Escaped newlines were properly handled inside preprocessor directives,
but not otherwise.
Seeing `continue` here suggests the code used to work a long time ago
but some loop refactoring broke it, as now it would not stay in the
loop unless in a preprocessor directive. Or maybe it only ever worked
for preprocessor directives, and the `continue` was superfluous?
Fixes#1370.
When plugin calls plugin_set_key_group() several times for the same
group (when creating keybindings dynamically and needs to reset them),
it crashes with the current code the second time it gets called.
The reason is that group->plugin_keys is an array into which entries of
group->key_items point and when calling
g_ptr_array_set_size(group->key_items, 0);
it calls free_key_binding() for every item - when these items are
deallocated by g_free(group->plugin_keys) previously, calls of
free_key_binding() reference an invalid memory.
Just first resizing group->key_items (and calling free_key_binding() for
its items) and freeing group->plugin_keys afterwards fixes the problem.
Add parfor to Matlab keyword list
Add `parfor` to the list of keywords to be highlighted in Matlab script sources. `parfor` is a Matlab keyword that can be used in place of `for` to achieve parallel processing.
Use a per-line cache to avoid re-computing the offset from the start of
the buffer each time. This dramatically speeds up multiple replacements
on large files.
X-Scintilla-Bug-ID: https://sourceforge.net/p/scintilla/bugs/1910/
On GTK2 and GTK3 < 3.3.6 there is no GtkAccessibleClass::widget_unset()
method, so we can't destroy our accessible object right away. So, to
avoid accessing a destroyed widget, we need to check whether the widget
still exists in the the ScintillaGTKAccessible destructor.
In other methods it's not necessary because the wrapping GObject class
makes sure not to forward other when the widget has been destroyed, but
we still have to destroy the C++ instance no matter what, so the check
has to be on this side.
Fixes#1410.
