auth_request.html
• Failure Log

————————————————————————————————————

• Push with failures bc failures
• Failure Log

Comment 29

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/c0454be735a8
Show internal error page for HTTP responses with error codes (4xx, 5xx) and Content-Length: 0 instead of blank page r=manuel,necko-reviewers,fluent-reviewers,jesup,bolsson,valentin,kershaw
https://hg.mozilla.org/integration/autoland/rev/1092114b16b5
Expose responseStatus and responseStatusText for about:neterror r=necko-reviewers,fluent-reviewers,bolsson,kershaw,webidl,saschanaz
https://hg.mozilla.org/integration/autoland/rev/10448d623bd4
Add a test for internal error page with error codes (4xx, 5xx) and Content-Length: 0 r=necko-reviewers,jesup,kershaw

Comment 30

[agoloman]

Backed out for causing multiple bc failures.

• Backout link
• Push with failures
• Failure Log

Comment 31

[Aaron Train [:aaronmt]]

This also breaks ui-test-arm jobs on CI in [1] verifyFirefoxSuggestHeaderForBrowsingDataSuggestionsTest and [2] verifyTabsSearchWithOpenTabsTest which load generic assets through a local web server (OK HTTP MockWebServer) in Fenix tests. The tabs are showing a 'Cannot Complete Request' error. Fenix is expected to show the name of the page (e.g, generic1.html) (loaded via local web server) instead of this error as a search suggestion.

[1] https://searchfox.org/mozilla-central/source/mobile/android/fenix/app/src/androidTest/java/org/mozilla/fenix/ui/SearchTest.kt#672
[2] https://searchfox.org/mozilla-central/source/mobile/android/fenix/app/src/androidTest/java/org/mozilla/fenix/ui/SearchTest.kt#758

• CI logs: https://treeherder.mozilla.org/logviewer?job_id=487147286&repo=autoland&lineNumber=3385
• Firebase Test Lab matrix link: https://console.firebase.google.com/project/moz-fenix/testlab/histories/bh.66b7091e15d53d45/matrices/6498108635348826662/details?stepId=bs.c6325c2449c71a89&testCaseId=4

Android APK: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/Wv5filItRUOz6kg9pVZ2dg/artifacts/public/build/target.arm64-v8a.apk
Android test APK: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/ODTdg2rpTpOky1hrLjGUGA/artifacts/public/build/target.noarch.apk
Reproducible locally via (after installing both APKs on device):

adb shell am instrument -w -e class org.mozilla.fenix.ui.SearchTest#verifyTabsSearchWithOpenTabsTest org.mozilla.fenix.debug.test/androidx.test.runner.AndroidJUnitRunner

Matrix has videos and screenshots

Comment 32

[Sean Kim [:sekim]]

Disabling below chunk of code strangely fixed the problem on Android (using #ifndef ANDROID).. Seems like generic1.html has a contentLength of 0 then returns responseCode above 400 (confirmed that rv did not fail). Also confirmed that the error passed here is ERROR_UNKNOWN on Fenix.

    if (NS_FAILED(rv) || contentLength <= 0) {
      if (responseCode >= 500) {
        return NS_ERROR_NET_ERROR_RESPONSE;
      }
      if (responseCode >= 400) {
        return NS_ERROR_NET_EMPTY_RESPONSE;
      }
    }

Responsible patch: https://phabricator.services.mozilla.com/D220193

Comment 33

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/0bda3da380ac
Show internal error page for HTTP responses with error codes (4xx, 5xx) and Content-Length: 0 instead of blank page r=manuel,necko-reviewers,fluent-reviewers,jesup,bolsson,valentin,kershaw
https://hg.mozilla.org/integration/autoland/rev/decca0d92c74
Expose responseStatus and responseStatusText for about:neterror r=necko-reviewers,fluent-reviewers,bolsson,kershaw,webidl,saschanaz
https://hg.mozilla.org/integration/autoland/rev/aede2883237a
Add a test for internal error page with error codes (4xx, 5xx) and Content-Length: 0 r=necko-reviewers,jesup,kershaw

Comment 34

[Norisz Fay [:noriszfay]]

Backed out for causing bc failures on browser_subdocument_downgrade.js

Backout link

Push with failures

Failure log

This failure is also caused by this bug.

Comment 35

[Sean Kim [:sekim]]

Attachment: Bug 1944358 - Set browser.http.blank_page_with_error_response.enabled pref to false by default r=#necko,maltejur

Comment 36

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/a4ecb980de87
Show internal error page for HTTP responses with error codes (4xx, 5xx) and Content-Length: 0 instead of blank page r=manuel,necko-reviewers,fluent-reviewers,jesup,bolsson,valentin,kershaw
https://hg.mozilla.org/integration/autoland/rev/55446c955dbd
Expose responseStatus and responseStatusText for about:neterror r=necko-reviewers,fluent-reviewers,bolsson,kershaw,webidl,saschanaz
https://hg.mozilla.org/integration/autoland/rev/8b56e7a25cdd
Add a test for internal error page with error codes (4xx, 5xx) and Content-Length: 0 r=necko-reviewers,jesup,kershaw
https://hg.mozilla.org/integration/autoland/rev/1c2e847c0632
Add a pref for showing error page for HTTP responses with error codes and Content-Length: 0 r=necko-reviewers,valentin

Comment 37

[Serban Stanca [:SerbanS]]

Backed out for causing multiple failures.

• Backout link

---

• Push with failures - wpt failures
• Failure Log
• Failure line: TEST-UNEXPECTED-ERROR | /loading/early-hints/404-with-early-hints.h2.window.html | Reached error page: about:neterror?e=httpErrorPage&u=https%3A//web-platform.test%3A9000/loading/early-hints/resources/404-with-early-hints.h2.py%3Fresource-url%3Dhttps%253A%252F%252Fweb-platform.test%253A9000%252Floading%252Fearly-hints%252Fresources%252Fsquare.png%253F7d1545a6-6f8c-4cf4-a293-a78777afebd8&c=UTF-8&d=https%3A//web-platform.test%3A9000/loading/early

---

• Push with failures - mochitests failures
• Failure Log
• Failure line: TEST-UNEXPECTED-FAIL | browser/components/tabbrowser/test/browser/tabs/browser_link_in_tab_title_and_url_prefilled_normal_page_blank_target.js | Test timed out -

---

• Push with failures - another wpt failures
• Failure Log
• Failure line: TEST-UNEXPECTED-TIMEOUT | /resource-timing/response-status-code.html | This test validates the response status of resources. 62 - Test timed out

Comment 38

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/014c9c77393a
Show internal error page for HTTP responses with error codes (4xx, 5xx) and Content-Length: 0 instead of blank page r=manuel,necko-reviewers,fluent-reviewers,jesup,bolsson,valentin,kershaw
https://hg.mozilla.org/integration/autoland/rev/41cc884f84c1
Expose responseStatus and responseStatusText for about:neterror r=necko-reviewers,fluent-reviewers,bolsson,kershaw,webidl,saschanaz
https://hg.mozilla.org/integration/autoland/rev/f8add6e2ef93
Add a test for internal error page with error codes (4xx, 5xx) and Content-Length: 0 r=necko-reviewers,jesup,kershaw
https://hg.mozilla.org/integration/autoland/rev/b29ef3f14bf7
Add a pref for showing error page for HTTP responses with error codes and Content-Length: 0 r=necko-reviewers,valentin,credential-management-reviewers,tabbrowser-reviewers,dimi,sthompson

Comment 39

[amarc]

Backed out for causing wd fails @ navigation.spec.js

• Backout link
• Push with failures
• Failure Log
• Failure line:

TEST-UNEXPECTED-FAIL | navigation Page.goto should not throw an error for a 404 response with an empty body (navigation.spec.js) | expected PASS
TEST-UNEXPECTED-FAIL | navigation Page.goto should not throw an error for a 500 response with an empty body (navigation.spec.js) | expected PASS 

Comment 40

[Andra Esanu (needinfo me)]

(In reply to Pulsebot from comment #27)

Backout by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/49ba5bf01339
Backed out 3 changesets for causing multiple failures. CLOSED TREE

Perfherder has detected a browsertime performance change from push 49ba5bf01339012469f5bb95aa4f1108f1dbe16f.

Improvements:

Ratio	Test	Platform	Options	Absolute values (old vs new)	Performance Profiles
4%	linkedin LastVisualChange	macosx1015-64-shippable-qr	cold fission webrender	2,306.98 -> 2,207.65	Before/After
3%	linkedin SpeedIndex	macosx1015-64-shippable-qr	fission warm webrender	604.34 -> 585.15	Before/After

Details of the alert can be found in the alert summary, including links to graphs and comparisons for each of the affected tests.

If you need the profiling jobs you can trigger them yourself from treeherder job view or ask a sheriff to do that for you.

You can run these tests on try with ./mach try perf --alert 43168

For more information on performance sheriffing please see our FAQ.

Comment 41

[Sean Kim [:sekim]]

Planning to first land the response status code patch since it is relatively less risky on regressions, then let's land the blank page patch here.

Comment 42

[Phabricator Automation]

Comment on attachment 9426931 [details]
Bug 1325876 - Expose responseStatus and responseStatusText for about:neterror r=#necko,manuel

Revision D223417 was moved to bug 1938947. Setting attachment 9426931 [details] to obsolete.

Comment 43

[Alex Finder]

(In reply to agoloman from comment #30)

Backed out for causing multiple bc failures.

• Backout link
• Push with failures
• Failure Log

Perfherder has detected a browsertime performance change from push 6712afee953a1d20ee8124093979a03daf112ca4.

Improvements:

Ratio	Test	Platform	Options	Absolute values (old vs new)	Performance Profiles
5%	linkedin PerceptualSpeedIndex	macosx1015-64-shippable-qr	fission warm webrender	514.02 -> 486.53	Before/After
4%	linkedin LastVisualChange	macosx1015-64-shippable-qr	fission warm webrender	1,637.37 -> 1,570.43	Before/After

Details of the alert can be found in the alert summary, including links to graphs and comparisons for each of the affected tests.

If you need the profiling jobs you can trigger them yourself from treeherder job view or ask a sheriff to do that for you.

You can run these tests on try with ./mach try perf --alert 43145

For more information on performance sheriffing please see our FAQ.

Comment 44

[Andra Esanu (needinfo me)]

(In reply to Pulsebot from comment #27)

Backout by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/49ba5bf01339
Backed out 3 changesets for causing multiple failures. CLOSED TREE

Perfherder has detected a mozperftest performance change from push 49ba5bf01339012469f5bb95aa4f1108f1dbe16f.

Improvements:

Ratio	Test	Platform	Options	Absolute values (old vs new)
5%	AndroidStartup:fenix cold_main_first_frame.median	android-hw-a55-14-0-aarch64-shippable		557.54 -> 528.54
5%	AndroidStartup:fenix cold_main_first_frame.median	android-hw-a55-14-0-aarch64-shippable		555.08 -> 526.83
4%	AndroidStartup:fenix cold_main_first_frame.mean	android-hw-a55-14-0-aarch64-shippable		551.16 -> 529.07

Details of the alert can be found in the alert summary, including links to graphs and comparisons for each of the affected tests.

If you need the profiling jobs you can trigger them yourself from treeherder job view or ask a sheriff to do that for you.

You can run these tests on try with ./mach try perf --alert 43216

For more information on performance sheriffing please see our FAQ.

Comment 45

[Alex Finder]

(In reply to Serban Stanca [:SerbanS] from comment #37)

Backed out for causing multiple failures.

• Backout link

---

• Push with failures - wpt failures
• Failure Log
• Failure line: TEST-UNEXPECTED-ERROR | /loading/early-hints/404-with-early-hints.h2.window.html | Reached error page: about:neterror?e=httpErrorPage&u=https%3A//web-platform.test%3A9000/loading/early-hints/resources/404-with-early-hints.h2.py%3Fresource-url%3Dhttps%253A%252F%252Fweb-platform.test%253A9000%252Floading%252Fearly-hints%252Fresources%252Fsquare.png%253F7d1545a6-6f8c-4cf4-a293-a78777afebd8&c=UTF-8&d=https%3A//web-platform.test%3A9000/loading/early

---

• Push with failures - mochitests failures
• Failure Log
• Failure line: TEST-UNEXPECTED-FAIL | browser/components/tabbrowser/test/browser/tabs/browser_link_in_tab_title_and_url_prefilled_normal_page_blank_target.js | Test timed out -

---

• Push with failures - another wpt failures
• Failure Log
• Failure line: TEST-UNEXPECTED-TIMEOUT | /resource-timing/response-status-code.html | This test validates the response status of resources. 62 - Test timed out

Perfherder has detected a awsy performance change from push e75215ee1dbf77b9f31e2627215a83dbc4e17137.

Regressions:

Ratio	Test	Platform	Options	Absolute values (old vs new)
3%	Images Tabs closed [+30s, forced GC]	windows11-64-2009-shippable-qr	fission tp6	1,630,515.33 -> 1,681,829.33

As author of one of the patches included in that push, we need your help to address this regression.
Details of the alert can be found in the alert summary, including links to graphs and comparisons for each of the affected tests. Please follow our guide to handling regression bugs and let us know your plans within 3 business days, or the patch(es) may be backed out in accordance with our regression policy.

If you need the profiling jobs you can trigger them yourself from treeherder job view or ask a sheriff to do that for you.

You can run these tests on try with ./mach try perf --alert 43247

For more information on performance sheriffing please see our FAQ.

Comment 46

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/42f5ed6abce5
Show internal error page for HTTP responses with error codes (4xx, 5xx) and Content-Length: 0 instead of blank page r=manuel,necko-reviewers,fluent-reviewers,jesup,bolsson,valentin,kershaw
https://hg.mozilla.org/integration/autoland/rev/ba3a433729a6
Add a test for internal error page with error codes (4xx, 5xx) and Content-Length: 0 r=necko-reviewers,jesup,kershaw
https://hg.mozilla.org/integration/autoland/rev/2dcbd388c95f
Add a pref for showing error page for HTTP responses with error codes and Content-Length: 0 r=necko-reviewers,valentin,credential-management-reviewers,tabbrowser-reviewers,dimi,sthompson

Comment 47

[Pulsebot]

Backout by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/2edc8f15bb01
Backed out 3 changesets for causing Linux related  puppeteer failures. CLOSED TREE

Comment 48

[Serban Stanca [:SerbanS]]

Backed out for causing Linux related puppeteer failures.

• Backout link
• Push with failures
• Failure Log
• Failure line: TEST-UNEXPECTED-FAIL | navigation Page.goto should not throw an error for a 404 response with an empty body (navigation.spec.js) | expected PASS

Comment 49

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/b7a277367308
Add a pref for showing internal error page for HTTP responses with error codes (4xx, 5xx) and Content-Length: 0 instead of blank page r=manuel,necko-reviewers,fluent-reviewers,jesup,bolsson,valentin,kershaw
https://hg.mozilla.org/integration/autoland/rev/38e555f37799
Add a test for internal error page with error codes (4xx, 5xx) and Content-Length: 0 r=necko-reviewers,jesup,kershaw

Comment 50

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/8f3190a3ce19
Lint fix. CLOSED TREE

Comment 51

[Cristian Tuns]

https://hg.mozilla.org/mozilla-central/rev/b7a277367308
https://hg.mozilla.org/mozilla-central/rev/38e555f37799
https://hg.mozilla.org/mozilla-central/rev/8f3190a3ce19

Comment 52

[Sean Kim [:sekim]]

The patch has been landed, however the error page becomes available only if browser.http.blank_page_with_error_response.enabled pref is set to false.

Comment 53

[Phabricator Automation]

Comment on attachment 9444240 [details]
Bug 1944358 - Set browser.http.blank_page_with_error_response.enabled pref to false by default r=#necko,maltejur

Revision D232429 was moved to bug 1944358. Setting attachment 9444240 [details] to obsolete.