Bug 35301

Summary:

[Qt] HTTP 307 after a 303 after a POST re-sends POST data from the original request

Product:

WebKit

Reporter:

Brady Eidson <beidson>

Component:

Page Loading

Assignee:

QtWebKit Unassigned <webkit-qt-unassigned>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

jturcotte, markus, robert

Priority:

Keywords:

Qt, QtTriaged

Version:

528+ (Nightly build)

Hardware:

All

OS:

All

Attachments:

Description	Flags
Patch	none
Patch	kenneth: review+

Brady Eidson

Reported 2010-02-23 09:33:07 PST

HTTP 307 after a 303 after a POST re-sends POST data from the original request. See https://bugs.webkit.org/show_bug.cgi?id=31410 for more discussion and the fix that went in for Mac and Windows. Qt needs to do the same.

Attachments
Patch (3.18 KB, patch) 2010-06-22 13:02 PDT, Robert Hogan	no flags	Details Formatted Diff Diff
Patch (4.11 KB, patch) 2010-06-24 13:54 PDT, Robert Hogan	kenneth: review+	Details Formatted Diff Diff
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.

Tor Arne Vestbø

Comment 1 2010-03-10 06:42:54 PST

Please follow the QtWebKit bug reporting guidelines when reporting bugs. See http://trac.webkit.org/wiki/QtWebKitBugs Specifically: - The 'QtWebKit' component should only be used for bugs/features in the public QtWebKit API layer, not to signify that the bug is specific to the Qt port of WebKit http://trac.webkit.org/wiki/QtWebKitBugs#Component - Add the keyword 'Qt' to signal that it's a Qt-related bug http://trac.webkit.org/wiki/QtWebKitBugs#Keywords

Robert Hogan

Comment 2 2010-06-22 13:02:12 PDT

Created attachment 59407 [details] Patch

Kenneth Rohde Christiansen

Comment 3 2010-06-22 14:36:18 PDT

Comment on attachment 59407 [details] Patch uh, seems like a QtNetwork bug. If you do this change, please add a comment and at best a link to the qt bug report.

Robert Hogan

Comment 4 2010-06-24 13:02:43 PDT

(In reply to comment #3) > (From update of attachment 59407 [details]) > uh, seems like a QtNetwork bug. I don't think it is - the reason checking m_method doesn't work for this test is because there are two redirects following an initial POST. The first one is managed successfully: the m_method for the original request is POST, the check detects this and overrides it with GET. For the second redirect, the check sees that m_method is now GET so it doesn't override the method in the newRequest it has created from ResourceHandle::request() (which is still the original value of POST) with GET. So during a redirect chain QNetworkReplyHandler needs to check the httpMethod() or lastHTTPMethod() rather than the m_method, because that will always contain the original request method in a redirect chain. I'll update the changelog with this and use httpMethod() instead - lastHTTPMethod() is Mac-specific really.

Robert Hogan

Comment 5 2010-06-24 13:54:20 PDT

Created attachment 59699 [details] Patch

Robert Hogan

Comment 6 2010-06-25 11:11:32 PDT

Committed r61875: <http://trac.webkit.org/changeset/61875>

Note You need to log in before you can comment on or make changes to this bug.