1025537 - Stop using gfxImageSurface in gfxBlur

Status: RESOLVED FIXED

(Core :: Graphics, defect)

Description

[Jonathan Watt [:jwatt]]

Stop using gfxImageSurface in gfxBlur

Comment 1

[Jonathan Watt [:jwatt]]

Attachment: patch

This code seems a bit pointless.

Comment 2

[Markus Stange [:mstange]]

Comment on attachment 8440317 [details] [diff] [review]
patch

Matt added this code, he'll know why it looks the way it does.

Comment 3

[Matt Woodrow (:mattwoodrow)]

Comment on attachment 8440317 [details] [diff] [review]
patch

Review of attachment 8440317 [details] [diff] [review]:
-----------------------------------------------------------------

::: gfx/thebes/gfxBlur.cpp
@@ +68,5 @@
>          gfxPlatform::GetPlatform()->CreateDrawTargetForData(mData, size,
>                                                              mBlur->GetStride(),
>                                                              SurfaceFormat::A8);
>      if (!dt) {
> +        return nullptr;

CreateDrawTargetForData will return nullptr if the content backend is d2d, so we need to have a fallback.

Comment 4

[Jonathan Watt [:jwatt]]

Should CreateDrawTargetForData maybe fall back to using BackendType::CAIRO in that case?

Comment 5

[Jonathan Watt [:jwatt]]

And do you have any idea why gfxPlatform::CreateDrawTargetForData creates a gfxImageSurface?

Comment 6

[Matt Woodrow (:mattwoodrow)]

(In reply to Jonathan Watt [:jwatt] from comment #4)
> Should CreateDrawTargetForData maybe fall back to using BackendType::CAIRO
> in that case?

Yeah that's probably ok for the gfxPlatform version, not the Moz2D one.

(In reply to Jonathan Watt [:jwatt] from comment #5)
> And do you have any idea why gfxPlatform::CreateDrawTargetForData creates a
> gfxImageSurface?

None at all. hg history might reveal something.

Comment 7

[Jonathan Watt [:jwatt]]

(In reply to Matt Woodrow (:mattwoodrow) from comment #6)
> None at all. hg history might reveal something.

I already checked that, but it didn't. I'll try replacing it and see if the sky falls in. :)

Comment 8

[Jonathan Watt [:jwatt]]

Attachment: patch

This seems to pass Try.

Comment 9

[Jonathan Watt [:jwatt]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/cd59e16a415d

Comment 10

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/cd59e16a415d