Re: DND in TreeView

[Deng Xiyue <manphiz gmail com>]

José Alburquerque <jaalburquerque cox net> writes:

> On Fri, 2009-03-20 at 07:48 +0800, Deng Xiyue wrote: 
>> José Alburquerque <jaalburquerque cox net> writes:
>> 
>> > On Fri, 2009-03-20 at 00:17 +0800, Deng Xiyue wrote:
>> >> Sorry for the delay in preparing a testcase for being busy in the past
>> >> weeks.  The test program is based on one of the examples with callbacks
>> >> connected to the aforementioned signals, cout the deleted and inserted
>> >> rows to stdout.  By DND from and to the last row the abnormal behavior
>> >> will be observed.  As the implementation of TreeView in GTKMM, I guess
>> >> the same behavior is expected with pure GTK+ code.  But I still like to
>> >> know whether this is the way to handle DND within TreeView before filing
>> >> bug report.  Thanks.
>> >
>> > Is it a lot of C++ code?  Maybe you can post the code and it can be
>> > tested to get a sense of what's happening.
>> >
>> 
>> I have attached the code in my last post, maybe it get stripped from
>> your INBOX?  It can be seen on the mailing list archive[1] though.
>
> Sorry.  Just began using evolution and honestly did not notice the
> attachment.
>
> As far as I can tell, the signals are called.  I've attached your source
> with some output statements in your signal handlers so you can confirm.
> Maybe something you're trying do is not quite working the way you
> expect.  Maybe ensure that the methods you're calling do what they're
> supposed to and that your logic is right.  HTH.
>

Yes, the signal handlers are called in normal case, but the abnormal
behavior will only be noticed when manipulating the last row:

1. When dragging the last row away, *_row_deleted is never emitted.

2. When dragging a row to the point after the current last row,
*_row_inserted erroneously indicates that it is inserted to position 0,
which is the first row, as well as a GTK+ assertion warning.

Please confirm.  Thanks.

Regards,
Deng Xiyue