Matthew D. Fuller wrote:
>
> On Thu, Apr 21, 2016 at 11:49:12PM +0100 I heard the voice of
> Aaron Sloman, and lo! it spake thus:
> >
> > I have now restarted, using this. It all seems to work, but with a
> > minor problem that I noticed in an earlier version running on this
> > machine:
> >
> > After restarting the new ctwm on SL or invoking a menu option to
> > re-read .ctwmrc, menus come up initially without the menu items
> > showing, only the menu name at the top and the menu frame.
>
> TLDR: Try setting NoBackingStore.
Thanks for the suggestion. I'll try that when next in my office.
Apologies for not getting round to your previously suggested test using
remote access. I had just been too swamped with deadlines and travel and
postponed the experiment.
> OK, I got an X server running on a CentOS 6 VM, and I get what looks
> like similar (though not identical) symptoms. What _I_ wind up with
> is the background of each line not being set (just left black) until
> the mouse goes over it.
That's what I found, but perhaps I misdescribed it.
> In point of fact, I've wondered in passing whether we should just make
> that the default. X servers paint awful fast now, and with the number
> of people using compositors, there's no point to it anyway. I'm more
> than half tempted to completely remove all use of it wholesale; I
> think the whole idea is pretty much a remnant of 80's hardware
> limitations. Any thoughts?
I don't know enough to have an informed opinion on that but do recall
deciding a few years ago that optimising code for refreshing the xterm
window used in a text editor was no longer necessary because refreshing the
whole window was so fast, even over a network connection.
Anyhow, I'll let you know later if that suggestion fixes the problem.
Curiously I've found in my .ctwmrc on a desktop pc running fedora 22 at
home that this is in my .ctwmrc
## ADDED 1 Dec 2008
NoBackingStore
But removing it doesn't seem to make any difference on that machine.
Perhaps there was a time when the problem now on SL existed on Fedora, and
settingthat variable fixed it. I have no recollection of adding that entry!
Thanks.
Aaron
