[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#238918: xserver-xfree86: resize events mismanaged on a compaq armada - 16mb ram pentium (60?)

Package: xserver-xfree86
Version: 4.2.1-16
Severity: normal

this one is fascinating: you can actually watch the numlock+capslock
keys flash on and off during the thirty seconds it takes to start up

the issue is that the machine is running so incredibly slowly that you
can actually see things like resizing and window creation occurring
right under your nose, on-screen.

for example, starting the editor "ted", you first get a window of
size 5x10, then a couple of seconds later when the menu is added by
the window manager, you get a window of size 30 x 20, and then a
couple of seconds after _that_ you get a window of size approx 300 x
300 as the ted program adds its File etc menu, and then it goes to
full size.

the bug that i am raising is with xbomb, which, on resize, actually
manages to kerchunk back to the ORIGINAL size after some strange
x and y refactoring.

out of about eight attempts to resize the xbomb package, only two
actually succeeded in resizes.  all other attempts resulted in first
the window being resized to one of the axes, followed by the other
one, followed by the window collapsing back to its original size.

i believe the issue to be that the X-server is running faster, and
at a higher priority, than the applications, and that the drawing event
queue is consequently distracting things.

if there is, in X, some sort of draw timing delay deliberately
introduced in order to avoid the above issues, it clearly isn't
working, and clearly isn't long enough!!!

i am, however, quite impressed that switching to console and back
again, and also suspend, works fine!  i thought at first, after the
first minute, and about the 20th flashing on a 1s on, 2s off duty cycle
of the caps and numlock keys, that X had crashed.  after about 
two minutes, X finally decided to spring into life.  after another
thirty seconds, i had the applications back, too.

paint drying isn't interesting, but watching X clunk along on such
an abominably slow machine (and actually work) is bizarre and horribly


-- System Information:
Debian Release: testing/unstable
Architecture: i386
Kernel: Linux highfield 2.4.23-1-686 #1 Sun Nov 30 20:51:10 EST 2003 i686

Versions of packages xserver-xfree86 depends on:
ii  debconf                     1.4.7        Debian configuration management sy
ii  libc6                       2.3.2.ds1-11 GNU C Library: Shared libraries an
ii  xserver-common              4.2.1-15     files and utilities common to all 
ii  zlib1g                      1:1.2.1-3    compression library - runtime

-- debconf information excluded

Reply to: