[Webkit-unassigned] [Bug 148388] We should also store the time information for recent searches

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=148388

--- Comment #17 from Darin Adler <darin at apple.com> ---
Comment on attachment 261270
  --> https://bugs.webkit.org/attachment.cgi?id=261270
Patch v3

View in context: https://bugs.webkit.org/attachment.cgi?id=261270&action=review

> Source/WebCore/platform/SearchPopupMenu.h:36
> +    virtual void saveRecentSearches(const AtomicString& name, const Vector<std::pair<String, double>>& searchItems) = 0;
> +    virtual void loadRecentSearches(const AtomicString& name, Vector<std::pair<String, double>>& searchItems) = 0;

This interface is confusing. Nothing about std::pair<String, double> indicates to me that one element is the search string and the other is a time. We’d probably be better off with a struct so we can name the two members.

Also, I suggest for new code we use std::chrono. Instead of double we would use std::chrono::time_point<std::chrono::system_clock>.

> Source/WebCore/rendering/RenderSearchField.cpp:95
> +    int size = static_cast<int>(m_recentSearches.size());
> +    for (int i = size - 1; i >= 0; --i) {
> +        if (m_recentSearches[i].first == value)
> +            m_recentSearches.remove(i);
> +    }

This is really ugly, especially the static_cast<int> part. We should try to do a little better.

> Source/WebCore/rendering/RenderSearchField.cpp:96
> +    m_recentSearches.insert(0, std::make_pair(value, (double)time(nullptr)));

We should definitely not be using the C standard library function time and converting it to a double! If we use std::chrono, then it would be std::chrono::system_clock::now().

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20150916/9820b361/attachment.html>