1258161 - XML parsing error in omni.ja!/chrome/toolkit/content/global/netError.xhtml

Status: RESOLVED FIXED

(Toolkit :: Safe Browsing, defect)

Description

[benl]

User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:43.0) Gecko/20100101 Firefox/43.0 SeaMonkey/2.40
Build ID: 20160120202951

Steps to reproduce:

Started Seamonkey 2.44a2 and tried to navigate to any URL.

Also happens on initial attempt to show http://www.seamonkey-project.org/start/



Actual results:

No page is displayed, instead this is shown:

XML Parsing Error: undefined entity Location: jar:file:///C:/Users/[user name]/AppData/Local/SeaMonkeyTest/omni.ja!/chrome/toolkit/content/global/netError.xhtml Line Number 296, Column 36:
        <h1 id="et_malwareBlocked">&malwareBlocked.title;</h1>
-----------------------------------^


Expected results:

Page at URL should be displayed.

User agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:47.0) Gecko/20100101 Firefox/47.0 SeaMonkey/2.44a2

Build identifier: 20160309193908

Comment 1

[benl]

Ok, so this is not that bad. I hadn't properly set the firewall to let Seamonkey through.

But it's still a little bug.

So, for the steps to reproduce section:

Please replace "any URL" with "any unreachable URL"

and for the "expected results" section:

please delete "Page at URL should be displayed."

and replace with "'Failed to Connect' page should be displayed."

Sorry about that.

Comment 2

[K Chayka]

Windows 7, Seamonkey 2.45a1, build 20160308001946

I get the same error but only on some sites. I think I was supposed to get some kind of DNS error coming back to the browser, since other browsers (e.g. Chrome) also get an error on the same URLs but indicate a DNS problem of some sort. However, I think my particular error might actually have been caused by a blacklisted IP address.

Regardless, as benl said, the XML error should not occur and a more meaningful error message should display.

Comment 3

[Frank-Rainer Grahl (:frg)]

Confirmed using Adrians unofficial 2.44a2 Windows x64 en-US build

>> User agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:47.0) Gecko/20100101 Firefox/47.0 >> SeaMonkey/2.44a2
>> Build identifier: 20160309193908

and private 2.45a1 Windows x64 en-US VS2015 build

>> User agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:48.0) Gecko/20100101 Firefox/48.0 >> SeaMonkey/2.45a1
>> Build identifier: 20160325110550

You can force showing the error if just enter an invalid url like about:lalal

Does not occur in 2.43. 

I am unsure what changed in the backend. Seamonkey does not have a netError.dtd which together with another change might have triggered this.

Comment 5

[Philip Chee]

Attachment: Restore removed strings that are actually used.

These strings are actually used. hint DXR hint MXR hint hint nudge nudge.

Comment 6

[Frank-Rainer Grahl (:frg)]

Are the removed strings in Big 1245992 still needed elsewhere or could they be just removed from netError.xhtml? The Safebrowsing feature itself in Firefox and Seamonkey doesn't need them.

Comment 7

[Frank-Rainer Grahl (:frg)]

Oops make that Bug 1245992

Comment 8

[Philip Chee]

Comment on attachment 8734773 [details] [diff] [review]
Restore removed strings that are actually used.

(In reply to Frank-Rainer Grahl from comment #6)
> Are the removed strings in Bug 1245992 still needed elsewhere or could they
> be just removed from netError.xhtml? The Safebrowsing feature itself in
> Firefox and Seamonkey doesn't need them.
Good question! New patch comming up

Comment 9

[Philip Chee]

Attachment: Patch remove unused strings from docshell/resources/content/netError.xhtml

Building now.

Comment 10

[Frank-Rainer Grahl (:frg)]

Revised patch works for me. Did the same in parallel after writing the comment to check it out. 

>> VS2015 x64 en-US build
>> User agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:48.0) Gecko/20100101 Firefox/48.0 SeaMonkey/2.45a1

Comment 11

[Panos Astithas (he/him) [:past] (please ni?)]

Comment on attachment 8734824 [details] [diff] [review]
Patch remove unused strings from docshell/resources/content/netError.xhtml

Review of attachment 8734824 [details] [diff] [review]:
-----------------------------------------------------------------

Firefox doesn't use this file, so our tests didn't catch this. Thanks!

Comment 12

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/2a8ca5caae8e

Comment 13

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/2a8ca5caae8e

Comment 15

[Rainer Bielefeld]

REPRODUCIBLE with  English SeaMonkey 2.45a1  (Windows NT 6.1; WOW64; rv:48.0)  Gecko/20100101 Firefox/48.0 Build 20160308001946  (Default Classic Theme)  on German WIN7 64bit.

I think we urgently need some backports for SeaMonkey

Comment 16

[Frank-Rainer Grahl (:frg)]

Only m-a is still affected.

Comment 17

[Philip Chee]

Comment on attachment 8734824 [details] [diff] [review]
Patch remove unused strings from docshell/resources/content/netError.xhtml

Approval Request Comment
[Feature/regressing bug #]: Bug 1259922 

[User impact if declined]: Visiting any invalid URL will result in XML Parsing Error instead of the proper error page.

[Describe test coverage new/current, TreeHerder]: Baked on mozilla-central.

[Risks and why]: No risk. Fix bustage caused by Bug 1259922

[String/UUID change made/needed]: None.

Comment 18

[Philip Chee]

(In reply to Frank-Rainer Grahl from comment #16)
> Only m-a is still affected.
> status-firefox47: --- → affected
Firefox has completely forked netError.* files. Which is why they didn't realize that they were breaking other consumers in Bug 1259922.

Comment 19

[Ritu Kothari (:ritu) (Inactive, please n-i to RyanVM, jcristau, or pascal)]

Hi Panos, can I get a second opinion from you on whether this fix needs to be uplifted to Aurora 47 or should we let it just ride the trains? Thanks!

Comment 20

[Panos Astithas (he/him) [:past] (please ni?)]

I think it's pretty safe to uplift this patch. As comment 18 says, it won't affect Firefox at all.

Comment 21

[Liz Henry (:lizzard) (relman/hg->git project)]

Bug 1259922 landed in (thunderbird) 48. How would that cause a regression in Firefox 47?  Are our releases not in sync now?

Marking 46 as unaffected.

Comment 22

[Frank-Rainer Grahl (:frg)]

Firefox 47 itself is not affected. It does not use the file. Just the mozilla-aurora tree which is at 47 contains this bug and there is no flag for the tree itself. Seamonkey 2.44 in c-a is affected and needs this fix because it does still use netErrror.xhtml from m-a.

Comment 23

[Philip Chee]

1. mozilla-aurora contains code some of which is used by Firefox, and some of it isn't.
2. The confusion is that there are no tracking flags for gecko-47 only for firefox-47.
3. This uplift request is for Gecko code that lives in mozilla-* but isn't used by Firefox-*
4. This code *is* used by SeaMonkey however so we need to unbreak SeaMonkey by uplifting to mozilla-aurora.
5. Thunderbird is ... confusing. They have *partially* forked some of the netError.* files. I suggest you ignore what is happening to the Thunderbird part of the comm-* trees as far as this patch is conserned.

Comment 24

[Ritu Kothari (:ritu) (Inactive, please n-i to RyanVM, jcristau, or pascal)]

Comment on attachment 8734824 [details] [diff] [review]
Patch remove unused strings from docshell/resources/content/netError.xhtml

NPOTB (thanks everyone who helped clarify), Aurora47+

Comment 25

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/89d0d3749675100aab0c87782f94100dcb2563ae

Comment 27

[Rainer Bielefeld]

See "Bug 1283423 - Getting this again: XML Parsing Error: undefined entity Location: ... "

Fix incomplete? Not yet pulled to SM? Something else?