Bug 198175 – [WHLSL] Add a helper for in-place AST mutation

• Component: WebGPU
• Status: RESOLVED FIXED

[Myles C. Maxfield]

Something like

template <typename Old, typename New, typename ...Args>
New* replace(Old& old, Args&&... args)
{
    static_assert(sizeof(New) <= sizeof(Old));
    return new (&old) New(std::forward<Args>(args)...);
}

[Saam Barati]

patch forthcoming

[Saam Barati]

Created attachment 370510 [details]
patch

[Myles C. Maxfield]

Comment on attachment 370510 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=370510&action=review

> Source/WebCore/Modules/webgpu/WHLSL/AST/WHLSLNode.h:56
> +    return new (&old) New(std::forward<Args>(args)...);

Don’t you have to save the address of “old” before you call its destructor?

[Saam Barati]

(In reply to Myles C. Maxfield from comment #3)
> Comment on attachment 370510 [details]
> patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=370510&action=review
> 
> > Source/WebCore/Modules/webgpu/WHLSL/AST/WHLSLNode.h:56
> > +    return new (&old) New(std::forward<Args>(args)...);
> 
> Don’t you have to save the address of “old” before you call its destructor?

Spoke with Myles offline. This code is probably fine (since dtor is just a method call), but we're just going to be explicit and do it before.

[Saam Barati]

Created attachment 370512 [details]
patch for landing

[WebKit Commit Bot]

Comment on attachment 370512 [details]
patch for landing

Clearing flags on attachment: 370512

Committed r245706: <https://trac.webkit.org/changeset/245706>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Radar WebKit Bug Importer]

<rdar://problem/51079016>