842418 - org.json.JSONException: Value null at minZoom of type org.json.JSONObject$1 cannot be converted to double @ ZoomConstraints.java:27

Status: RESOLVED FIXED

(Firefox for Android Graveyard :: Toolbar, defect)

Description

[Aaron Train [:aaronmt]]

02-18 16:24:17.382 E/GeckoApp( 8616): org.json.JSONException: Value null at minZoom of type org.json.JSONObject$1 cannot be converted to double
02-18 16:24:17.382 E/GeckoApp( 8616): 	at org.json.JSON.typeMismatch(JSON.java:96)
02-18 16:24:17.382 E/GeckoApp( 8616): 	at org.json.JSONObject.getDouble(JSONObject.java:412)
02-18 16:24:17.382 E/GeckoApp( 8616): 	at org.mozilla.gecko.ZoomConstraints.<init>(ZoomConstraints.java:27)
02-18 16:24:17.382 E/GeckoApp( 8616): 	at org.mozilla.gecko.GeckoApp.handleMessage(GeckoApp.java:860)
02-18 16:24:17.382 E/GeckoApp( 8616): 	at org.mozilla.gecko.BrowserApp.handleMessage(BrowserApp.java:491)
02-18 16:24:17.382 E/GeckoApp( 8616): 	at org.mozilla.gecko.util.EventDispatcher.dispatchEvent(EventDispatcher.java:75)
02-18 16:24:17.382 E/GeckoApp( 8616): 	at org.mozilla.gecko.GeckoAppShell.handleGeckoMessage(GeckoAppShell.java:1811)
02-18 16:24:17.382 E/GeckoApp( 8616): 	at org.mozilla.gecko.mozglue.GeckoLoader.nativeRun(Native Method)
02-18 16:24:17.382 E/GeckoApp( 8616): 	at org.mozilla.gecko.mozglue.GeckoLoader.nativeRun(Native Method)
02-18 16:24:17.382 E/GeckoApp( 8616): 	at org.mozilla.gecko.GeckoAppShell.runGecko(GeckoAppShell.java:305)
02-18 16:24:17.382 E/GeckoApp( 8616): 	at org.mozilla.gecko.GeckoThread.run(GeckoThread.java:113)

https://tbpl.mozilla.org/php/getParsedLog.php?id=19858238&tree=Mozilla-Inbound

Comment 1

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

My guess is this is a regression from bug 840722 and might have something to do with the prototype setting the values to null:
http://mxr.mozilla.org/mozilla-central/source/mobile/android/chrome/content/browser.js#5612

Should we pick better defaults? Maybe use Nan, which we use in other places and might be better for JSON:
http://mxr.mozilla.org/mozilla-central/source/mobile/android/chrome/content/browser.js#3777

Comment 2

[Tetsuharu OHZEKI [:tetsuharu] (UTC+9)]

The property of ViewportMetadata instance always gets any values which does not inherit from ViewportMetadata.prototype.
http://hg.mozilla.org/mozilla-central/file/e8f8a3f6f1f6/mobile/android/chrome/content/browser.js#l5599

However, the constructor don't check parameter's member values. If aMetadata.minZoom = "bar" is passed, the constructor will set "bar" to his instance's |minZoom| as string.

I agree that we should specify a value which is more suitable type for ViewportMetadata.prototype. But I think also we need a stack trace information in details for resolving the root of this bug.

Comment 3

[Kartikaya Gupta (email:[e-mail])]

From a code inspection I don't see any code path that could result in this. If nobody else can see one either I think we should back out bug 840722 until we can figure it out.

Comment 4

[Tetsuharu OHZEKI [:tetsuharu] (UTC+9)]

Attachment: specify value types

Backout bug 840722, or specify a value type.

Comment 6

[Kartikaya Gupta (email:[e-mail])]

(In reply to Tetsuharu OHZEKI [:saneyuki_s] from comment #4)
> Created attachment 715592 [details] [diff] [review]
> specify value types
> 
> Backout bug 840722, or specify a value type.

I'm leery of making a change like this without properly understanding the codepath that is causing the problem. I'll back out bug 840722.

Comment 7

[Kartikaya Gupta (email:[e-mail])]

Fixed by backout of bug 840722: https://hg.mozilla.org/integration/mozilla-inbound/rev/7210fcb55aae

Comment 8

[Tetsuharu OHZEKI [:tetsuharu] (UTC+9)]

(In reply to Kartikaya Gupta (email:[e-mail]) from comment #6)
> 
> I'm leery of making a change like this without properly understanding the
> codepath that is causing the problem. I'll back out bug 840722.

OK. I understand.

Comment 9

[Tetsuharu OHZEKI [:tetsuharu] (UTC+9)]

This is the result of attachment 715592 [details] [diff] [review] on try:
(I wrote for future reference)
https://tbpl.mozilla.org/?tree=Try&rev=08c78b02409c

Comment 10

[Kartikaya Gupta (email:[e-mail])]

The full log for the "rc" test on that try build still has the same exceptions, so the patch doesn't fix the underlying problem.

Comment 11

[Tetsuharu OHZEKI [:tetsuharu] (UTC+9)]

I fully agree the backout of bug 840722.

Comment 12

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/7210fcb55aae

Comment 13

[Scoobidiver (away)]

Push to Aurora:
https://hg.mozilla.org/releases/mozilla-aurora/rev/60d687faf510