Bug 222441 – Streamline decoding NSURL so it doesn't require allocating NSData and autoreleasing

• Status: RESOLVED
• Resolution: INVALID
• Component: New Bugs

[Darin Adler]

Streamline decoding NSURL so it doesn't require allocating NSData and autoreleasing

[Darin Adler]

Created attachment 421577 [details]
Patch

[Sam Weinig]

Comment on attachment 421577 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=421577&action=review

> Source/WTF/wtf/cocoa/NSURLExtras.mm:153
> +NSURL *URLWithData(const void* bytes, size_t length, NSURL *baseURL)

What is the reason behind using const void* here rather than const uint8_t*? With string code, I like to be more specific if I can to avoid people accidentally passing in UChar* to a function like this.

[Darin Adler]

Comment on attachment 421577 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=421577&action=review

>> Source/WTF/wtf/cocoa/NSURLExtras.mm:153
>> +NSURL *URLWithData(const void* bytes, size_t length, NSURL *baseURL)
> 
> What is the reason behind using const void* here rather than const uint8_t*? With string code, I like to be more specific if I can to avoid people accidentally passing in UChar* to a function like this.

I was trying to side step the three different types for bytes and was a little bit influenced by the types in NSData methods. But your argument seems compelling, and CFData functions take uint8_t*. I will go with uint8_t instead.

[Darin Adler]

Looks like Brent’s patch had to be rolled out. I’ll have to wait to land this until it’s landed again.

[Radar WebKit Bug Importer]

<rdar://problem/75061347>

[Darin Adler]

Not needed any more. Brent’s latest is using CFURLCreateAbsoluteURLWithBytes now, not URLWithData, so there’s no need for this.