[chromium] Make openFileSystem check for permission first
Created attachment 93232 [details] Patch
Created attachment 93233 [details] Patch
James/Tony: please review others: FYI since you work on the FS code. I had done this pattern for all the other storage APIs, and just realized I missed FS. The motivation is that the code in src\content shouldn't know about content settings, so splitting the opening into two steps helps that. The chrome side is http://codereview.chromium.org/7012007
Comment on attachment 93233 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=93233&action=review > Source/WebKit/chromium/public/WebCommonWorkerClient.h:90 > + virtual bool allowFileSystem(WebFrame*) The WebFrame param does not belong in this interface. I know it's in the allowDatabase() method, but it doesn't belong there either. The origin of the worker can be determined w/o that parameter just fine since there's a 'client' per worker. > Source/WebKit/chromium/src/WebWorkerBase.cpp:134 > + static PassRefPtr<AllowFileSystemMainThreadBridge> create(WebWorkerBase* worker, const WTF::String& mode, WebCommonWorkerClient* commonClient, WebFrame* frame) frame not needed here either
Created attachment 93292 [details] Patch
(In reply to comment #4) > (From update of attachment 93233 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=93233&action=review > > > Source/WebKit/chromium/public/WebCommonWorkerClient.h:90 > > + virtual bool allowFileSystem(WebFrame*) > > The WebFrame param does not belong in this interface. I know it's in the allowDatabase() method, but it doesn't belong there either. The origin of the worker can be determined w/o that parameter just fine since there's a 'client' per worker. > > > Source/WebKit/chromium/src/WebWorkerBase.cpp:134 > > + static PassRefPtr<AllowFileSystemMainThreadBridge> create(WebWorkerBase* worker, const WTF::String& mode, WebCommonWorkerClient* commonClient, WebFrame* frame) > > frame not needed here either Done for both. I had done it like that to be consistent with allowDatabase, but this is fine as well since that's the same url as openFileSystem uses.
Comment on attachment 93292 [details] Patch R=me
Committed r86356: <http://trac.webkit.org/changeset/86356>
Comment on attachment 93292 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=93292&action=review > Source/WebKit/chromium/src/WebWorkerClientImpl.h:102 > + virtual bool allowFileSystem(WebFrame*) nix WebFrame here too... this wants to be overriding the WebWorkerClient method, right?
(In reply to comment #9) > (From update of attachment 93292 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=93292&action=review > > > Source/WebKit/chromium/src/WebWorkerClientImpl.h:102 > > + virtual bool allowFileSystem(WebFrame*) > > nix WebFrame here too... this wants to be overriding the WebWorkerClient method, right? oops, that was an oversight, will commit in a followup.
http://trac.webkit.org/changeset/86356 might have broken SnowLeopard Intel Release (WebKit2 Tests) The following tests are not passing: fast/lists/inlineBoxWrapperNullCheck.html