RESOLVED FIXED 189528
Expose fewer of URL's internal members 
https://bugs.webkit.org/show_bug.cgi?id=189528
Summary Expose fewer of URL's internal members
Alex Christensen
Reported 2018-09-11 15:53:11 PDT
Expose fewer of URL's internal members
Attachments
Patch (4.57 KB, patch)
2018-09-11 15:53 PDT, Alex Christensen 		no flags
DetailsFormatted DiffDiff
Patch (5.02 KB, patch)
2018-09-12 10:16 PDT, Alex Christensen 		no flags
DetailsFormatted DiffDiff
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.
Alex Christensen
Comment 1 2018-09-11 15:53:41 PDT
Created attachment 349484 [details] Patch
Alex Christensen
Comment 2 2018-09-12 10:16:50 PDT
Created attachment 349557 [details] Patch
WebKit Commit Bot
Comment 3 2018-09-12 14:01:13 PDT
Comment on attachment 349557 [details] Patch Clearing flags on attachment: 349557 Committed r235949: <https://trac.webkit.org/changeset/235949>
WebKit Commit Bot
Comment 4 2018-09-12 14:01:15 PDT
All reviewed patches have been landed. Closing bug.
Radar WebKit Bug Importer
Comment 5 2018-09-12 14:03:34 PDT
<rdar://problem/44393455>
Darin Adler
Comment 6 2018-09-13 18:55:01 PDT
Comment on attachment 349557 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=349557&action=review > Source/WebCore/platform/URL.cpp:434 > +unsigned URL::hostStart() const > +{ > + return (m_passwordEnd == m_userStart) ? m_passwordEnd : m_passwordEnd + 1; > +} What is the value in de-inlining this? Does it improve performance or code size or something?
Alex Christensen
Comment 7 2018-09-14 10:53:01 PDT
It's now only used in URL.cpp. Moving it makes URL.h less of a mess, and I believe it would always be inlined in release builds.
Darin Adler
Comment 8 2018-09-15 08:27:47 PDT
(In reply to Alex Christensen from comment #7) > It's now only used in URL.cpp. Moving it makes URL.h less of a mess, and I > believe it would always be inlined in release builds. Makes sense. I have to remember that in modern times the inline keyword now means "OK to have this in a header / multiple translation units" rather than literally "compile this function inline". Which means we probably *never* need to specify inline for a function template.
Note You need to log in before you can comment on or make changes to this bug.