Re: [ctwm] Strange effect when upgrading NetBSD and X

2009-06-05 Thread Rhialto
On Wed 13 May 2009 at 21:10:53 +0200, Rhialto wrote:
 And indeed, that fixed it! Really weird. I want to try to find which
 option makes the difference... but I don't have time for that now :-(

Returning to my previous problem: it is indeed the backing store option
that causes the problem. To be precise:

Section Device
# Backingstore causes ctwm to misbehave; it looks mostly like it is
# expecting backing store to work when it doesn't. Various things
# are not re-painted on exposure.
# The option isn't documented in Xorg(1) or savage(1), but it is
# recognised by the server, judged by this but in Xorg.log:
# (**) SAVAGE(0): Option BackingStore
# (**) SAVAGE(0): Backing store enabled
#Option Backingstore

Identifier  Card0
Driver  savage
VendorName  S3 Inc.
BoardName   VT8636A [ProSavage KN133] AGP4X VGA Controller (TwisterK)
BusID   PCI:1:0:0
EndSection

Today I updated my main machine, and I saw the same problem. The
solution was also the same. Finally I tried putting in the Backingstore
option, and the problem reappeared.

Some googling suggests that it is a sort-of-known problem:
http://www.nvnews.net/vbulletin/showthread.php?t=125433

KDE4 crashes : Option BackingStore must be false with Xorg 1.5
Hi,

For people experiencing crashes in KDE4 (e.g. clicking on a konsole
menu), check in your xorg.conf file your options. If you have:

Option BackingStore true

then this option is known to lead to crashes with Xorg 1.5.
Put it to false, it solved the crashes for me.

Shoddy linux-quality programming. Backing store is supposed to work!

-Olaf.
-- 
___ Olaf 'Rhialto' Seibert-- You author it, and I'll reader it.
\X/ rhialto/at/xs4all.nl  -- Cetero censeo authored delendum esse.



Re: [ctwm] Strange effect when upgrading NetBSD and X

2009-05-13 Thread Richard Levitte
I run Debian [unstable] on my laptop, and switched to X.org a long
time ago.  Running 1.6.1 for the moment.

I have seen no changes in ctwm's behavior, at least when it comes to
the 3D effects (I think the only things I noticed were some distinct
font changes ;-)).

Is your ctwm compiled by you, or is it a package that comes with
NetBSD?  Could it be that there's been some configuration changes
somewhere that went unnoticed?

In message 20090512202328.ge...@falu.nl on Tue, 12 May 2009 22:23:28 +0200, 
Rhialto rhia...@falu.nl said:

rhialto I have updated my laptop from NetBSD 4.0 to 5.0, which also means X has
rhialto changed from XFree86 4.5.0 to X.org 1.4.2.
rhialto 
rhialto Now my WorkSpaceManager window and the pop-up menus look incorrect.
rhialto 
rhialto The pop-up menus are missing their 3D highlighting lines around the
rhialto sides.
rhialto 
rhialto The WorkSpaceManager seems to get no expose (redraw) events. The little
rhialto window titles and the 3D highlighting lines along the sides of the
rhialto little windows disappear if you put something on top. I can force a
rhialto redraw by selecting things inside, or simpler, switching to button
rhialto view and back (pressing and releasing Control inside the
rhialto WorkSpaceManager window).
rhialto 
rhialto http://www.falu.nl/~rhialto/ctwm-xorg/win1.png shows what it looks 
like,
rhialto and 
rhialto http://www.falu.nl/~rhialto/ctwm-xorg/win2.png shows the difference on
rhialto my un-updated machine.
rhialto 
rhialto Has anyone seen anything like this?



Re: [ctwm] Strange effect when upgrading NetBSD and X

2009-05-13 Thread Rhialto
On Wed 13 May 2009 at 08:06:10 +0200, Richard Levitte wrote:
 I run Debian [unstable] on my laptop, and switched to X.org a long
 time ago.  Running 1.6.1 for the moment.
 
 I have seen no changes in ctwm's behavior, at least when it comes to
 the 3D effects (I think the only things I noticed were some distinct
 font changes ;-)).
 
 Is your ctwm compiled by you, or is it a package that comes with
 NetBSD?  Could it be that there's been some configuration changes
 somewhere that went unnoticed?

It is a ctwm compiled by myself. First I used my old binary, but nothing
changed when I recompiled with X11R7 libraries.

Using the NetBSD package should not be much different (except that it is
not newer than 3.8a) because I made it myself :-)

With X.org I even use the same XF86Config file as before (I changed some
path names from /usr/X11R6 to /usr/X11R7 but that was all). At some
point in the past, I must have specified a non-standard location for it,
and apparently that is still used; but I forgot where exactly I
specified that :-)

Maybe I should try to have X.org generate its own config file and see if
it makes a difference.

Unfortunately I don't have much time to debug in the next 2 weeks, but I
know I can make ctwm log lots of event-related debug info, so it should
be possible to find out if it actually receives the Expose events or
not. (I did find out in the past that it is better to display those
messages on another display than the one that ctwm is managing!)

-Olaf.
-- 
___ Olaf 'Rhialto' Seibert-- You author it, and I'll reader it.
\X/ rhialto/at/xs4all.nl  -- Cetero censeo authored delendum esse.



[ctwm] Strange effect when upgrading NetBSD and X

2009-05-12 Thread Rhialto
I have updated my laptop from NetBSD 4.0 to 5.0, which also means X has
changed from XFree86 4.5.0 to X.org 1.4.2.

Now my WorkSpaceManager window and the pop-up menus look incorrect.

The pop-up menus are missing their 3D highlighting lines around the
sides.

The WorkSpaceManager seems to get no expose (redraw) events. The little
window titles and the 3D highlighting lines along the sides of the
little windows disappear if you put something on top. I can force a
redraw by selecting things inside, or simpler, switching to button
view and back (pressing and releasing Control inside the
WorkSpaceManager window).

http://www.falu.nl/~rhialto/ctwm-xorg/win1.png shows what it looks like,
and 
http://www.falu.nl/~rhialto/ctwm-xorg/win2.png shows the difference on
my un-updated machine.

Has anyone seen anything like this?

-Olaf.
-- 
___ Olaf 'Rhialto' Seibert-- You author it, and I'll reader it.
\X/ rhialto/at/xs4all.nl  -- Cetero censeo authored delendum esse.