| Summary: | [Qt] Use DirectConnection for the QNetworkReply in the QNetworkReplyHandler | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | WebKit | Reporter: | Holger Freyther <zecke> | ||||||
| Component: | Platform | Assignee: | Nobody <webkit-unassigned> | ||||||
| Status: | RESOLVED FIXED | ||||||||
| Severity: | Normal | CC: | hausmann, markus, webkit.review.bot | ||||||
| Priority: | P2 | ||||||||
| Version: | 528+ (Nightly build) | ||||||||
| Hardware: | PC | ||||||||
| OS: | OS X 10.5 | ||||||||
| Attachments: |
|
||||||||
style-queue ran check-webkit-style on attachment 44224 [details] without any errors.
Landed it in r52089. Waiting for the bot to run the tests before closing the bug. mgoetz just highlighted that I have blindly converted everything to DirectConnection. With some luck I'm not reverting bug #25556. Once we are confident that it is not crashing we should remove the signal and do a direct function call again. This patch depends on a bugfix inside Qt 4.6. It should hit qt.gitorious.org in the next days. http://pastebin.ca/1714463 Created attachment 44852 [details]
[Qt] Followup for direct signals and slots connections in QNetworkReplyHandler
style-queue ran check-webkit-style on attachment 44852 [details] without any errors.
Comment on attachment 44852 [details]
[Qt] Followup for direct signals and slots connections in QNetworkReplyHandler
Good idea and good save :), r=me
|
Created attachment 44224 [details] Use direct Qt signal connections We should use DirectConnection to achieve a improvement in latency on network jobs. This also means that the loading will be more stable across different runs. The only possible problem is that it might cause crashes on Qt4.5 and Qt4.6... we will have to figure if it is crashing or not.