Make sure the SMA window has really at least 2 seconds

This could cause problems on very unstable connections.  When next()
wasn't called for a long time, and then it was called very frequently
(so time() returned 4x the same value) we were clean()ing the
window.dequeue so much so we got "last" equal to "deque[0][0]",
therefore counter/(last - deque[0][0]) could cause a division by zero.

In theory, this would also solve some temporary situations when
subsequent calls to time() did not return consecutive values:

    >>> a = time() ; b = time() ; a > b
    True

Resolves: rhbz#2103093
Version: 1.6-4

0001-fixup-moving-average-window.patch

@@ -77,7 +77,7 @@ index b434b30..71b2039 100644
 +        if self.max_items:
 +            while len(self.deque) > self.max_items:
 +                self.pop()
-+        while len(self.deque) > 2 and self.last - self.deque[0][0] > float(self.max_seconds):
++        while len(self.deque) > 2 and self.last - self.deque[1][0] > float(self.max_seconds):
 +            self.pop()
 +
 +    def next(self, n, t):

python-progress.spec

@@ -20,7 +20,7 @@
 
 Name:           python-%{pypi_name}
 Version:        1.6
-Release:        3%{?dist}
+Release:        4%{?dist}
 Summary:        Easy to use progress bars
 
 License:        ISC
@@ -98,6 +98,9 @@ rm -rf %{pypi_name}.egg-info
 
 
 %changelog
+* Sat Jul 02 2022 Pavel Raiskup <praiskup@redhat.com> - 1.6-4
+- fix for very frequent next() call, rhbz#2103093
+
 * Mon Jun 13 2022 Python Maint <python-maint@redhat.com> - 1.6-3
 - Rebuilt for Python 3.11