Re: pre-announce exmh-2.7.2

OK - I fixed up the detached message display and detached folder
display so they use "Exwin_Toplevel" instead of "Widget_Toplevel"
so their saved position is used correctly.  I also got the bindings
and focus right so you can use your keyboard commands when the
keyboard focus is on the scan listing (ftoc) window.

The only glitch, which I cannot easily fix, is that the size is
not remembered, only the position.  To have the size you like
remembered, you need to use the preferences settings to explicitly
set mtextLines, mtextWidth, and ftocLines.

I also cannot prevent you from using the weird, but functional,
configuration of having both a detached message and ftoc display.

There is a new tarball, dated 1/7/2005, and I really need to
be "done" with this for now.

>>>[email protected] said:
 > On Fri, 07 Jan 2005 14:08:28 PST, Brent Welch said:
 > 
 > > I'd hesitate to change the implementation of exwin.tcl for
 > > remembering window locations.  That has worked for years for
 > 
 > I hesitated as well..  thus the "somebody check this" patch. ;)
 > 
 > > all other exmh windows - what is it about the message window
 > > that is different ?  I believe that if you use the "Dismiss"
 > > button in the msg window then the location will be remembered
 > > correctly.  If you want the geometry to be saved when the
 > > user destroys the window, you need a 
 > >   wm protocol .msgframe WM_DELETE_WINDOW {Exwin_Dismiss .msgframe}
 > > hook that calls Exwin_Dismiss
 > 
 > The odd part is that it *is* getting saved:
 > 
 > % grep msgframe .exmh/exmh-defaults 
 > msgframe.geometry:      900x800+200+10
 > *msgframe.position:     +100+124
 > 
 > (and 'msgframe' is also in the saved *exwinPaths line, which is too long
 > to 
 > paste usefully here).
 > 
 > So it looks like the geometry/position isn't *applied* when the window
 > is
 > re-created at next startup.  Which points at Exwin_Toplevel, where the
 > geometry is applied....
 > 
 > As an aside, I suspect that both problems with .msgframe will *also*
 > apply to .ftocframe if that's a toplevel - I suspect that it's just not
 > as heavily used so nobody's noticed on .ftocframe.
 > 
 > > please don't "fix" exin.tcl, OK?
 > > I've added the WM_DELETE_WINDOW hook
 > 
 > Good, that probably should be there regardless..
 > 
 > > Exwin_Dismiss will be perfect because it'll unmap the window.
 > > However, you need something to re-show the menu and position it,
 > > which is the Exwin_Toplevel command.
 > > 
 > > I'll fiddle with this.
 > 
 > OK, I wasn't *positive* that Exwin_Toplevel was safe to re-invoke, and
 > I had *no* idea where to put the check - the place we're currently dying
 > is probably *way* too late...
 > 

--
Brent Welch
Software Architect, Panasas Inc
Delivering the premier storage system for scalable Linux clusters

www.panasas.com
[email protected]