Bug 162247 – Make URLSearchParams spec-compliant

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

[Alex Christensen]

Make URLSearchParams spec-compliant

[Alex Christensen]

Created attachment 289296 [details]
Patch

[Chris Dumez]

Comment on attachment 289296 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=289296&action=review

> Source/WebCore/html/DOMURL.cpp:94
>  void DOMURL::setQuery(const String& query)

Who is calling this method? I quickly grep'd and did not find any obvious call site.

> Source/WebCore/html/DOMURL.cpp:98
> +        m_searchParams->setContents(URLSearchParams::create(query).get());

Shouldn't this get use search() so that we get the *parsed* query from m_url?

> Source/WebCore/html/DOMURL.cpp:129
> +        m_searchParams = URLSearchParams::create("?" + m_url.query(), this);

Can we just use search() instead of "?" + m_url.query(). Seems simpler.

> Source/WebCore/html/URLSearchParams.h:47
> +    void setContents(const Vector<std::pair<String, String>>& pairs) { m_pairs = pairs; }

Instead of doing can we have an updateFromAssocatedURL() method that does the right thing? I don't like that we have to construct a URLSearchParam every time the URL gets updated.

[Build Bot]

Comment on attachment 289296 [details]
Patch

Attachment 289296 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.webkit.org/results/2109124

New failing tests:
imported/w3c/web-platform-tests/url/urlsearchparams-constructor.html

[Build Bot]

Created attachment 289299 [details]
Archive of layout-test-results from ews107 for mac-yosemite-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: ews107  Port: mac-yosemite-wk2  Platform: Mac OS X 10.10.5

[Build Bot]

Comment on attachment 289296 [details]
Patch

Attachment 289296 [details] did not pass ios-sim-ews (ios-simulator-wk2):
Output: http://webkit-queues.webkit.org/results/2109105

New failing tests:
imported/w3c/web-platform-tests/url/urlsearchparams-constructor.html

[Build Bot]

Created attachment 289300 [details]
Archive of layout-test-results from ews125 for ios-simulator-elcapitan-wk2

The attached test failures were seen while running run-webkit-tests on the ios-sim-ews.
Bot: ews125  Port: ios-simulator-elcapitan-wk2  Platform: Mac OS X 10.11.6

[Build Bot]

Comment on attachment 289296 [details]
Patch

Attachment 289296 [details] did not pass mac-debug-ews (mac):
Output: http://webkit-queues.webkit.org/results/2109268

New failing tests:
imported/w3c/web-platform-tests/url/urlsearchparams-constructor.html

[Build Bot]

Created attachment 289302 [details]
Archive of layout-test-results from ews117 for mac-yosemite

The attached test failures were seen while running run-webkit-tests on the mac-debug-ews.
Bot: ews117  Port: mac-yosemite  Platform: Mac OS X 10.10.5

[Alex Christensen]

(In reply to comment #2)
> Comment on attachment 289296 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=289296&action=review
> 
> > Source/WebCore/html/DOMURL.cpp:94
> >  void DOMURL::setQuery(const String& query)
> 
> Who is calling this method? I quickly grep'd and did not find any obvious
> call site.
URLSearchParams::updateURL

[Chris Dumez]

Comment on attachment 289296 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=289296&action=review

>>> Source/WebCore/html/DOMURL.cpp:94
>>>  void DOMURL::setQuery(const String& query)
>> 
>> Who is calling this method? I quickly grep'd and did not find any obvious call site.
> 
> URLSearchParams::updateURL

Well then isn't it super weird for URLSearchParams to call DOMURL::setQuery() and for DOMURL to update its URLSearchParams object back again? :)

[Alex Christensen]

And ideally without crashing...

[Alex Christensen]

Created attachment 289309 [details]
Patch

[Chris Dumez]

Comment on attachment 289309 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=289309&action=review

> Source/WebCore/html/DOMURL.cpp:97
> +    if (m_searchParams)

Why is this still updating search params if it is only called by search params

> Source/WebCore/html/URLSearchParams.cpp:120
> +void URLSearchParams::updateFromAssociatedURL(const DOMURL& url)

I don't think this should take a parameter?

[Alex Christensen]

(In reply to comment #13)
> > Source/WebCore/html/URLSearchParams.cpp:120
> > +void URLSearchParams::updateFromAssociatedURL(const DOMURL& url)
> 
> I don't think this should take a parameter?
It's nice to not have a null check and to have an assertion.

[Chris Dumez]

(In reply to comment #14)
> (In reply to comment #13)
> > > Source/WebCore/html/URLSearchParams.cpp:120
> > > +void URLSearchParams::updateFromAssociatedURL(const DOMURL& url)
> > 
> > I don't think this should take a parameter?
> It's nice to not have a null check and to have an assertion.

You do not need a null check, just an assertion.

[Alex Christensen]

Created attachment 289316 [details]
Patch

[Chris Dumez]

Comment on attachment 289316 [details]
Patch

R=me

[Build Bot]

Comment on attachment 289316 [details]
Patch

Attachment 289316 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.webkit.org/results/2110154

New failing tests:
imported/w3c/web-platform-tests/url/urlsearchparams-constructor.html

[Build Bot]

Created attachment 289325 [details]
Archive of layout-test-results from ews104 for mac-yosemite-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: ews104  Port: mac-yosemite-wk2  Platform: Mac OS X 10.10.5

[Alex Christensen]

Created attachment 289326 [details]
Patch

[Sam Weinig]

Comment on attachment 289326 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=289326&action=review

> Source/WebCore/html/DOMURL.h:61
> +    ~DOMURL();

We usually put the destructor up toward the top, below the creates().

> Source/WebCore/html/URLSearchParams.h:38
> +    void URLDestroyed() { m_associatedURL = nullptr; }

initial capital letter makes me sad. associatedURLDestroyed()? WeakPtr<>?

[Alex Christensen]

http://trac.webkit.org/changeset/206168

[Alex Christensen]

http://trac.webkit.org/changeset/206179