Bug 58905 – [Qt][Texmap] Do not create a texture if it will not be used.

• Status: RESOLVED
• Resolution: INVALID
• Component: Platform

[Young Han Lee]

[Qt][Texmap] Do not create a texture if it will not be used.

[Young Han Lee]

Created attachment 90215 [details]
Patch

[Young Han Lee]

Currently, all TextureMapperNodes create it's own texture even if the texture will not be used.
A lot of memory is being wasted due to this.
    
Whether a TextureMapperNode will be painted or not is determined in the sync-phase, before a texture is created. So, do not create a texture if it is not really needed.

[Laszlo Gombos]

Comment on attachment 90215 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=90215&action=review

Overall change looks good, r- to fix the typo and change come up with a better function name.

> Source/WebCore/ChangeLog:14
> +        No new tests required as no functionaliry changes.

Typo.

> Source/WebCore/platform/graphics/texmap/TextureMapperNode.h:176
> +    inline bool shouldBePainted() const

I think we should have a more descriptive name - what about haveContent() ?

[Young Han Lee]

Created attachment 91281 [details]
Patch

haveContent() looks better! thx.

[Antonio Gomes]

Comment on attachment 91281 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=91281&action=review

> Source/WebCore/platform/graphics/texmap/TextureMapperNode.h:176
> +    inline bool haveContent() const

it should be named hasContent, or something better and more descriptive.

[Young Han Lee]

Created attachment 91417 [details]
Fixed

Fixed. I didn't find a better name.

[Alexis Menard (darktears)]

Comment on attachment 91417 [details]
Fixed

View in context: https://bugs.webkit.org/attachment.cgi?id=91417&action=review

> Source/WebCore/platform/graphics/texmap/TextureMapperNode.cpp:859
> +    if (m_state.dirty && hasContent())

I'm wondering if that check should be done before. Why entering here if anyway there is nothing to render?

[Young Han Lee]

This problem is resolved by https://bugs.webkit.org/show_bug.cgi?id=60439