Closed Bug 783582 Opened 12 years ago Closed 2 years ago

BrowserQuest spends 75% of its time in PaintGradient

Categories

(Websites Graveyard :: BrowserQuest, defect)

ARM
Android
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED INCOMPLETE

People

(Reporter: vlad, Assigned: mbest)

References

Details

On a Nexus 7 with a very recent nightly, BrowserQuest spends 75% of its time in PaintBackground -> PaintGradient. On desktop, 25% of the time is there as well. On destkop since I have full stack traces, PaintBackground -> nsImageRenderer::Draw -> PaintGradient in case that matters. I believe this is because the body element has: background-image: radial-gradient(transparent 40%, rgb(0, 0, 0) 125%), url("http://cdn.mozilla.net/browserquest/img/1/wood.png"), none;
One could also reasonably ask why that background is being repainted all the time, given that it's not changing at all as you play the game.. the game is played in a canvas, so maybe we should cache the background rendering into a layer?
(Also note: my desktop has d2d/d3d disabled)
We should not be invalidating the background which should be retained in a thebes layer. Is this is expected? Does DLBI fix it?
Anyone know if there's someone on the BrowserQuest side that we can Cc here? Would be nice to just remove the gradient on mobile, perhaps using a CSS selector...
We need to get rid of that gradient everywhere; it's causing problems on older desktop machines anyway. However! We should create a benchmark/testcase that does something similar so that we can actually get a fix in for the gradients.
Assignee: nobody → mbest
Did DLBI fix the problem?
Flags: needinfo?(mbest)
I'll verify and get back to you.
Flags: needinfo?(mbest)
Here are the two guys that currently control the repo for future reference: Guillaume Lecollinet – guillaume@littleworkshop.fr Franck Lecollinet – franck@littleworkshop.fr I have also sent them an email with a link to this bug and a request to see if it was corrected. Last I looked at this issue, Rob Hawkes was going to look into this but he is no longer with Mozilla.
Vlad is looking at this, we have verified that the bug is still in place. Vlad is attempting to get access to the Github repo that contains the source so that we can correct the problem.
Any news? Did you get access to the repo?
Flags: needinfo?(vladimir)
I just got access to the repo. There are a few issues currently. 1. We don't own the server that browserquest.mozilla.org points to. This makes updated deployments difficult. The third party that wrote this currently hosts this and I don't know if we have a contract with them for a certain amount of time. 2. I don't know who can modify the domain browserquest.mozilla.org to point to a Mozilla owned server, possibly run on Stackato. 3. There is a thriving community that's been adding to browserquest: https://github.com/browserquest/BrowserQuest They have added a bit, but I've also noticed a few regressions. Either way, they have too many patches to upstream all at once. It might be nice to try to upstream some of their patches, or have our own fork.
Flags: needinfo?(vladimir)
Filed bug 906169 for the underlying graphics problem.
Component: Graphics → BrowserQuest
Product: Core → Websites
Product: Websites → Websites Graveyard

This bug lies at rest in the graveyard.

Status: NEW → RESOLVED
Closed: 2 years ago
Resolution: --- → INCOMPLETE
You need to log in before you can comment on or make changes to this bug.