1004988 - [Homescreen] several issues in the icon placement

Status: RESOLVED DUPLICATE of bug 1008161

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

Description

[Julien Wajsberg [:julienw]]

Attachment: first attachment

I use a Peak on v1.4, with the image images-peak-v1.4-2014-04-29.Gecko-44787e8.Gaia-01b7794.zip.

In case it's useful, this Peak was initially migrated from v1.3.

I'll focus on the issues that are not likely Peak-related.

STR1:
* install at least 12 applications/bookmarks

Exected:
* they all fit on the first page

Actual:
* only 11 icons (in addition to the 4 icons from e.me) fit on the first page, the 12th icons is installed on the second page
* If I try to move an icon back on the first page to fill the void, then one icon is moved to the second page.
* see first attachment

STR2:
* install enough applications to fill the second page too
* install one more application

Expected:
* the icon is installed on a newly opened 3rd page

Actual:
* no icon is displayed for the last application
* still in the Marketplace the icon is "launch" instead of "free" or "install". And I can launch it.
* I moved an icon to a 3rd page, and then the last application icon appeared on the 2nd page.

Note that one of the application was installed through the App Manager, I don't know if that's relevant. I also didn't try to install yet one more application.

Could be 2 different bugs, so please clone the bug if necessary.

Comment 1

[Cristian Rodriguez (:crdlc)]

Attachment: 1004988.patch

Julien, could you test this patch? thanks a lot

Comment 2

[Julien Wajsberg [:julienw]]

Comment on attachment 8416424 [details] [diff] [review]
1004988.patch

After applying the patch, the icons in the last line of the first page moved. But then the problem reappeared when I added 3 new icons, only 2 were shown, and the last one went to the next page.

Also, when the second page was full (again, because the last line moved too), I installed an app from the Marketplace, and the second issue was still there too (no icon for the installed app). So I'm confident this patch doesn't change anything for my issues, sorry.

Also it's very easy to make the Homescreen crash while moving an icon from page 2 to page 1. I have errors like this in logcat:

05-06 17:15:05.501  4002  4002 E msm7627a.gralloc: gralloc_register_buffer: gralloc_map failed
05-06 17:15:05.501  4002  4002 W GraphicBufferMapper: registerBuffer(0x4419eba0) failed -12 (Out of memory)
05-06 17:15:05.501  4002  4002 E memalloc: /dev/pmem: Failed to map buffer size:36671488 offset:36634624 fd:367 Error: Out of memory
05-06 17:15:05.501  4002  4002 E msm7627a.gralloc: Could not mmap handle 0x4419eba0, fd=367 (Out of memory)
05-06 17:15:05.501  4002  4002 E libgenlock: perform_lock_unlock_operation: GENLOCK_IOC_LOCK failed (lockType0x1, err=Bad file number fd=367)
05-06 17:15:05.501  4002  4002 E msm7627a.gralloc: gralloc_lock: genlock_lock_buffer (lockType=0x2) failed
05-06 17:15:05.501  4002  4002 E memalloc: /dev/pmem: Failed to map buffer size:36671488 offset:36634624 fd:367 Error: Out of memory
05-06 17:15:05.501  4002  4002 E msm7627a.gralloc: Could not mmap handle 0x4419eba0, fd=367 (Out of memory)
05-06 17:15:05.501  4002  4002 E libgenlock: perform_lock_unlock_operation: GENLOCK_IOC_LOCK failed (lockType0x1, err=Bad file number fd=367)
05-06 17:15:05.501  4002  4002 E msm7627a.gralloc: gralloc_lock: genlock_lock_buffer (lockType=0x2) failed
05-06 17:15:05.501  4002  4002 E memalloc: /dev/pmem: Failed to map buffer size:36671488 offset:36634624 fd:367 Error: Out of memory
05-06 17:15:05.501  4002  4002 E msm7627a.gralloc: Could not mmap handle 0x4419eba0, fd=367 (Out of memory)
05-06 17:15:05.501  4002  4002 E libgenlock: perform_lock_unlock_operation: GENLOCK_IOC_LOCK failed (lockType0x1, err=Bad file number fd=367)
05-06 17:15:05.501  4002  4002 E msm7627a.gralloc: gralloc_lock: genlock_lock_buffer (lockType=0x2) failed
05-06 17:15:05.501  4002  4002 E memalloc: /dev/pmem: Failed to map buffer size:36671488 offset:36634624 fd:367 Error: Out of memory
05-06 17:15:05.501  4002  4002 E msm7627a.gralloc: Could not mmap handle 0x4419eba0, fd=367 (Out of memory)
05-06 17:15:05.501  4002  4002 E libgenlock: perform_lock_unlock_operation: GENLOCK_IOC_LOCK failed (lockType0x1, err=Bad file number fd=367)
05-06 17:15:05.501  4002  4002 E msm7627a.gralloc: gralloc_lock: genlock_lock_buffer (lockType=0x2) failed
05-06 17:15:05.501  4002  4002 E memalloc: /dev/pmem: Failed to map buffer size:36671488 offset:36634624 fd:367 Error: Out of memory
05-06 17:15:05.501  4002  4002 E msm7627a.gralloc: Could not mmap handle 0x4419eba0, fd=367 (Out of memory)
05-06 17:15:05.501  4002  4002 E libgenlock: perform_lock_unlock_operation: GENLOCK_IOC_LOCK failed (lockType0x1, err=Bad file number fd=367)
05-06 17:15:05.501  4002  4002 E msm7627a.gralloc: gralloc_lock: genlock_lock_buffer (lockType=0x2) failed
05-06 17:15:05.501  4002  4002 E memalloc: /dev/pmem: Failed to map buffer size:36671488 offset:36634624 fd:367 Error: Out of memory
05-06 17:15:05.501  4002  4002 E msm7627a.gralloc: Could not mmap handle 0x4419eba0, fd=367 (Out of memory)
05-06 17:15:05.501  4002  4002 E libgenlock: perform_lock_unlock_operation: GENLOCK_IOC_LOCK failed (lockType0x1, err=Bad file number fd=367)
05-06 17:15:05.501  4002  4002 E msm7627a.gralloc: gralloc_lock: genlock_lock_buffer (lockType=0x2) failed
05-06 17:15:05.501  4002  4002 W GraphicBufferMapper: lock(...) failed -22 (Invalid argument)
05-06 17:15:05.511  4002  4002 I Gecko   : [Child 4002] ###!!! ABORT: NULL actor value passed to non-nullable param: file /home/geeksphone/FOS/peak/objdir-gecko/ipc/ipdl/PLayerTransactionChild.cpp, line 2138
05-06 17:15:05.511  4002  4002 E Gecko   : mozalloc_abort: [Child 4002] ###!!! ABORT: NULL actor value passed to non-nullable param: file /home/geeksphone/FOS/peak/objdir-gecko/ipc/ipdl/PLayerTransactionChild.cpp, line 2138

Comment 3

[Cristian Rodriguez (:crdlc)]

I don't have a Peak device so feel free to fix it or someone with this device. I don't see issues with my hamachi, sorry

Comment 4

[Jason Smith [:jsmith]]

QA Wanted to check to see if we can reproduce this on a Open C running 1.4.

Comment 5

[Julien Wajsberg [:julienw]]

I think the crash is unrelated to this bug and could be specific to the Peak (it has an history of weird graphics issues).

I'm surprised that the other issues are not happening on other phones though... Maybe this happens only on HD phones?

Comment 6

[Josh Schmitt [Joshs]]

(In reply to Jason Smith [:jsmith] from comment #4)
> QA Wanted to check to see if we can reproduce this on a Open C running 1.4.

Issue repros on the Open_C 1.4 Build with the latest Open_C base image.

Moving an app to the first page from the second will move the last icon from the first page to the second page.

1.4 Environmental Variables:
Device: Open_C 1.4
BuildID: 20140506000202
Gaia: b1242f33981024de59b8b4c26bacff8b876211b1
Gecko: fe4080728c60
Version: 30.0
Firmware Version: FFOS_US_EBAY_P821A10V1.0.0B06_LOG_DL

Comment 7

[Jason Smith [:jsmith]]

This is likely a regression, but let's double check.

Can you check to see if this reproduces on 1.3?

Comment 8

[Julien Wajsberg [:julienw]]

Can you also check the STR2 on Open C 1.4?

Comment 9

[Julien Wajsberg [:julienw]]

IMO the STR1 is not _that_ important (the icon is still visible) and I wouldn't make it a blocker; however the STR2 is critical because the icon is not visible at all.

Comment 10

[Jason Smith [:jsmith]]

(In reply to Julien Wajsberg [:julienw] (away May 8th) from comment #8)
> Can you also check the STR2 on Open C 1.4?

We need separate bugs for these issues. This is going to be confusing triage to analyze the bug otherwise.

Comment 11

[Jason Smith [:jsmith]]

Julien - Can you refile this bug as two separate bugs for each problem separately?

Comment 12

[Julien Wajsberg [:julienw]]

Not now, I was too busy with 1.3t+ issues and needs to go now. It's easy to do if you want to do it since I clearly separated the 2 STR in the description, but I'll do it friday otherwise.

Comment 13

[Joshua Mitchell (Inactive)]

(In reply to Jason Smith [:jsmith] from comment #7)
> This is likely a regression, but let's double check.
> 
> Can you check to see if this reproduces on 1.3?

This issue DOES reproduce on the latest Open C 1.3 (base build)
1.3 Environmental Variables:
Device: Open_C 1.3
BuildID: 20140505052400
Gaia: Unknown Git commit; build date shown here.
Gecko:
Version: 28.0
Firmware Version: FFOS_US_EBAY_P821A10V1.0.0B06_LOG_DL

Comment 14

[Julien Wajsberg [:julienw]]

Filed bug 1008161 for STR2, let's focus on STR1 here.

Comment 15

[Julien Wajsberg [:julienw]]

The patch that I tested from bug 1008161 fixed this as well, so I'll dupe there.