949bce1...
by
Michael Catanzaro <email address hidden>
Prepare 3.24.5
a63373a...
by
Michael Catanzaro <email address hidden>
header-bar: Don't disable navigation actions when history is cleared
There's no reason to do this. The history service is not needed for the
back/forward list to work properly, so no reason to disable back/forward
after clearing history. And it's disabling the stop/reload action, too,
which doesn't make any sense. Especially since there is no code to ever
enable it again.
a3d68ae...
by
Michael Catanzaro <email address hidden>
Plug leak in bookmarks code *properly*
It's not NULL-safe. I should know this by now....
54568fa...
by
Michael Catanzaro <email address hidden>
Plug leak in bookmarks import code
Discovered by Daniel Boles
03ef4ea...
by
Michael Catanzaro <email address hidden>
json-glib should not be required
It's really only needed for sync, so stop using it unconditionally.
4a7de81...
by
Michael Catanzaro <email address hidden>
Make bookmarks import not crash when there's no Firefox profile
914ed77...
by
Michael Catanzaro <email address hidden>
location-controller: Decode URI in get_location_cb
Otherwise, pressing Escape inside the location entry causes the URI to
become percent-encoded. Oops.
4a940d9...
by
Michael Catanzaro <email address hidden>
search-engine-manager: get_address() must return NULL
The search-engine-dialog relies on get_address() returning NULL to
detect new search engines. It doesn't function properly otherwise. All
code using this function just needs to be prepared to handle a NULL
address.
In contrast, get_address_for_default_engine() really does need to always
return a result. At least, it's more convenient this way, so that
calling code doesn't have to handle that manually. Rename the function
to make this more clear and reduce confusion, as it would be odd to have
two parallel functions function differently in this respect.