Bug 25409

Summary: [GTK] incorrectly fallsback on cancellation
Product: WebKit Reporter: Gustavo Noronha (kov) <gustavo>
Component: WebKitGTKAssignee: Nobody <webkit-unassigned>
Status: RESOLVED FIXED    
Severity: Normal CC: xan.lopez
Priority: P2 Keywords: Gtk
Version: 528+ (Nightly build)   
Hardware: PC   
OS: All   
Attachments:
Description Flags
Fix fallback and cancellation handling
none
Do not fallback on cancellation and load interruption caused by policy change.
xan.lopez: review+
Check if we should use the fallback content in didFailLoad. xan.lopez: review+

Description Gustavo Noronha (kov) 2009-04-26 10:05:36 PDT
On some sites, there are some frames appearing with the fallback content when they shouldn't. The error message that is presented is that the load was cancelled. Also, our shouldFallBack implementation seems to miss some cancellation cases, which makes some tests fail.
Comment 1 Gustavo Noronha (kov) 2009-04-26 10:28:24 PDT
Created attachment 29803 [details]
Fix fallback and cancellation handling

 WebKit/gtk/ChangeLog                               |   16 ++++++++++++++++
 WebKit/gtk/WebCoreSupport/FrameLoaderClientGtk.cpp |    8 +++++++-
 2 files changed, 23 insertions(+), 1 deletions(-)
Comment 2 Gustavo Noronha (kov) 2009-04-26 17:28:09 PDT
Created attachment 29809 [details]
Do not fallback on cancellation and load interruption caused by policy change.

 WebKit/gtk/ChangeLog                               |   14 ++++++++++++++
 WebKit/gtk/WebCoreSupport/FrameLoaderClientGtk.cpp |    3 +--
 2 files changed, 15 insertions(+), 2 deletions(-)
Comment 3 Gustavo Noronha (kov) 2009-04-26 17:28:13 PDT
Created attachment 29810 [details]
Check if we should use the fallback content in didFailLoad.

 WebKit/gtk/ChangeLog                               |   14 ++++++++++++++
 WebKit/gtk/WebCoreSupport/FrameLoaderClientGtk.cpp |    6 ++++++
 2 files changed, 20 insertions(+), 0 deletions(-)
Comment 4 Xan Lopez 2009-04-26 22:40:48 PDT
Comment on attachment 29809 [details]
Do not fallback on cancellation and load interruption caused by policy change.

Looks good.
Comment 5 Xan Lopez 2009-04-26 22:44:07 PDT
Comment on attachment 29810 [details]
Check if we should use the fallback content in didFailLoad.

Perhaps you can add the same comment to the signal emission than Jan did, to be consistent. r=me otherwise.
Comment 6 Gustavo Noronha (kov) 2009-04-27 07:54:55 PDT
Landed as r42888, r42889 with Xan's comment addressed.