michael@0: TODO
michael@0: ====
michael@0: 
michael@0: A collection of ideas and notes about stuff to implement in future versions.
michael@0: "#NNN" occurrences refer to bug tracker issues at:
michael@0: https://code.google.com/p/psutil/issues/list
michael@0: 
michael@0: 
michael@0: HIGHER PRIORITY
michael@0: ===============
michael@0: 
michael@0:  * #387: system-wide connections (netstat).
michael@0:  * OpenBSD support.
michael@0: 
michael@0:  * #371: CPU temperature (apparently OSX and Linux only; on Linux it requires
michael@0:    lm-sensors lib).
michael@0: 
michael@0:  * #250: net ifaces speed.
michael@0: 
michael@0:  * (Linux) resource limit get/set - see man prlimit.
michael@0: 
michael@0:  * Process.name on Windows is slow:
michael@0:    http://stackoverflow.com/questions/6587036/
michael@0: 
michael@0:  * Windows binary for Python 3.3 64-bit.
michael@0: 
michael@0:  * #269: expose network ifaces RX/TW queues.
michael@0: 
michael@0: 
michael@0: LOWER PRIORITY
michael@0: ==============
michael@0: 
michael@0:  * #355: Android support.
michael@0: 
michael@0:  * #276: GNU/Hurd support.
michael@0: 
michael@0:  * NetBSD support?
michael@0: 
michael@0:  * DranflyBSD support?
michael@0: 
michael@0:  * AIX support?
michael@0: 
michael@0:  * examples/pidof.py (same as 'pidof' cli tool)
michael@0: 
michael@0:  * examples/pstree.py (same as 'pstree' cli tool)
michael@0:     * get_threads() should also return thread names in order to implement it
michael@0: 
michael@0:  * examples/taskmgr-gui.py (using tk).
michael@0: 
michael@0:  * system-wide # open file descriptors:
michael@0:     * https://jira.hyperic.com/browse/SIGAR-30
michael@0:     * http://www.netadmintools.com/part295.html
michael@0: 
michael@0:  * Number of system threads.
michael@0:     * Windows: http://msdn.microsoft.com/en-us/library/windows/desktop/ms684824(v=vs.85).aspx
michael@0: 
michael@0:  * #357: what CPU a process is on.
michael@0: 
michael@0:  * thread names:
michael@0:     * https://code.google.com/p/plcrashreporter/issues/detail?id=65
michael@0: 
michael@0: 
michael@0: DEBATABLE
michael@0: =========
michael@0: 
michael@0:  * [Linux]: process cgroups (http://en.wikipedia.org/wiki/Cgroups). They look
michael@0:    similar to prlimit() in terms of functionality but uglier (they should allow
michael@0:    limiting per-process network IO resources though, which is great). Needs
michael@0:    further reading.
michael@0: 
michael@0:  * cpu_percent(): current default interval is 0.1 so that by default it will
michael@0:    produce a meaningful value. It represents a trap in case the user iterates
michael@0:    over multiple processes though, as it introduces a big slowdown.
michael@0:    Should it default to 0.0?
michael@0: 
michael@0:  * Rename connection ntuple's fields 'local_address', 'remote_address' to
michael@0:    'laddr', 'raddr'  (note in accordance with http://bugs.python.org/issue17675)
michael@0: 
michael@0:  * Process per-cpus percent (XXX Windows only?), see:
michael@0:    https://groups.google.com/forum/?fromgroups#!topic/psutil/ErrKTxAbu50
michael@0: 
michael@0:  * Should we expose OS constants (psutil.WINDOWS, psutil.OSX etc.)?
michael@0: 
michael@0:  * Python 3.3. exposed different sched.h functions:
michael@0:    http://docs.python.org/dev/whatsnew/3.3.html#os
michael@0:    http://bugs.python.org/issue12655
michael@0:    http://docs.python.org/dev/library/os.html#interface-to-the-scheduler
michael@0:    It might be worth to take a look and figure out whether we can include some
michael@0:    of those in psutil.
michael@0:    Also, we can probably reimplement wait_pid() on POSIX which is currently
michael@0:    implemented as a busy-loop.
michael@0: 
michael@0:  * Certain systems (XXX figure out which ones exactly) provide CPU times about
michael@0:    process children. On those systems Process.get_cpu_times() might return
michael@0:    a (user, system, user_children, system_children) ntuple.