935782 - [B2G][SMS] Unable to send MMS.

Status: RESOLVED WORKSFORME

(Firefox OS Graveyard :: Gaia::SMS, defect)

Description

[Lucas Aliaga]

Attachment: sendingMMS.txt

Unable to send any file over MMS including: Gallery photos, videos, music, camera photos. The message has a loading wheel that spins for several minutes until it fails.

Repro Steps:
1. Update Buri device to 1.3 buildID: 20131106040203.
2. Tap SMS > New contact and add a real contact.
3. Tap the Attachment button at the top of the screen.
4. Attach any short music/video file or picture.
5. Tap send.

Actual Results:
Message does not send.

Expected Results:
MMS to send to the recipient in a timely manner.

Repro Rate: 5/5 100%

Environmental Variables:
Builds ID: 20131106040203
Gecko: 9ba3faa35c96
Gaia: 3b5fe429f2414f2a9d7241b311b399033bb10612
Base Image: 20131015

See attached: Logcat "sendingMMS.txt"

Comment 1

[Julien Wajsberg [:julienw]]

bug 930882 was backed out not long before this build, and was causing this problem. Could it be it?

Comment 2

[Zac C (:zac)]

Still failing today unfortunately on build:

Device Hamachi
Gecko  http://hg.mozilla.org/mozilla-central/rev/70de5e24d79b
Gaia  42bbe26a72e030faf07a6fc297f61a3a8ccda25b
BuildID 20131107040200
Version 28.0a1

Comment 3

[David Scravaglieri [:scravag]]

Julien, could you investigate this deeper ?

Comment 4

[Montoya Clemmons (MClemmons)]

The last working 1.3 build for this issue occurs with below environmental variables:

Environmental Variables:
Works for 1.3 Build 20131101040203
Device: example: Buri v1.3 Mozilla RIL
BuildID: 20131101040203
Gaia: ccdf357ea150fc7d8b8a4b74c7adf31e7a57e465
Gecko: abe6790a5dd8
Version: 28.0a1

The MMS feature failed for this issue with the below 1.3 environmental variables but in a different fashion than the issue reported in Comment 0. Instead, the behavior is the attachment is removed and the To: field phone number is removed and the message does not send. This behavior was reproducible from the range of Environmental Variables from 11-2 through 11-4

Environmental Variables:
 (first failed in this fashion – message/attachment vanishes and does not send)
Device: Buri v1.3  Mozilla RIL
BuildID: 20131102080838
Gaia: 00cd7534afb8d727531c7a939e54e382b3be713a
Gecko: d8fd5706493e
Version: 28.0a1

The following 1.3 build first shows the behavior reported in Comment 0 (The message has a loading wheel that spins for several minutes until it fails)

Environmental Variables:
Device: Buri v1.3 Mozilla RIL
BuildID: 20131105040208
Gaia: 5c2929cd15ac020f85026030387aed9eb322362d
Gecko: 770de5942471
Version: 28.0a1

I was not able to reproduce this issue for Build 1.2
 
Environmental Variables:
Device: Buri v1.2 Mozilla RIL
BuildID: 20131107004003
Gaia: 590eb598aacf1e2136b2b6aca5c3124557a365ca
Gecko: 26f1e160e696
Version: 26.0

Comment 5

[Julien Wajsberg [:julienw]]

Taking for next sprint.

Comment 6

[Jason Smith [:jsmith]]

This feels like a gecko regression more likely than a Gaia regression. There is one commit in the regression range timeframe related to MMS (bug 923023), but I don't think it's likely the cause of this. Maybe this is a RIL regression?

Comment 7

[Jason Smith [:jsmith]]

Very likely caused by bug 854326.

Comment 8

[Gene Lian (I already quit Mozilla)]

(In reply to Jason Smith [:jsmith] from comment #7)
> Very likely caused by bug 854326.

I think it's too early to say so. I did carefully test all the functions before landing. Needless to say we have other patches landed after bug 854326 [1], which are supposed to be tested before landing as well.

In this week, we did encounter a Gaia regression that would make SMS fail to send. That is, the SMIL library cannot be recognized and loaded. Steve knows what's happening. I'm not sure if he's already opened/solved that bug or not.

[1] http://hg.mozilla.org/mozilla-central/filelog/a07aebef20e7/dom/mobilemessage/src/gonk/MmsService.js

Comment 9

[Gene Lian (I already quit Mozilla)]

> In this week, we did encounter a Gaia regression that would make SMS fail to
                                                                   ^^^
Sorry, I mean MMS.

> send. That is, the SMIL library cannot be recognized and loaded. Steve knows
> what's happening. I'm not sure if he's already opened/solved that bug or not.

Comment 10

[Gene Lian (I already quit Mozilla)]

I just manually checked out the latest codes form central/master. I can successfully send out MMS. Suspecting it's a Gaia regression mentioned at comment #8, which has been solved as well. Steve?

Gecko
----------
changeset:   154188:a07aebef20e7
tag:         tip
parent:      154164:4d7480679baa
parent:      154187:92a85e196ba3
user:        Ryan VanderMeulen <[e-mail]>
date:        Fri Nov 08 14:53:47 2013 -0500
summary:     Merge b-i to m-c.

Gaia
----------
commit f86f7f5c64eb6fd01ab17f721b542e17ed1c9b7e
Merge: 6a4f159 ff6f518
Author: Kevin Grandon <[e-mail]>
Date:   Fri Nov 8 17:15:33 2013 -0800

Comment 11

[Gene Lian (I already quit Mozilla)]

It's not a regression caused by bug 854326. Please let me know if you disagree. Thanks!

Comment 12

[Gene Lian (I already quit Mozilla)]

Which SIM do QA use? I'm wondering it's SIM dependent... If the regression is really due to bug 854326, there might be something wrong with the MMS APN settings reading. Just randomly guessing. Can you guys turn on the MMS debugging?

Sadly, Ctai and I are going to have business trips. I'm not available until next Wednesday. I'd highly appreciate if Vicamo or someone could take over for a while...

Comment 13

[Tony Chung [:tchung]]

(In reply to Gene Lian [:gene] (needinfo? encouraged) from comment #12)
> Which SIM do QA use? I'm wondering it's SIM dependent... If the regression
> is really due to bug 854326, there might be something wrong with the MMS APN
> settings reading. Just randomly guessing. Can you guys turn on the MMS
> debugging?

My guess is the reporter is using a US AT&T SIM.   mclemmons, please reproduce this again with the MMS debugging log enabled and capture the logcat into this bug.

> 
> Sadly, Ctai and I are going to have business trips. I'm not available until
> next Wednesday. I'd highly appreciate if Vicamo or someone could take over
> for a while...

Yes, please.  we cannot wait until wednesday for this bug to be looked at again.

Comment 14

[Julien Wajsberg [:julienw]]

I'm building latest central to test with a French SIM as an additional data point.

Comment 15

[Julien Wajsberg [:julienw]]

Gene, I just had an idea: is it possible that the gecko changes needs to be done in the commercial RIL too ?

Comment 16

[Julien Wajsberg [:julienw]]

I could send a MMS with latest gecko/latest gaia using a self-built gecko on a geeksphone peak with a french SIM from Orange.

I'm more and more thinking this is a comm RIL issue.

Comment 17

[Gene Lian (I already quit Mozilla)]

(In reply to Julien Wajsberg [:julienw] from comment #16)
> I could send a MMS with latest gecko/latest gaia using a self-built gecko on
> a geeksphone peak with a french SIM from Orange.
> 
> I'm more and more thinking this is a comm RIL issue.

We did have a couple of changes for:

1. RadioInterfaceLayer.js [1]

Bug 854326 - B2G Multi-SIM: support multiple SIM cards for SMS/MMS (part 5-2, MMS APN settings).
Bug 854326 - B2G Multi-SIM: support multiple SIM cards for SMS/MMS (part 5-1, clean up the APN setting logic).
Bug 854326 - B2G Multi-SIM: support multiple SIM cards for SMS/MMS (part 4, handle service ID).
Bug 854326 - B2G Multi-SIM: support multiple SIM cards for SMS/MMS (part 3, add ICC ID).
Bug 854326 - B2G Multi-SIM: support multiple SIM cards for SMS/MMS (part 2, IPDL for IPC).

2. nsIRadioInterfaceLayer.idl [2]

Bug 854326 - B2G Multi-SIM: support multiple SIM cards for SMS/MMS (part 5-2, MMS APN settings).
Bug 854326 - B2G Multi-SIM: support multiple SIM cards for SMS/MMS (part 4, handle service ID).


Obviously, COMM RIL needs the corresponding changes to make its RIL.js/RIL.idl compatible with MMS. According to the previous discussions and agreements with the vendor at Norway's work week, what we only need to do is to mark the component as RIL and the vendor will try to align our changes as possible as they can.

Is comment #0 reported by the COMM RIL? If yes, then it's a COMM RIL issue that needs to be fixed for sure.

Hi Vicamo, please correct me if I'm wrong. Thanks!

[1] http://hg.mozilla.org/mozilla-central/filelog/16949049f03d/dom/system/gonk/RadioInterfaceLayer.js
[2] http://hg.mozilla.org/mozilla-central/filelog/16949049f03d/dom/system/gonk/nsIRadioInterfaceLayer.idl

Comment 18

[Gene Lian (I already quit Mozilla)]

(In reply to Julien Wajsberg [:julienw] from comment #16)
> I could send a MMS with latest gecko/latest gaia using a self-built gecko on
> a geeksphone peak with a french SIM from Orange.

Thanks for the testing. It works for me by the Taiwan's local SIM. Waiting to see the MMS debugging results by US AT&T SIM to ensure it's a SIM dependent issue or not. Also, please use Mozilla RIL to test. Thanks!

Comment 19

[Jason Smith [:jsmith]]

Marking QA Wanted for comment 18.

Comment 20

[Vicamo Yang [:vicamo][:vyang]]

(In reply to Gene Lian [:gene] (needinfo? encouraged) from comment #17)
> (In reply to Julien Wajsberg [:julienw] from comment #16)
> Obviously, COMM RIL needs the corresponding changes to make its
> RIL.js/RIL.idl compatible with MMS. According to the previous discussions
> and agreements with the vendor at Norway's work week, what we only need to
> do is to mark the component as RIL and the vendor will try to align our
> changes as possible as they can.

Yes, unless someone open sources COMM RIL and gets it merged into Mozilla's code base, I don't know why that term keeps popping up on Mozilla's Bugzilla.  We move RIL bugs to RIL components because they should have been there, not asked by anyone else.

Comment 21

[Steve Chung [:steveck]]

(In reply to Gene Lian [:gene] (needinfo? encouraged) from comment #10)
> I just manually checked out the latest codes form central/master. I can
> successfully send out MMS. Suspecting it's a Gaia regression mentioned at
> comment #8, which has been solved as well. Steve?
> 
I could send mms in 11/10 master build. The regression you metioned should be bug 930882 and it's already backout and fixed.

Comment 22

[Zac C (:zac)]

(In reply to Gene Lian [:gene] (needinfo? encouraged) from comment #18)
> Thanks for the testing. It works for me by the Taiwan's local SIM. Waiting
> to see the MMS debugging results by US AT&T SIM to ensure it's a SIM
> dependent issue or not. Also, please use Mozilla RIL to test. Thanks!

Hey Gene, some of the devices this bug was found on run with AT&T, some run on T-Mo. I have to check over the results and see whethere there is correlation.

Comment 23

[Zac C (:zac)]

Just checked, on the device that initiated this bug it was a T-Mobile USA Sim.

The automated test has been passing all weekend and locally to me too so I cannot see any reason not to re-enable it.

I'm not 100% sure what the regression or backout was so can I leave this bug to someone else to close out?

Comment 24

[Jason Smith [:jsmith]]

Since you were able to reproduce original bug Zac, I'm going use comment 23 indicate that we have proof we can't reproduce this anymore. I'll close it as works for me. If someone reproduces this again today, feel free to reopen.

Comment 25

[Anshul]

(In reply to Vicamo Yang [:vicamo][:vyang] from comment #20)
> (In reply to Gene Lian [:gene] (needinfo? encouraged) from comment #17)
> > (In reply to Julien Wajsberg [:julienw] from comment #16)
> > Obviously, COMM RIL needs the corresponding changes to make its
> > RIL.js/RIL.idl compatible with MMS. According to the previous discussions
> > and agreements with the vendor at Norway's work week, what we only need to
> > do is to mark the component as RIL and the vendor will try to align our
> > changes as possible as they can.
> 
> Yes, unless someone open sources COMM RIL and gets it merged into Mozilla's
> code base, I don't know why that term keeps popping up on Mozilla's
> Bugzilla.  We move RIL bugs to RIL components because they should have been
> there, not asked by anyone else.
Please have the vendor file an SR to deal with issues related to Comm RIL. We have been trying to educate everyone on this since quite some time already.

Comment 26

[Montoya Clemmons (MClemmons)]

In response to Comment 13

 Tony Chung [:tchung] 2013-11-09 11:16:57 PST

My guess is the reporter is using a US AT&T SIM.   mclemmons, please reproduce this again with the MMS debugging log enabled and capture the logcat into this bug.

I am still trying to get the MMS debugging log enabled feature on my device to provide this information. I am working with others here trying to get this feature enabled. As of today 11-12-2013, I am still able to reproduce the issue consistently from the STR in comment 0.

Comment 27

[Jason Smith [:jsmith]]

(In reply to mclemmons from comment #26)
> In response to Comment 13
> 
>  Tony Chung [:tchung] 2013-11-09 11:16:57 PST
> 
> My guess is the reporter is using a US AT&T SIM.   mclemmons, please
> reproduce this again with the MMS debugging log enabled and capture the
> logcat into this bug.
> 
> I am still trying to get the MMS debugging log enabled feature on my device
> to provide this information. I am working with others here trying to get
> this feature enabled. As of today 11-12-2013, I am still able to reproduce
> the issue consistently from the STR in comment 0.

I think there's a separate bug filed for what you might be seeing - see bug 937894.

Comment 28

[Montoya Clemmons (MClemmons)]

In response to Comment 27.

Thank you. I was able to flash the device to the build listed in Comment 0 and use the port value in step 6 of STR in https://bugzilla.mozilla.org/show_bug.cgi?id=937894 and this corrected the problem. I was able to send an MMS and see it being received on the other device every time.