| Summary: | [Android] Android port requires platform-specific implementation of initializeMaximumHTTPConnectionCountPerHost | ||||||
|---|---|---|---|---|---|---|---|
| Product: | WebKit | Reporter: | Steve Block <steveblock> | ||||
| Component: | WebCore Misc. | Assignee: | Steve Block <steveblock> | ||||
| Status: | RESOLVED FIXED | ||||||
| Severity: | Normal | CC: | abarth, android-webkit-unforking, ap, beidson, commit-queue, eric, fishd, mjs, phanna, steveblock | ||||
| Priority: | P2 | ||||||
| Version: | 528+ (Nightly build) | ||||||
| Hardware: | Other | ||||||
| OS: | Other | ||||||
| Attachments: |
|
||||||
|
Description
Steve Block
2009-11-23 10:45:33 PST
Created attachment 43720 [details] Patch 1 for Bug 31810 I thought the point of this default was less about what the loader was capable of, and more about "being a good web citizen" or something. Donno. I'm not an expert here. (In reply to comment #2) > I thought the point of this default was less about what the loader was capable > of, and more about "being a good web citizen" or something. Donno. I'm not an > expert here. The comment in loader.cpp for maxRequestsInFlightPerHost, and the comment in ResourceRequestBase.cpp for initializeMaximumHTTPConnectionCountPerHost both seem to suggest that the objective of this limit is to match the capability of the network layer. Can you point me to an expert who can confirm the intended use? Maciej or Bradey should know. style-queue successfully ran check-webkit-style on attachment 43720 [details] without any errors
Comment on attachment 43720 [details] Patch 1 for Bug 31810 This patch is fine. Given such a high limit on connection counts per host, you might also consider disabling the REQUEST_MANAGEMENT_ENABLED define in loader.cpp. Comment on attachment 43720 [details] Patch 1 for Bug 31810 Clearing flags on attachment: 43720 Committed r52005: <http://trac.webkit.org/changeset/52005> All reviewed patches have been landed. Closing bug. |