Bug 11152 - Loading locks up the browser on mobile devices
Summary: Loading locks up the browser on mobile devices
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Page Loading (show other bugs)
Version: 420+
Hardware: Other All
: P2 Normal
Assignee: David Carson
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2006-10-04 08:01 PDT by David Carson
Modified: 2006-10-05 14:43 PDT (History)
1 user (show)

See Also:

Attachments
patch (2.18 KB, patch)
2006-10-04 08:11 PDT, David Carson 		no flags Details | Formatted Diff | Diff
Corrected patch after Zalan's comment (2.17 KB, patch)
2006-10-04 09:21 PDT, David Carson 		darin: review+
Details | Formatted Diff | Diff
Show Obsolete (1) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description David Carson 2006-10-04 08:01:03 PDT 
When loading large sites on mobile devices, the browser often feels like it locks up. The user can not navigate the partially loaded content.
Comment 1 David Carson 2006-10-04 08:11:17 PDT 
Created attachment 10900 [details]
patch

The tokenizer chunk size is quite large for low CPU devices, which results in the browser spending all it's time in the tokenizer. With a smaller chunk size, the tokenizer yields more often allowing the browser to respond to key events, etc. With the smaller chunk size, the tokenizer should not need to wait as before resuming tokenizing.
Comment 2 zalan 2006-10-04 09:16:40 PDT 
'#define MOBILE 0' makes '#ifdef MOBILE' true, use #if MOBILE instead.
Comment 3 David Carson 2006-10-04 09:21:00 PDT 
Created attachment 10901 [details]
Corrected patch after Zalan's comment

Changed #ifdef MOBILE to #if MOBILE
Comment 4 Darin Adler 2006-10-05 07:49:56 PDT 
Comment on attachment 10901 [details]
Corrected patch after Zalan's comment

The only thing I'm concerned about here is that there's a comment that mentions 200ms in the non-mobile case, but for some reason the we are using 300ms for the mobile case.

r=me
Comment 5 Mark Rowe (bdash) 2006-10-05 14:43:55 PDT 
Landed in r16821.