Bug 130438 – Small cleanup of empty string

• Component: Web Template Framework
• Status: RESOLVED FIXED

[Gavin Barraclough]

Some prep for https://bugs.webkit.org/show_bug.cgi?id=128624.

[Gavin Barraclough]

Created attachment 227144 [details]
Fix

[Andreas Kling]

The PCRE workaround removal has been attempted before, see bug 123265.

[Gavin Barraclough]

(In reply to comment #2)
> The PCRE workaround removal has been attempted before, see bug 123265.

Hmmm, interesting. :-(

The tests mentioned in the bug pass for me, maybe I should try again. :-/

[Anders Carlsson]

Comment on attachment 227144 [details]
Fix

View in context: https://bugs.webkit.org/attachment.cgi?id=227144&action=review

> Source/WTF/wtf/text/StringImpl.h:164
> +        , m_data8(reinterpret_cast<LChar*>(1))

Even though this is what the comment said, I think it's awkward. I think you should do what we do for the StringImpl(CreateEmptyUniqueTag) constructor instead:

// We expect m_length to be initialized to 0 as we use it
// to represent a null terminated buffer.
, m_data8(reinterpret_cast<const LChar*>(&m_length))

> Source/WTF/wtf/text/StringStatics.cpp:46
> +    DEPRECATED_DEFINE_STATIC_LOCAL(StringImpl, emptyString, (ConstructEmptyString));

Please change this to

static NeverDestroyed<StringImpl> emptyString(ConstructEmptyString);

instead.

[Gavin Barraclough]

Created attachment 227148 [details]
Patch with review comments for EWS

[Gavin Barraclough]

Fixed in r165906