The ResoureResponse class has a HTTP status text field which is not set with the curl backend.
Created attachment 203945 [details] proposed patch Sadly the curl does not provide a way to extract the http status text so a little workaround is needed.
Comment on attachment 203945 [details] proposed patch I've just found out this patch will result a failing http test. So removing the flags and trying to fix it.
Created attachment 204022 [details] fixed patch
Comment on attachment 204022 [details] fixed patch View in context: https://bugs.webkit.org/attachment.cgi?id=204022&action=review > Source/WebCore/platform/network/curl/ResourceHandleManager.cpp:321 > + else if (header.startsWith("HTTP", false)) { what does false mean? add /* ... */ ? > Source/WebCore/platform/network/curl/ResourceHandleManager.cpp:323 > + // This is (probably) the first line of the response. > + // Try to extract the http status text from this. not very trustworthy comment, can you be more specific? Attempt to extract HTTP status...
(In reply to comment #4) > (From update of attachment 204022 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=204022&action=review > > > Source/WebCore/platform/network/curl/ResourceHandleManager.cpp:321 > > + else if (header.startsWith("HTTP", false)) { > > what does false mean? add /* ... */ ? The second argument specifies if the string matching should be case sensitive or not. I've never seen that someone adds the comment for this method. But I can do. > > Source/WebCore/platform/network/curl/ResourceHandleManager.cpp:323 > > + // This is (probably) the first line of the response. > > + // Try to extract the http status text from this. > > not very trustworthy comment, can you be more specific? > > Attempt to extract HTTP status... In case of the curl backend the FOLLOWLOCATION is set to 1, that'll mean that the libcurl will internally follow the redirections. Thus the headerCallback method will be called more than one times with the line starting "HTTP".
Comment on attachment 204022 [details] fixed patch Peter: Can you please revise the comments on line 322-323 with the additional detail you added in the bug? Once you do, I'll r+ this. I don't think you need to add a comment to the 'startsWith' method, as it's used all-over-the-place.
(In reply to comment #6) > (From update of attachment 204022 [details]) > Peter: Can you please revise the comments on line 322-323 with the additional detail you added in the bug? Once you do, I'll r+ this. I don't think you need to add a comment to the 'startsWith' method, as it's used all-over-the-place. Okay, will do.
Created attachment 204714 [details] updated patch
Comment on attachment 204714 [details] updated patch r=me.
Comment on attachment 204714 [details] updated patch Clearing flags on attachment: 204714 Committed r151668: <http://trac.webkit.org/changeset/151668>
All reviewed patches have been landed. Closing bug.