Bug 6528

Summary:

remove all use of QDict from WebCore

Product:

WebKit

Reporter:

Darin Adler <darin>

Component:

WebKit Misc.

Assignee:

Darin Adler <darin>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

andersca

Priority:

Version:

420+

Hardware:

Mac

OS:

OS X 10.4

Attachments:

Description	Flags
remove QDict, use HashMap and HashSet instead	hyatt: review+
follow-on to fix regression in layout test	andersca: review+

Darin Adler

Reported 2006-01-13 22:57:49 PST

We want to get rid of QDict.

Attachments
remove QDict, use HashMap and HashSet instead (107.93 KB, patch) 2006-01-13 22:58 PST, Darin Adler	hyatt: review+	Details Formatted Diff Diff
follow-on to fix regression in layout test (26.95 KB, patch) 2006-01-15 13:21 PST, Darin Adler	andersca: review+	Details Formatted Diff Diff
View All Add attachment proposed patch, testcase, etc.

Darin Adler

Comment 1 2006-01-13 22:58:34 PST

Created attachment 5657 [details] remove QDict, use HashMap and HashSet instead

Dave Hyatt

Comment 2 2006-01-14 01:28:17 PST

There's tons of whitespace changes that are irrelevant to the patch. Did you mean to include these (is this just because you de-tabified the files when you touched them)?

Dave Hyatt

Comment 3 2006-01-14 01:29:25 PST

Comment on attachment 5657 [details] remove QDict, use HashMap and HashSet instead r=me, may be worth running performance tests just to verify everything is ok with all the hash tables performance-wise. Whitespace/tab changes are fine too.

Darin Adler

Comment 4 2006-01-14 10:01:05 PST

Right, Dave, the whitespace changes are because I de-tab-ified files as I touched them. I guess I will have to run the performance tests.

Anders Carlsson

Comment 5 2006-01-15 03:44:03 PST

Maybe a CountedHashSet could be used for the m_idCount table instead?

Darin Adler

Comment 6 2006-01-15 06:51:16 PST

Good idea. I used HashCountedSet for m_idCount. I also did performance tests, and it's working well.

Darin Adler

Comment 7 2006-01-15 13:21:12 PST

Created attachment 5700 [details] follow-on to fix regression in layout test

Anders Carlsson

Comment 8 2006-01-15 13:33:20 PST

Comment on attachment 5700 [details] follow-on to fix regression in layout test r=me

Note You need to log in before you can comment on or make changes to this bug.