Add *new_only* parameter for process_iter()

CREDITS

@@ -656,3 +656,7 @@ I: 1646
 N: Javad Karabi
 W: https://github.com/karabijavad
 I: 1648
+
+N: Mike Hommey
+W: https://github.com/glandium
+I: 1665

HISTORY.rst

@@ -1,6 +1,6 @@
 *Bug tracker at https://github.com/giampaolo/psutil/issues*
 
-5.6.8 (unreleased)
+5.7.0 (unreleased)
 ==================
 
 XXXX-XX-XX
@@ -12,6 +12,7 @@ XXXX-XX-XX
   directory for additional data.  (patch by Javad Karabi)
 - 1652_: [Windows] dropped support for Windows XP and Windows Server 2003.
   Minimum supported Windows version now is Windows Vista.
+- 1667_: added process_iter(new_only=True) parameter.
 
 **Bug fixes**
 
@@ -22,6 +23,8 @@ XXXX-XX-XX
   current user and os.getpid().
 - 1660_: [Windows] Process.open_files() complete rewrite + check of errors.
 - 1662_: [Windows] process exe() may raise WinError 0.
+- 1665_: [Linux] disk_io_counters() does not take into account extra fields
+  added to recent kernels.  (patch by Mike Hommey)
 
 5.6.7
 =====

docs/index.rst

@@ -855,7 +855,7 @@ Functions
   .. versionchanged::
     5.6.0 PIDs are returned in sorted order
 
-.. function:: process_iter(attrs=None, ad_value=None)
+.. function:: process_iter(attrs=None, ad_value=None, new_only=False)
 
   Return an iterator yielding a :class:`Process` class instance for all running
   processes on the local machine.
@@ -871,24 +871,10 @@ Functions
   the resulting dict is stored as a ``info`` attribute which is attached to the
   returned :class:`Process`  instances.
   If *attrs* is an empty list it will retrieve all process info (slow).
+  If *new_only* is true this function will yield only new processes which
+  appeared since the last time it was called.
   Example usage::
 
-    >>> import psutil
-    >>> for proc in psutil.process_iter():
-    ...     try:
-    ...         pinfo = proc.as_dict(attrs=['pid', 'name', 'username'])
-    ...     except psutil.NoSuchProcess:
-    ...         pass
-    ...     else:
-    ...         print(pinfo)
-    ...
-    {'name': 'systemd', 'pid': 1, 'username': 'root'}
-    {'name': 'kthreadd', 'pid': 2, 'username': 'root'}
-    {'name': 'ksoftirqd/0', 'pid': 3, 'username': 'root'}
-    ...
-
-  More compact version using *attrs* parameter::
-
     >>> import psutil
     >>> for proc in psutil.process_iter(attrs=['pid', 'name', 'username']):
     ...     print(proc.info)
@@ -909,19 +895,28 @@ Functions
      3: {'name': 'ksoftirqd/0', 'username': 'root'},
      ...}
 
-  Example showing how to filter processes by name::
+  Example showing how to filter processes by name (see also
+  `process filtering <#filtering-and-sorting-processes>`__ section for more
+  examples)::
 
     >>> import psutil
     >>> [p.info for p in psutil.process_iter(attrs=['pid', 'name']) if 'python' in p.info['name']]
     [{'name': 'python3', 'pid': 21947},
      {'name': 'python', 'pid': 23835}]
 
-  See also `process filtering <#filtering-and-sorting-processes>`__ section for
-  more examples.
+  Get new processes only (since last call)::
+
+    >>> import psutil
+    >>> for proc in psutil.process_iter(attrs=['pid', 'name'], new_only=True):
+    ...     print(proc.info)
+    ...
 
   .. versionchanged::
     5.3.0 added "attrs" and "ad_value" parameters.
 
+  .. versionchanged::
+    5.7.0 added "new_only" parameter.
+
 .. function:: pid_exists(pid)
 
   Check whether the given PID exists in the current process list. This is

psutil/__init__.py

@@ -227,7 +227,7 @@
 
 __all__.extend(_psplatform.__extra__all__)
 __author__ = "Giampaolo Rodola'"
-__version__ = "5.6.8"
+__version__ = "5.7.0"
 version_info = tuple([int(num) for num in __version__.split('.')])
 
 _timer = getattr(time, 'monotonic', time.time)
@@ -1407,7 +1407,7 @@ def pid_exists(pid):
 _lock = threading.Lock()
 
 
-def process_iter(attrs=None, ad_value=None):
+def process_iter(attrs=None, ad_value=None, new_only=False):
     """Return a generator yielding a Process instance for all
     running processes.
 
@@ -1427,6 +1427,8 @@ def process_iter(attrs=None, ad_value=None):
     to returned Process instance.
     If *attrs* is an empty list it will retrieve all process info
     (slow).
+    If *new_only* is true this function will yield only new processes
+    which appeared since the last time it was called.
     """
     def add(pid):
         proc = Process(pid)
@@ -1448,8 +1450,10 @@ def remove(pid):
         remove(pid)
 
     with _lock:
-        ls = sorted(list(_pmap.items()) +
-                    list(dict.fromkeys(new_pids).items()))
+        ls = list(dict.fromkeys(new_pids).items())
+        if not new_only:
+            ls += list(_pmap.items())
+        ls.sort()
 
     for pid, proc in ls:
         try:

psutil/tests/test_system.py

@@ -85,7 +85,7 @@ def test_process_iter(self):
             with self.assertRaises(psutil.AccessDenied):
                 list(psutil.process_iter())
 
-    def test_prcess_iter_w_params(self):
+    def test_prcess_iter_w_attrs(self):
         for p in psutil.process_iter(attrs=['pid']):
             self.assertEqual(list(p.info.keys()), ['pid'])
         with self.assertRaises(ValueError):
@@ -105,6 +105,22 @@ def test_prcess_iter_w_params(self):
                 self.assertGreaterEqual(p.info['pid'], 0)
             assert m.called
 
+    def test_process_iter_new_only(self):
+        ls1 = list(psutil.process_iter(attrs=['pid']))
+        ls2 = list(psutil.process_iter(attrs=['pid'], new_only=True))
+        self.assertGreater(len(ls1), len(ls2))
+        # assume no more than 3 new processes were created in the meantime
+        self.assertIn(len(ls2), [0, 1, 2, 3, 4, 5])
+
+        sproc = get_test_subprocess()
+        ls = list(psutil.process_iter(attrs=['pid'], new_only=True))
+        self.assertIn(len(ls2), [0, 1, 2, 3, 4, 5])
+        for p in ls:
+            if p.pid == sproc.pid:
+                break
+        else:
+            self.fail("subprocess not found")
+
     def test_wait_procs(self):
         def callback(p):
             pids.append(p.pid)