821498 - Pending Reports are not getting sent after fix for Bug 814078

Status: RESOLVED FIXED

(Firefox OS Graveyard :: General, defect)

Description

[Marcia Knous [:marcia]]

Seen while smoketesting yesterday using the 20121212 build.

I crashed several times during the course of the smoketest, and only one report submitted. I still have 5 reports in the "pending" folder. I tried two different fast wifi access points (Haxxor Room in MV + my home network). Even after connecting to these networks, the reports in pending queue were never submitted.

Comment 1

[Robert Kaiser]

Hub, can you look into this?
Nominating for bb because this is blocking getting good crash data.

Comment 2

[David Bolter [:davidb] (NeedInfo me for attention)]

(Hub's back from PTO tomorrow)

Comment 3

[Hubert Figuiere [:hub]]

Are these smoke test running on Unagi?

Comment 4

[Robert Kaiser]

In an ongoing email thread, Marcia said this:

> Example:
> 
> 1. Wifi off. Generate an OS crash (auto reboot).
> 2. Crash shows up up in pending directory
> 3. Turn wifi on
> 4. Crash stays in pending directory

I guess that's in line with the problem she reported here.
Also, I take it that the steps discussed in bug 814078 comment #33 and later will fix this issue here as well, so I asked to do this work in this bug.

Comment 5

[Hubert Figuiere [:hub]]

there was a bug where things didn't get sent. Also make sure that prefs.js in the profile has a line like this:

user_pref("app.reportCrashes", true);

It is set by the settings.

Comment 6

[Hubert Figuiere [:hub]]

Attachment: Bug 821498 - Make sure to always send the queued reports too.

This patch should fix a few cases I missed. Not sure if this fixes Marcia's case. But it sure make things better.

Comment 7

[Robert Kaiser]

(In reply to Hub Figuiere [:hub] from comment #5)
> user_pref("app.reportCrashes", true);

Well, that also should be set with the prompt that appears when you encounter your first crash and acknowledge to always send them. :)

Comment 8

[Marcia Knous [:marcia]]

Yes, all the smoketests run on Unagi and the only devices QA are focusing on ATM are Unagi.

I was able to repeat the STR in Comment 4 moments ago and the same thing happened with an OS crash. I believe Naoki said that is not the case with App crashes but we will have to confirm.

Y(In reply to Hub Figuiere [:hub] from comment #3)
> Are these smoke test running on Unagi?

Comment 9

[[:fabrice] Fabrice Desré]

Comment on attachment 694554 [details] [diff] [review]
Bug 821498 - Make sure to always send the queued reports too.

Review of attachment 694554 [details] [diff] [review]:
-----------------------------------------------------------------

r=me, but I'd like to know the answer to my question ;)

::: b2g/chrome/content/shell.js
@@ +148,4 @@
>        if (network.state == Ci.nsINetworkInterface.NETWORK_STATE_CONNECTED
>            && network.type == Ci.nsINetworkInterface.NETWORK_TYPE_WIFI) {
>          shell.CrashSubmit.submit(aCrashID);
> +        shell.submitQueuedCrashes();

Isn't the aCrashID crash in the pending queue at this point? If so, we can just use submitQueuedCrashes() here.

Comment 10

[Hubert Figuiere [:hub]]

Attachment: Bug 821498 - Make sure to always send the queued reports too.

Reworked patch follow fabrice question

Comment 11

[Hubert Figuiere [:hub]]

(In reply to Marcia Knous [:marcia] from comment #8)
> Yes, all the smoketests run on Unagi and the only devices QA are focusing on
> ATM are Unagi.

OK, I just wanted to make sure. Thanks.

Comment 12

[Hubert Figuiere [:hub]]

Comment on attachment 695487 [details] [diff] [review]
Bug 821498 - Make sure to always send the queued reports too.

Adding Vivien to the review in case Fabrice is on holiday. 

This new patch just answer Fabrice question and take proper action.

Comment 13

[:Margaret Leibovic]

Comment on attachment 695487 [details] [diff] [review]
Bug 821498 - Make sure to always send the queued reports too.

This patch addresses Fabrice's comment and looks good to me.

Comment 14

[Hubert Figuiere [:hub]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/422309d2a072

Comment 15

[Graeme McCutcheon [:graememcc]]

https://hg.mozilla.org/mozilla-central/rev/422309d2a072

Comment 16

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/32f9669a3063
https://hg.mozilla.org/releases/mozilla-b2g18/rev/af84ccfa189d