When we check start & end time selector's sensitivity before
notifying the property change, the check is overriden by the
binding.
Fix that by performing the checks after the notification of the
property change.
This commit sets up the edit dialog from the window's ui
resource file, in a series of commits that tries to move
all C code to the UI definitions.
Also, GcalManager is a full GObject class, not a simple
pointer. As such, update the property spec to be of type
object, not pointer.
By using these properties, we gain the possibility to
bind them to the widgets and save some needless code.
This commit adds GcalEditDialog::writable and ::manager
properties. Other needed properties will be eventually
added in the subsequent commits.
This commit updates GcalEditDialog's code for
it to be a final class. As such, we don't need
the GcalEditDialogPrivate type - we can simply
move our data to the GcalEditDialog struct.
Because we're using a simple (gint) cast to round y value of
the clicked point, a negative y was rounding to a valid cell
position and showing the popover even when the user clicked
on the blank cells before the actual month view.
Fix that by properly checking if y is negative.
https://bugzilla.gnome.org/show_bug.cgi?id=755910
Latest GNOME applications follow a clear style of
searching and displaying results, with a no results
page.
To create a coherent usage between GNOME applications,
add the No results page to search popover.
Two things changes:
a) the actual crash cause in on_client_readonly_changed(), to test whether
the 'unit' is valid at all before dereferencing it
b) the callback doesn't seem to be disconnected at all, thus added a disconnect
of it on the source removal as well
The b) might be enough, but just to be sure I included also a).
https://bugzilla.gnome.org/show_bug.cgi?id=751211
For this to be made, some significant changes on the inners
of GcalMonthView had to be done. Later on, I'll do a complete
rewrite of it, deprecating the use of 'sw' and 'k'.
As stated on the bug below Calendar crashes in some locales because of the
parsing of the string representing the date. Since parsing this string can be
very tricky, we decided to make a simple translatable string with the format it
is needed and th translators can set the proper order according to his locale.
The translatable string goes as "%d/%m/%y" being %d the representation of the
day, %m the representation of the month and, %y, the year. The translators
should place it in the correct order for its locale, for instance en_US, should
place it: "%m/%d/%y"
The validation of the translated string is lacking some strength. It could use
some improving.
Bug: https://bugzilla.redhat.com/show_bug.cgi?id=1226823
