<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">Le 19/09/2018 à 12:25, Samuel Gougeon a
écrit :<br>
</div>
<blockquote cite="mid:9a57ac26-b16e-7ce0-4a46-37a30cd6b6ec@free.fr"
type="cite">Le 19/09/2018 à 12:02, Stéphane Mottelet a écrit :
<br>
<blockquote type="cite">Le 19/09/2018 à 11:46, Samuel Gougeon a
écrit :
<br>
.../...
<br>
<blockquote type="cite">
<br>
At first sight, Octave's result does not look more consistent
than Scilab's one.
<br>
But following its own ceiling/flooring rules, yet it is
consistent.
<br>
</blockquote>
What is puzzling is that Scilab implements a *mix* of rules
comming from different software. I am wondering about the true
reason:
<br>
<br>
Scilab:
<br>
<br>
--> int8(-128)/int8(0)
<br>
ans =
<br>
<br>
-128
<br>
</blockquote>
<br>
This is in Scilab 6. In Scilab 6, int8(-%inf) has been set to the
int8 floor, and int8(%inf) to the int8 ceil. It is more consistent
than the 5.5 behavior (and is now documented in the 6.0 branch). <br>
</blockquote>
<br>
Here is the Scilab 6.0.0 changelog in the int8 ... iuint64 help page
for 6.0.2:<br>
<img src="cid:part1.C9452C5E.A1D86AB7@free.fr" alt="" height="160"
width="276"><br>
<br>
Would you prefer fixing int32(%inf) in order to keep all answers to
0: this would avoid any saturating behavior for the infinite values
as for finite ones ; or these new settings?<br>
<br>
<br>
</body>
</html>