Re: [PATCH] Keep history of useful set_filename() calls

[Vincent Legoll <vincent legoll gmail com>]

On Tue, Mar 24, 2009 at 8:33 AM, Kai Willadsen <kai willadsen gmail com> wrote:
> 2009/3/24 Vincent Legoll <vincent legoll gmail com>:
>> Here is a patch that make meld save history of set_filename()
>> calls, which is IMHO a sensible thing to do.
>>
>> The condition of path existence before saving history is debatable
>> but I don't think we should keep the non existent pathes.
>
> I think this would go better in __entry_activate_signal(), otherwise
> manually entered paths won't get saved. Something like:
>
>     def __entry_activate_signal(self, widget, *data):
> +        # Move current to top of history
> +        self.__gentry.prepend_text(self.gtk_entry.get_text())
>         self.emit("activate")
>
> should work.
>
> I know I said before that we should not keep non-existent paths, but I
> think too much of that logic lies outside of the History* widgets to
> be able to do it reliably inside. For example, the os.path.exists()
> check would break with URI input. Maybe for now just leave the check
> out and apply something like the above, which looks like a reasonable
> 80% solution.

I'm OK with this, I'll respin the series this way

-- 
Vincent Legoll