annotate Demo/threads/find.py @ 5817:48b94de85bf3 legacy-trunk

Extensive changes by AMK.
author Guido van Rossum <guido@python.org>
date Thu, 17 Jul 1997 16:29:42 +0000
parents dc5c11177339
children 3c08fb9e134f
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
3874
dc5c11177339 Remove comment "SGI only"
Guido van Rossum <guido@python.org>
parents: 1633
diff changeset
1 # A parallelized "find(1)" using the thread module.
1431
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
2
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
3 # This demonstrates the use of a work queue and worker threads.
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
4 # It really does do more stats/sec when using multiple threads,
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
5 # although the improvement is only about 20-30 percent.
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
6
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
7 # I'm too lazy to write a command line parser for the full find(1)
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
8 # command line syntax, so the predicate it searches for is wired-in,
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
9 # see function selector() below. (It currently searches for files with
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
10 # group or world write permission.)
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
11
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
12 # Usage: parfind.py [-w nworkers] [directory] ...
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
13 # Default nworkers is 4, maximum appears to be 8 (on Irix 4.0.2)
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
14
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
15
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
16 import sys
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
17 import getopt
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
18 import string
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
19 import time
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
20 import os
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
21 from stat import *
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
22 import thread
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
23
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
24
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
25 # Work queue class. Usage:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
26 # wq = WorkQ()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
27 # wq.addwork(func, (arg1, arg2, ...)) # one or more calls
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
28 # wq.run(nworkers)
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
29 # The work is done when wq.run() completes.
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
30 # The function calls executed by the workers may add more work.
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
31 # Don't use keyboard interrupts!
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
32
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
33 class WorkQ:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
34
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
35 # Invariants:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
36
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
37 # - busy and work are only modified when mutex is locked
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
38 # - len(work) is the number of jobs ready to be taken
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
39 # - busy is the number of jobs being done
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
40 # - todo is locked iff there is no work and somebody is busy
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
41
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
42 def __init__(self):
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
43 self.mutex = thread.allocate()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
44 self.todo = thread.allocate()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
45 self.todo.acquire()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
46 self.work = []
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
47 self.busy = 0
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
48
1633
f2602288381a Fixed for stricter argument matching rules
Guido van Rossum <guido@python.org>
parents: 1553
diff changeset
49 def addwork(self, func, args):
f2602288381a Fixed for stricter argument matching rules
Guido van Rossum <guido@python.org>
parents: 1553
diff changeset
50 job = (func, args)
1431
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
51 self.mutex.acquire()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
52 self.work.append(job)
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
53 self.mutex.release()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
54 if len(self.work) == 1:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
55 self.todo.release()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
56
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
57 def _getwork(self):
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
58 self.todo.acquire()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
59 self.mutex.acquire()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
60 if self.busy == 0 and len(self.work) == 0:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
61 self.mutex.release()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
62 self.todo.release()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
63 return None
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
64 job = self.work[0]
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
65 del self.work[0]
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
66 self.busy = self.busy + 1
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
67 self.mutex.release()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
68 if len(self.work) > 0:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
69 self.todo.release()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
70 return job
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
71
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
72 def _donework(self):
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
73 self.mutex.acquire()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
74 self.busy = self.busy - 1
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
75 if self.busy == 0 and len(self.work) == 0:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
76 self.todo.release()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
77 self.mutex.release()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
78
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
79 def _worker(self):
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
80 while 1:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
81 job = self._getwork()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
82 if not job:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
83 break
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
84 func, args = job
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
85 apply(func, args)
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
86 self._donework()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
87
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
88 def run(self, nworkers):
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
89 if not self.work:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
90 return # Nothing to do
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
91 for i in range(nworkers-1):
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
92 thread.start_new(self._worker, ())
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
93 self._worker()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
94 self.todo.acquire()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
95
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
96
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
97 # Main program
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
98
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
99 def main():
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
100 nworkers = 4
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
101 opts, args = getopt.getopt(sys.argv[1:], '-w:')
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
102 for opt, arg in opts:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
103 if opt == '-w':
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
104 nworkers = string.atoi(arg)
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
105 if not args:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
106 args = [os.curdir]
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
107
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
108 wq = WorkQ()
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
109 for dir in args:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
110 wq.addwork(find, (dir, selector, wq))
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
111
1553
b70f08907546 Use time.time() instead of time.millitimer()
Guido van Rossum <guido@python.org>
parents: 1431
diff changeset
112 t1 = time.time()
1431
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
113 wq.run(nworkers)
1553
b70f08907546 Use time.time() instead of time.millitimer()
Guido van Rossum <guido@python.org>
parents: 1431
diff changeset
114 t2 = time.time()
1431
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
115
1553
b70f08907546 Use time.time() instead of time.millitimer()
Guido van Rossum <guido@python.org>
parents: 1431
diff changeset
116 sys.stderr.write('Total time ' + `t2-t1` + ' sec.\n')
1431
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
117
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
118
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
119 # The predicate -- defines what files we look for.
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
120 # Feel free to change this to suit your purpose
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
121
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
122 def selector(dir, name, fullname, stat):
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
123 # Look for group or world writable files
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
124 return (stat[ST_MODE] & 0022) != 0
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
125
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
126
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
127 # The find procedure -- calls wq.addwork() for subdirectories
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
128
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
129 def find(dir, pred, wq):
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
130 try:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
131 names = os.listdir(dir)
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
132 except os.error, msg:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
133 print `dir`, ':', msg
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
134 return
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
135 for name in names:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
136 if name not in (os.curdir, os.pardir):
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
137 fullname = os.path.join(dir, name)
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
138 try:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
139 stat = os.lstat(fullname)
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
140 except os.error, msg:
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
141 print `fullname`, ':', msg
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
142 continue
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
143 if pred(dir, name, fullname, stat):
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
144 print fullname
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
145 if S_ISDIR(stat[ST_MODE]):
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
146 if not os.path.ismount(fullname):
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
147 wq.addwork(find, (fullname, pred, wq))
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
148
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
149
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
150 # Call the main program
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
151
d7168cc2be07 Initial revision
Guido van Rossum <guido@python.org>
parents:
diff changeset
152 main()