Lisp HUG Maillist Archive

high load measure with LW5

Is linux 2.6 buggy wrt. measuring load, or is it something in the way
LW 5 threads works that will make the load appear artificially high?

Example:

top - 15:38:27 up 150 days, 13:49,  3 users,  load average: 13.44, 10.57, 8.72
Tasks: 123 total,   1 running, 122 sleeping,   0 stopped,   0 zombie
Cpu(s): 12.1% us,  7.5% sy,  0.0% ni, 80.1% id,  0.0% wa,  0.0% hi,  0.2% si

This is a dual-dual Opteron box which is not really heavy loaded at
all (notice that top says it's 80.1% idle!), but since I started
running a LW5-based server on it, it's load as measured by uptime or
top is skyrocketing several times a day.

(The LW5-based server works very well - it is more responsive and
 spends far less time in global GC than a similar LW4-based server on
 the same machine)
-- 
  (espen)

Unexpected result from rich-text-pane-operation :get-selection

I have an instance of a rich-text-pane which is displayed with some text in 
it.

If I do:
REPL> (apply-in-pane-process x #'rich-text-pane-operation (slot-value x 'rich) 
:set-selection 0 20)

I see the selection in the pane as I would expect.

Then if I do:
REPL> (apply-in-pane-process x #'rich-text-pane-operation (slot-value x 'rich) 
:get-selection)
--> NIL

I thought I would receive the bounding indices of the selection.  Is there 
something else I need to do in order to get those?  Thank you.

Mitch
Updated at: 2020-12-10 08:47 UTC