This is a small utility to embed in your python applications. It will help you trace what's happening. You can specify what slot in a class you want to record the trace to, kill traces, etc.
You should tweak this code to your liking, but I mostly used it on fairly intensive computing to measure performances about a slice of the code. Kill paths help you limit the depth of the tracing.
You should not use this decorator on multi-threaded applications...
Assuming the following snippet:
from tracer import trace, TraceUtils
@trace(record_all=True)
class foo:
def __init__(self):
pass
def call_me(self):
print "I'm being called"
@trace(slots=('call_too',))
class bar:
def __init__(self):
pass
def call_too(self):
print "call_too"
self.wont_trace()
def wont_trace(self):
print "Don't trace me..."
if __name__ == '__main__':
o = foo()
m = bar()
i = 0
while i < 10:
o.call_me()
m.call_too()
i += 1
TraceUtils.print_trace()You'll get the following trace:
<<>> foo::call_me(1) (0.074104)
<<>> bar::call_too(3) (0.074142)
<<>> foo::call_me(5) (0.074158)
<<>> bar::call_too(7) (0.074169)
<<>> foo::call_me(9) (0.074182)
<<>> bar::call_too(11) (0.074193)
<<>> foo::call_me(13) (0.074206)
<<>> bar::call_too(15) (0.074216)
<<>> foo::call_me(17) (0.074228)
<<>> bar::call_too(19) (0.074239)
<<>> foo::call_me(21) (0.074252)
<<>> bar::call_too(23) (0.074262)
<<>> foo::call_me(25) (0.074274)
<<>> bar::call_too(27) (0.074285)
<<>> foo::call_me(29) (0.074298)
<<>> bar::call_too(31) (0.074308)
<<>> foo::call_me(33) (0.074320)
<<>> bar::call_too(35) (0.074330)
<<>> foo::call_me(37) (0.074344)
<<>> bar::call_too(39) (0.074354)