Bug 164952 – Use Vector::uncheckedAppend() in more places

• Status: RESOLVED
• Resolution: FIXED
• Component: WebCore Misc.

[Daniel Bates]

We can use Vector::uncheckedAppend() whenever the number of items that will be appended to a vector is equal to or less than its capacity. Many call sites that call Vector::reserveInitialCapacity() and then use Vector::append() can be written to use Vector::uncheckedAppend().

[Daniel Bates]

Created attachment 295187 [details]
Patch

[Darin Adler]

Comment on attachment 295187 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=295187&action=review

> Source/WebCore/contentextensions/DFAMinimizer.cpp:90
> +        m_sets.uncheckedAppend(SetDescriptor({ 0, size, 0 }));

Peculiar that we reserve capacity for "size" elements, but here we do only a single uncheckedAppend. Did we reserve too much capacity?

Syntax is peculiar; I think it should just be:

    m_sets.uncheckedAppend(SetDescriptor { 0, size, 0 });

No need to have both parentheses and braces.

[Daniel Bates]

(In reply to comment #2)
> Comment on attachment 295187 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=295187&action=review
> 
> > Source/WebCore/contentextensions/DFAMinimizer.cpp:90
> > +        m_sets.uncheckedAppend(SetDescriptor({ 0, size, 0 }));
> 
> Peculiar that we reserve capacity for "size" elements, but here we do only a
> single uncheckedAppend. Did we reserve too much capacity?
> 

We may have reserved too much though it is not immediately obvious from reading the code. I hope you do not mind that I defer proving the bound of m_set to bug #165472.

> Syntax is peculiar; I think it should just be:
> 
>     m_sets.uncheckedAppend(SetDescriptor { 0, size, 0 });
> 
> No need to have both parentheses and braces.

Will fix before landing.

[Daniel Bates]

Committed r209400: <http://trac.webkit.org/changeset/209400>