813469 - [E-Mail] Can't open an e-mail from the Inbox email list

Status: VERIFIED FIXED

(Firefox OS Graveyard :: Gaia::E-Mail, defect, P1)

Description

[David Scravaglieri [:scravag]]

gecko-aurora: 5f19747d3410
gaia: b3344d5021137b59e98855d2c01badc8c3bf27c4

Steps:

1 - Open E-mail app
2 - Go to Inbox (if needed)
3 - Tap a message to open it

Expected:
The message is shown

Actual:
Nothing is shown, nothing happens.

Comment 1

[Jim Porter (:squib)]

Do you have a logcat? My guess is that an error is getting thrown because of an unusual message.

Comment 2

[Andrew Sutherland [:asuth] (he/him)]

I am able to duplicate this with the most recent unagi nightly.  Unfortunately, identifying the actual problem is troublesome since unit tests for GELAM are not showing breakage on IMAP, b2g-desktop is hosed on my linux desktop, and the fancy log dumping to device storage is failing on a type error.

The 'of course' news is that I added a whole bunch of try handlers to https://github.com/mozilla-b2g/gaia-email-libs-and-more/pull/82 to deal with problems like this and make it easier for us to see them.  I'll merge all trunk state to my branches and see if I can shake out the problem, etc.

Comment 3

[Andrew Sutherland [:asuth] (he/him)]

Per my mail to dev-b2g, last night's unagi nightly appears to have been built from a post-uplift aurora.  It's possible there may be complications/breakage related to that which could explain why unit tests backed by mozilla-beta work fine.  I'll spin a mozilla-aurora xpcshell and see if it shows any badness.

Comment 4

[Andrew Sutherland [:asuth] (he/him)]

The aurora xpcshell runs were clean after performing a minor for related to how JSM's work (we need to do this.BLAH to export BLAH rather than having it sucked out of the global scope which is now instead a function/object scope.)

Comment 5

[David Scravaglieri [:scravag]]

Works now on

gecko-aurora: 2f801d18884d
gaia: 65852cb879a3952ba6f6e4184a1444c147e3c90d

Comment 6

[Dylan Oliver [:doliver]]

Can this bug be resolved then?

Comment 7

[David Scravaglieri [:scravag]]

Doesn't work anymore today

gecko-beta: d82726e5866f
gaia: 6f2e22fda2b6a0b76f97cdfcb829b350c920af8a

Comment 8

[Andrew Sutherland [:asuth] (he/him)]

(In reply to dscravaglieri from comment #7)
> Doesn't work anymore today

Do you have a logcat from the run?  I landed an improvement to the logging for these type of failures last night, so there should have been some log output.

Also, we haven't actually reset the database since 11/16.  Have you reset your profile or otherwise reset the mail database since you encountered the problem on the messed-up nightly on 11/20?

Comment 9

[David Scravaglieri [:scravag]]

Attachment: adb logcat

The scenario is:
Open email
Inbox list
Tap first mail to open it

Expected
See email body

Actual
Nothing

Comment 10

[David Scravaglieri [:scravag]]

And Yes, I have reset my profile and I test each days on fresh gecko-beta.

Comment 11

[Andrew Sutherland [:asuth] (he/him)]

The loop subscripting variable added by the linkification code has the same name as the outer subscripting variable.  Whether the bug triggers is a function of how many content blocks were discovered during quote chewing and how many links were discovered in the text.  A message with many links in it and with few quoted blocks or signature blocks is very likely to cause the bug to trigger.

Back-end unit tests don't cover this because the bug only exists in the front-end.  The integration test would need the IMAP fake-server implemented to do this since ActiveSync favors HTML.  I'll see how hard it is to spin that up.

Comment 12

[Andrew Sutherland [:asuth] (he/him)]

Fix landed on gaia/master:
https://github.com/mozilla-b2g/gaia/pull/6699

Comment 13

[Nick Davidson]

Issue is fixed - verified on Unagi build 20121231070201