<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
<p>Well, that could explain all the wierd stuff I've been seeing
leading me to come up with that very simple example. FWIW,
lispworks clim2 works fine.</p>
<p>Thanks,</p>
<div class="moz-cite-prefix">On 1/3/2021 11:01 AM, Daniel Kochmański
wrote:<br>
</div>
<blockquote type="cite"
cite="mid:zp3Gs7z2-2EENsNqo6-I_hZokQqi2l0_4Hbv0gwOoMun2a1EOycsqe02taTXJzv4fM5IRRPkyWlds16eU6FdNTLiVyWzj-LGWVeMV_hHWjk=@turtleware.eu">
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
<br>
<br>
<br>
Accidently send to the sender, not the mailing list, forwarding<br>
-------- Oryginalna wiadomość --------<br>
3 sty 2021, 16:52, Daniel Kochmański napisał(a):
<blockquote class="protonmail_quote"><br>
<p dir="ltr">If the beer is at stake, I shall take a guess!</p>
<p dir="ltr">When 1+ is used, we draw beyond the sheet region.
That means, that *if*<br>
we call change-space-requirements it will cause a call to
resize-sheet,<br>
which in turn will: a) clear the screen, b) call
redisplay-frame-pane.</p>
<p dir="ltr">Since the sheet region has changed, redisplaying
the frame pane will<br>
expand it even further, so we go back to the *if* part.</p>
<p dir="ltr">Now, regarding the *if*, we have a bogus method:</p>
<p dir="ltr">(defmethod stream-read-gesture :before ((stream
clim-stream-pane)<br>
&key timeout peek-p<br>
input-wait-test<br>
input-wait-handler<br>
pointer-button-press-handler)<br>
(declare (ignore timeout peek-p input-wait-test
input-wait-handler<br>
pointer-button-press-handler))<br>
(force-output stream)<br>
;; make the output visible<br>
(let ((w (bounding-rectangle-max-x (stream-output-history
stream)))<br>
(h (bounding-rectangle-max-y (stream-output-history stream))))<br>
(unless (region-contains-region-p (sheet-region stream)<br>
(make-rectangle* 0 0 w h))<br>
(change-space-requirements stream)<br>
(finish-output stream)<br>
(redisplay-frame-pane *application-frame* stream))))</p>
<p dir="ltr">that causes bogus redisplays before each gesture
read, so when we move<br>
the pointer over the window, we enter the loop:</p>
<p dir="ltr">1. display (and expand the output history)<br>
2. move the pointer<br>
3. change space requirements (and resize)<br>
4. goto 1</p>
<p dir="ltr">as a bonus point we have one more unnecessary
redisplay.</p>
<p dir="ltr">The fun part is that I have changes stashed that
remove this method.</p>
<p dir="ltr">All best,<br>
Daniel</p>
<p dir="ltr">--<br>
Daniel Kochmański ;; aka jackdaniel | Przemyśl, Poland<br>
TurtleWare - Daniel Kochmański | <a
href="http://www.turtleware.eu" moz-do-not-send="true">www.turtleware.eu</a></p>
<p dir="ltr">"Be the change that you wish to see in the world."
- Mahatma Gandhi</p>
<p dir="ltr">‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐<br>
On Sunday, 3 January 2021 16:18, Paul Werkowski
<a class="moz-txt-link-rfc2396E" href="mailto:pw@snoopy.qozzy.com"><pw@snoopy.qozzy.com></a> wrote:</p>
<p dir="ltr">> Run the frame as is. Grab the lower corner
with mouse pointer. Move it<br>
> around.<br>
><br>
> Then change either or both of the (1+ ...) forms to (1-
...) and see<br>
> what happens.<br>
><br>
> A free virtual beer to whomever solves it.</p>
</blockquote>
</blockquote>
</body>
</html>