Closed Bug 970735 Opened 10 years ago Closed 10 years ago

[OOM]High def webpage take a long time to load

Categories

(Web Compatibility :: Site Reports, defect, P2)

ARM
Gonk (Firefox OS)
defect

Tracking

(Not tracked)

RESOLVED WONTFIX

People

(Reporter: guigs, Unassigned)

References

()

Details

Attachments

(1 file)

Steps to reproduce: 
1. there is an error due to picture. 
2. i press reload a couple of time but not working anymore.
 the error is https://m.ak.fbcdn.net/sphotos-a.ak/hphotos-ak-prn1/t1/1908483_209927625874419_1956547350_n.jpg


Example: 
    qcom model
   Boot2Gecko1.1.0hd-GP



url example
[http://www.geekbackpacker.com/kalasin.php]

[https://support.mozilla.org/en-US/questions/985717]
OS: Mac OS X → Gonk (Firefox OS)
Hardware: x86 → ARM
Error message: Well, this is embarrassing. We tried to display this Web page, but it's not responding
Flags: needinfo?(nhirata.bugzilla)
It's a tab crash; possibly due to out of memory.  Can you reproduce the issue and then send us the dmesg log?

adb shell dmesg > dmesg.log 

and attach the dmesg.log to this bug.
Flags: needinfo?(rmcguigan)
1.2 prerelease the page does not load on the first try, if reload it loads for about 30 seconds then displays the error logs attached
Flags: needinfo?(rmcguigan)
Attached file dmesg.log
dmesg.log after crashing the browser a few times. =)
Looks like it did get killed while loading.  It is an out of memory issue from what I see:
<3>[67422.277745] Out of memory: Kill process 1564 (Browser) score 838 or sacrifice child
Flags: needinfo?(nhirata.bugzilla)
Keywords: perf
Summary: High def webpage take a long time to load → [OOM]High def webpage take a long time to load
Not sure if memshrink/perf stuff is in your court or not, Mike... if not, can you move it to the right person?
Flags: needinfo?(mlee)
Flags: needinfo?(mlee)
Keywords: footprint
Priority: -- → P2
Whiteboard: [c=memory p= s= u=] [MemShrink]
Just to clarify, does this fail on v1.1 or just v1.2+?  A very tall page like this could trigger bug 947523 which would present this way.
Assignee: nobody → bkelly
Whiteboard: [c=memory p= s= u=] [MemShrink] → [c=memory p= s= u=] [MemShrink:P2]
(In reply to Ben Kelly [:bkelly] from comment #7)
> Just to clarify, does this fail on v1.1 or just v1.2+?  A very tall page
> like this could trigger bug 947523 which would present this way.

Agreed - we should find this out first.

Note - This could end up being a TE problem if the website isn't designed for mobile.

QA Wanted to check the behavior on 1.1.
Keywords: qawanted
This is a case of the website not being designed for mobile.  The page has ~50 images, most of which range from 1.5MB to 3MB compressed.  This means we end up using ~85MB just to load the image resources before we even get around to decoding anything.  Unfortunately we just don't have enough memory for that.

Jason, what is the correct way to mark this as a tech evangelism issue?
Flags: needinfo?(jsmith)
Keywords: qawanted
(In reply to Ben Kelly [:bkelly] from comment #9)
> This is a case of the website not being designed for mobile.  The page has
> ~50 images, most of which range from 1.5MB to 3MB compressed.  This means we
> end up using ~85MB just to load the image resources before we even get
> around to decoding anything.  Unfortunately we just don't have enough memory
> for that.
> 
> Jason, what is the correct way to mark this as a tech evangelism issue?

Let's move it over to TE --> Mobile component.
Component: Gaia::Browser → Mobile
Flags: needinfo?(jsmith)
Keywords: footprint, perf
Product: Firefox OS → Tech Evangelism
Whiteboard: [c=memory p= s= u=] [MemShrink:P2]
Thanks Jason.  I'm unassigning myself here.
Assignee: bkelly → nobody
(In reply to Jason Smith [:jsmith] from comment #8)
> (In reply to Ben Kelly [:bkelly] from comment #7)
> > Just to clarify, does this fail on v1.1 or just v1.2+?  A very tall page
> > like this could trigger bug 947523 which would present this way.
> 
> Agreed - we should find this out first.
> 
> Note - This could end up being a TE problem if the website isn't designed
> for mobile.
> 
> QA Wanted to check the behavior on 1.1.

www.geekbackpacker.com/singburi.php  also same problem. i saw in Andriod no problem :(
At some point, every phone or device is going to hit hardware limitations. If all of the following are true:
* The Android phone you tested on has more memory than the Firefox OS phone
* It's not realistic to try to optimise our memory usage more to handle this state
* Firefox doesn't crash

then this is just one of those cases we probably can't make work. 

(On my HTC One, in Android stock browser, the Kalasin page loads very slowly, and the browser is barely responding for a long time. When it starts responding again, I can scroll down a couple of pages, then the photos stop rendering or render only partially, and the text is unclear. A bit further down, the page is just white. So I definitely see problems here.)

If it were some big, professional site that served a separate version for smartphones, we could reach out and ask them to give us the right version. However, this site seems more like somebody's blog and he probably hasn't made a mobile version of it in the first place. I don't think it makes much sense to ask somebody to remove images from his blog or make an alternate mobile-friendly page just because of a random memory limitation in today's phone harware - 

Resolving, but thanks for reporting it anyway.
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → WONTFIX
Product: Tech Evangelism → Web Compatibility
Component: Mobile → Site Reports
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: