HashSet should work with move only types

[Anders Carlsson]

HashSet should work with move only types

[Anders Carlsson]

Created attachment 211361 [details]
Patch

[WebKit Commit Bot]

Attachment 211361 [details] did not pass style-queue:

Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'Source/WTF/ChangeLog', u'Source/WTF/wtf/HashSet.h', u'Source/WTF/wtf/HashTable.h', u'Tools/ChangeLog', u'Tools/TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj', u'Tools/TestWebKitAPI/Tests/MoveOnly.h', u'Tools/TestWebKitAPI/Tests/WTF/HashSet.cpp', u'Tools/TestWebKitAPI/Tests/WTF/Vector.cpp']" exit_code: 1
Source/WTF/wtf/HashTable.h:404:  Missing spaces around &&  [whitespace/operators] [3]
Source/WTF/wtf/HashSet.h:77:  Missing spaces around &&  [whitespace/operators] [3]
Total errors found: 2 in 8 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Geoffrey Garen]

Comment on attachment 211361 [details]
Patch

r=me

[Anders Carlsson]

Committed r155577: <http://trac.webkit.org/changeset/155577>

[Darin Adler]

Comment on attachment 211361 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=211361&action=review

>> Source/WTF/wtf/HashSet.h:77
>>          AddResult add(const ValueType&);
>> +        AddResult add(ValueType&&);
> 
> Missing spaces around &&  [whitespace/operators] [3]

Do we need both of these? Why isn’t the second one alone sufficient?

> Source/WTF/wtf/HashTable.h:399
>          AddResult add(const ValueType& value) { return add<IdentityTranslatorType>(Extractor::extract(value), value); }
> +        AddResult add(ValueType&& value) { return add<IdentityTranslatorType>(Extractor::extract(value), std::move(value)); }

Do we need both of these? Why isn’t the second one alone sufficient?

[Anders Carlsson]

(In reply to comment #5)
> (From update of attachment 211361 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=211361&action=review
> 
> >> Source/WTF/wtf/HashSet.h:77
> >>          AddResult add(const ValueType&);
> >> +        AddResult add(ValueType&&);
> > 
> > Missing spaces around &&  [whitespace/operators] [3]
> 
> Do we need both of these? Why isn’t the second one alone sufficient?

We do need both. Since ValueType is a class template parameter, ValueType&& is only going to bind to rvalues.

If ValueType were a member function template parameter, it would be able to bind to both rvalues and rvalues, and then we wouldn’t need the extra overload.