Re: [lmi] UpdateUI problems in main_wx.cpp

[Greg Chicares]

On 2008-08-26 23:08Z, Vadim Zeitlin wrote:
> On Sat, 23 Aug 2008 02:54:01 +0000 Greg Chicares <address@hidden> wrote:
[...]
> GC> Furthermore, all of this seems to be independent of your patch here:
> GC>   http://savannah.nongnu.org/bugs/?17757
> 
>  Yes, that patch is completely orthogonal.

And actually now that specific patch is no longer needed, because it was
better to remove the wxID_SAVE handler from lmi and just defer to the
builtin wx doc-view handler, as previously discussed.

> It does however also illustrate
> an important idea which it took me some time to realize so it might be
> useful if I explained it here:

[Snip illuminating discussion.]

>  To wrap it up, the patch mentioned above is an application of this idea to
> this particular case but it might be useful elsewhere in LMI as well.
> Basically anywhere where some condition is checked at the beginning of an
> event handler before it does anything, you can also check this condition
> somewhere else and only Connect() the handler when it's true (and sometimes
> also Disconnect() it when the condition becomes false).

Agreed. I don't notice any instance of this usage in lmi now, but it's
good to keep in mind for the future.

> GC> But I plan not to change HEAD for about one week, lest I destabilize
> GC> the 2008-09-01 release.
> 
>  Understood.

Meanwhile, some instability has intruded from outside the universe we
can control, but today at last I've been able to pick this up again.
Changes committed to cvs in [20080917T1654Z, 20080917T1729Z] should
resolve the EVT_UPDATE_UI(wxID_SAVE...) issue and defect report #17757
as well; I'll mark that report as ready for formal testing, with a
comment referring to this email.