1001195 - FxAccountStatusFragment: crash in java.lang.Error: Error: could not match input at org.json.simple.parser.Yylex.yylex(Yylex.java)

Status: RESOLVED FIXED

(Firefox for Android Graveyard :: Android Sync, defect)

Description

[Richard Newman [:rnewman]]

This bug was filed from the Socorro interface and is 
report bp-b5271ca5-b677-4d4b-918a-5931b2140418.
=============================================================

https://crash-stats.mozilla.com/report/index/b5271ca5-b677-4d4b-918a-5931b2140418

java.lang.Error: Error: could not match input
	at org.json.simple.parser.Yylex.yylex(Yylex.java:681)
	at org.json.simple.parser.JSONParser.parse$66bc622e(JSONParser.java:118)
	at org.mozilla.gecko.sync.ExtendedJSONObject.<init>(ExtendedJSONObject.java:172)
	at org.mozilla.gecko.sync.ExtendedJSONObject.<init>(ExtendedJSONObject.java:181)
	at org.mozilla.gecko.sync.ExtendedJSONObject.parseJSONObject(ExtendedJSONObject.java:142)
	at org.mozilla.gecko.sync.SyncConfiguration.getEngineNamesFromPref(SyncConfiguration.java:331)
	at org.mozilla.gecko.fxa.activities.FxAccountStatusFragment.updateSelectedEngines(FxAccountStatusFragment.java:340)
	at org.mozilla.gecko.fxa.activities.FxAccountStatusFragment.refresh(FxAccountStatusFragment.java:316)
	at org.mozilla.gecko.fxa.activities.FxAccountStatusFragment.refresh(FxAccountStatusFragment.java:262)
	at org.mozilla.gecko.fxa.activities.FxAccountStatusActivity.onResume(FxAccountStatusActivity.java:82)

Comment 1

[Mark Finkle (:mfinkle) (use needinfo?)]

This the #9 crasher on release (Fx42)

Comment 2

[Nick Alexander :nalexander [he/him]]

(In reply to Mark Finkle (:mfinkle) from comment #1)
> This the #9 crasher on release (Fx42)

I just looked through the crash reports.  Almost all of these crashes are with MediaTek 67xx series 64-bit SoCs.  Luckily, the MT6752 device I ordered should arrive tomorrow (http://www.gsmarena.com/meizu_m1_note-6890.php).  Hopefully, we can get to the bottom of this!

Comment 3

[:Margaret Leibovic]

(In reply to Nick Alexander :nalexander from comment #2)
> (In reply to Mark Finkle (:mfinkle) from comment #1)
> > This the #9 crasher on release (Fx42)
> 
> I just looked through the crash reports.  Almost all of these crashes are
> with MediaTek 67xx series 64-bit SoCs.  Luckily, the MT6752 device I ordered
> should arrive tomorrow (http://www.gsmarena.com/meizu_m1_note-6890.php). 
> Hopefully, we can get to the bottom of this!

Have you had the chance to look at this? It's #12 on release, #15 on beta, so not the biggest crash, but would be good to address.

Comment 4

[:Margaret Leibovic]

This is now #8 on the beta 9 crash list.

Comment 5

[Nick Alexander :nalexander [he/him]]

(In reply to :Margaret Leibovic from comment #4)
> This is now #8 on the beta 9 crash list.

I don't have the crashstats fu to verify this, but this crash while hit the same folks repeatedly until they stop using Sync (and probably stop using Firefox).  So the impact is large but on a limited set of people.

We have no way to message "catastrophic failure in Sync" because all the places that can react to that (and there are places built to react to that!) *are* the places that are catastrophically failing.

I have patches in flight to remove Old Sync, which will reduce the consumers of the offending code; and am about to dig back into Bug 1204559, which should address the issue once and for all.  Unfortunately, the scope of changes is very large and we really shouldn't expect to uplift it past Aurora.

Comment 6

[:Margaret Leibovic]

(In reply to Nick Alexander :nalexander from comment #5)
> (In reply to :Margaret Leibovic from comment #4)
> > This is now #8 on the beta 9 crash list.
> 
> I don't have the crashstats fu to verify this, but this crash while hit the
> same folks repeatedly until they stop using Sync (and probably stop using
> Firefox).  So the impact is large but on a limited set of people.

kbrosnan, can you help us figure out if these crashes are coming mostly from the same users?

This could be a good point to raise as a reason why our release crash rates might differ from beta, if this particular crash is somehow affecting a small set of users who a more likely to use beta than release.

Comment 7

[Kevin Brosnan [Ex-Mozilla]]

Around 1.5 crashes per install in 43. Here is the data from 28 days from the crashes per install section of the crash reports page.

Product         Ver     Crashes Installations
FennecAndroid   43.0    27489   17973
FennecAndroid   42.0.2  2635    2012
FennecAndroid   44.0b9  130     57

For our top two crashes the ratio is 1.0 and 1.1 crash per install for comparison (android.database.CursorWindowAllocationException: Cursor... and android.database.sqlite.SQLiteClosable...).

Comment 10

[:Grisha Kruglov]

Looking at the crash signature, there have been no reports of this past 51.0.3. Something must have fixed it, perhaps. Closing the bug for now.