Cynthia Tang [:cynthiatang] Reporter
	Description • 8 years ago

# Test Case
STR
1. Launch the browser with blank page
2. Open the google doc with 200+ pages
3. Press ctrl+end (to the last page)
4. close the browser

# Hardware
OS: Ubuntu 14.04 LTS 64-bit
CPU: i7-3770 3.4GMhz
Memory: 16GB Ram
Hard Drive: 1TB SATA HDD
Graphics: GK107 [GeForce GT 640]/ GF108 [GeForce GT 440/630]

# Browsers
Firefox version: Nightly 50.0a1 (2016/06/28)
Chrome version: 50.0.2661.75

# Result
Browser | Run time (median value) 
Firefox | 21,711.11 ms
Chrome  | 26,672.22 ms

# Video
https://www.youtube.com/watch?v=LYBBG3c4ZFQ

# Profiler
https://cleopatra.io/#report=8112066661d2bb52e0501258969b7a651bca74b2

# Test Script:
https://github.com/Mozilla-TWQA/Hasal/blob/master/tests/test_firefox_gdoc_pageend.sikuli/test_firefox_gdoc_pageend.py

	Bobby Chien
	Comment 1 • 8 years ago

Mark as minor case, and marked wontfix in this iteration. Revisit here if necessary.

This bug fixing iteration will focus on critical bug 1264535, bug 1269666, bug 1269684, bug 1269690, bug 1269695, bug 1269698.

	Andrew Overholt [:overholt]
	Comment 2 • 8 years ago

Can you please provide a link to the test document used?

Not much in the profile to go on here ... :smaug likes using his profiler (the name of which I forget) so please needinfo him once you've provided the URL so he can reproduce locally.

Thanks!

	Cynthia Tang [:cynthiatang] Reporter
	Comment 3 • 8 years ago

Hi Andrew,
The test document's url is https://docs.google.com/document/d/1EPSmGqm2r4Qq42B4t1VOYacjTlL0JVuC8JSlUvoIhss/edit

Thanks!

	Jerry Shih[:jerry] (UTC+8) (inactive)
	Comment 4 • 8 years ago

Olli, here is the test url[1]. Maybe you could find another info with your profiler.
Please check comment 2.

[1]
https://docs.google.com/document/d/1EPSmGqm2r4Qq42B4t1VOYacjTlL0JVuC8JSlUvoIhss/edit

	Olli Pettay [:smaug][bugs@pettay.fi]
	Comment 5 • 8 years ago

Page load time seems to be mostly JS, and then there is some reflow too, and page scroll looks the same. (I was using Zoom)
Would be good to get some JS engine dev to look at this. Perhaps jandem?

	Andrew Overholt [:overholt]
	Comment 6 • 8 years ago

Naveed can perhaps help.

	Naveed Ihsanullah [:naveed]
	Comment 7 • 8 years ago

Am I reading the profiler output correct? Most of the time is spent in Content syscall inside PollWrapper?

There do appear to be a bunch of bailouts. Are those JIT bails?

Might be interesting comparing trace of Chrome to see where they spend time. Might give us areas that lend themselves to optimizations.

	Bobby Chien
	Comment 8 • 8 years ago

more finding on this issue?

	Jan de Mooij [:jandem]
	Comment 9 • 8 years ago

Hannes maybe you can help profile this? I have a bunch of crashes to investigate.

	Steve Fink [:sfink] [:s:]
	Comment 10 • 8 years ago

Could you clarify:

(In reply to Cynthia Tang [:cynthiatang] from comment #0)
> # Result
> Browser | Run time (median value) 
> Firefox | 21,711.11 ms
> Chrome  | 26,672.22 ms

Those numbers seem to be saying that Firefox is *faster*, not slower. Is that correct, or are the numbers reversed?

	Cynthia Tang [:cynthiatang] Reporter
	Comment 11 • 8 years ago

Hi Steve,
Sorry, it's my fault. I just re-checked the result. It should be ...

Browser | Run time (median value) 
Chrome  | 21,711.11 ms
Firefox | 26,672.22 ms

Thank you.

	Bobby Chien
	Comment 12 • 8 years ago

Steve, do we have update on this issue?

	Hannes Verschore [:h4writer]
	Comment 13 • 8 years ago

Clearing myself. IIUC sfink and sstangl are looking into the google docs issues.

	Nicolas B. Pierron [:nbp]
	Comment 14 • 6 years ago

Nobody looked into these bugs for a while, moving to P2 and waiting for [qf] re-evaluation before moving back to P1.

	Daniel Holbert [:dholbert]
	Comment 15 • 6 years ago

[qf-] because opening 200+ tabs is not a typical use case.  Other blockers of bug 1260981 may be more relevant for qf.