##// END OF EJS Templates
Merge pull request #1155 from minrk/postexec...
Merge pull request #1155 from minrk/postexec Be less aggressive in de-registering failing post-execute functions so that normal errors don't disable them permanently. Two changes: 1. Don't unregister failing post-exec callbacks automatically. Instead, print a message regarding the failure, pointing to new `disable_failing_post_execute` trait for skipping failing callbacks. - When this flag is False (the default), failing callbacks will continue to be called. - When True, behavior is unchanged from previous, where callbacks are only allowed to fail once. 2. protect `flush_figures()` post-exec function from user error. Invalid matplotlib data may raise inside `print_figure()`. flush_figures() is a post-exec function, so user errors should not raise. Instead, call `get_ipython().showtraceback()` if called from IPython, raising as before otherwise. closes #1154

File last commit:

r4872:34c10438
r5737:2c683b72 merge
Show More
ipy_server.py
37 lines | 1023 B | text/x-python | PythonLexer
""" Simple TCP socket server that executes statements in IPython instance.
Usage:
import ipy_server
ipy_server.serve_thread(16455)
Now, to execute the statements in this ipython instance, open a TCP socket
(port 16455), write out the statements, and close the socket.
You can use e.g. "telnet localhost 16455" or a script to do this.
This is a bit like 'M-x server-start" or gnuserv in the emacs world.
"""
from IPython.core import ipapi
ip = ipapi.get()
import SocketServer
# user-accessible port
PORT = 8099
class IPythonRequestHandler(SocketServer.StreamRequestHandler):
def handle(self):
#print "connection from", self.client_address
inp = self.rfile.read().replace('\r\n','\n')
#print "Execute",inp
ip.runlines(inp)
def serve(port = PORT):
server = SocketServer.TCPServer(("", port), IPythonRequestHandler)
print "ipy_server on TCP port", port
server.serve_forever()
def serve_thread(port = PORT):
import thread
thread.start_new_thread(serve, (port,))