1141819 - Don't implicitly dispatch a runnable to the main thread in RasterImage::StartDecoding

Status: RESOLVED FIXED

(Core :: Graphics: ImageLib, defect)

Description

[Seth Fowler [:seth] [:s2h]]

After bug 1139225, there's no longer a reason for RasterImage::StartDecoding to work implicitly off-main-thread. We should remove the implicit runnable dispatch code, so we don't encourage people to call this method off-main-thread.

Note that if anyone does do so, we'll know, because there's an assertion in RasterImage::RequestDecodeForSize (which is used to implement RasterImage::StartDecoding) which will fire if it runs off-main-thread. So we don't need a new assertion here.

Comment 1

[Seth Fowler [:seth] [:s2h]]

Attachment: Don't implicitly dispatch a runnable to the main thread in RasterImage::StartDecoding

Here's the patch.

Comment 2

[Seth Fowler [:seth] [:s2h]]

Whoops, this got mixed in with a large stack of commits I just pushed to inbound by mistake. I thought this had already been r+'d.

bb86ca5e5552	Seth Fowler — Bug 1141819 - Don't implicitly dispatch a runnable to the main thread in RasterImage::StartDecoding. r=baku

I don't want to back out because doing so will require another round of testing.

Andrea, if you have time, could you review this soon? I'll push a followup with any changes you request. If you don't have time, please let me know and I'll ask someone else to review.

Comment 3

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/bb86ca5e5552

Comment 4

[Andrea Marchesini [:baku]]

Comment on attachment 8575634 [details] [diff] [review]
Don't implicitly dispatch a runnable to the main thread in RasterImage::StartDecoding

Review of attachment 8575634 [details] [diff] [review]:
-----------------------------------------------------------------

Thanks for the explanation by email.