I intend to do some work to cleanup/fix issues with current failing tests in the GTK+ port. I'm opening this bug to track progress, instead of opening one for each thing.
Created attachment 29797 [details] fix timeout in post-goback-same-url.html WebKitTools/ChangeLog | 10 ++++++++++ .../DumpRenderTree/gtk/LayoutTestControllerGtk.cpp | 15 +++++++++++++-- 2 files changed, 23 insertions(+), 2 deletions(-)
Comment on attachment 29797 [details] fix timeout in post-goback-same-url.html This fixes LayoutTestController::queueLoad to "resolve" relative URLs. It's a pitty it seems we have no sane API to do such things in glib, so I did it using some string handling. It fixes the timeout for that test.
Created attachment 29804 [details] fix timeout in post-goback-same-url.html WebKitTools/ChangeLog | 10 ++++++++++ .../DumpRenderTree/gtk/LayoutTestControllerGtk.cpp | 15 +++++++++++++-- 2 files changed, 23 insertions(+), 2 deletions(-)
Comment on attachment 29804 [details] fix timeout in post-goback-same-url.html Fixes the way we figure out the size of the string to init the GString, as Xan pointed out the other method would go wrong with utf8 file names.
Comment on attachment 29804 [details] fix timeout in post-goback-same-url.html Looks good.
I actually found that my method has problems with other tests that had escaped my tests, unfortunately. I cooked a new patch using SoupURI, which thankfully has a way of creating SoupURIs with base/relative.
Created attachment 29811 [details] Use SoupURI to resolve URLs WebKitTools/ChangeLog | 10 ++++++++++ .../DumpRenderTree/gtk/LayoutTestControllerGtk.cpp | 18 ++++++++++-------- 2 files changed, 20 insertions(+), 8 deletions(-)
Comment on attachment 29811 [details] Use SoupURI to resolve URLs Do we need any error-checking in queueLoad? r=me
Landed as r42890, with a null-check for absoluteURI.