cpython/Tools/scripts/objgraph.py
Thomas Wouters 49fd7fa443 Merge p3yk branch with the trunk up to revision 45595. This breaks a fair
number of tests, all because of the codecs/_multibytecodecs issue described
here (it's not a Py3K issue, just something Py3K discovers):
http://mail.python.org/pipermail/python-dev/2006-April/064051.html

Hye-Shik Chang promised to look for a fix, so no need to fix it here. The
tests that are expected to break are:

test_codecencodings_cn
test_codecencodings_hk
test_codecencodings_jp
test_codecencodings_kr
test_codecencodings_tw
test_codecs
test_multibytecodec

This merge fixes an actual test failure (test_weakref) in this branch,
though, so I believe merging is the right thing to do anyway.
2006-04-21 10:40:58 +00:00

216 lines
5.9 KiB
Python
Executable File

#! /usr/bin/env python
# objgraph
#
# Read "nm -o" input (on IRIX: "nm -Bo") of a set of libraries or modules
# and print various interesting listings, such as:
#
# - which names are used but not defined in the set (and used where),
# - which names are defined in the set (and where),
# - which modules use which other modules,
# - which modules are used by which other modules.
#
# Usage: objgraph [-cdu] [file] ...
# -c: print callers per objectfile
# -d: print callees per objectfile
# -u: print usage of undefined symbols
# If none of -cdu is specified, all are assumed.
# Use "nm -o" to generate the input (on IRIX: "nm -Bo"),
# e.g.: nm -o /lib/libc.a | objgraph
import sys
import os
import getopt
import re
# Types of symbols.
#
definitions = 'TRGDSBAEC'
externals = 'UV'
ignore = 'Nntrgdsbavuc'
# Regular expression to parse "nm -o" output.
#
matcher = re.compile('(.*):\t?........ (.) (.*)$')
# Store "item" in "dict" under "key".
# The dictionary maps keys to lists of items.
# If there is no list for the key yet, it is created.
#
def store(dict, key, item):
if dict.has_key(key):
dict[key].append(item)
else:
dict[key] = [item]
# Return a flattened version of a list of strings: the concatenation
# of its elements with intervening spaces.
#
def flat(list):
s = ''
for item in list:
s = s + ' ' + item
return s[1:]
# Global variables mapping defined/undefined names to files and back.
#
file2undef = {}
def2file = {}
file2def = {}
undef2file = {}
# Read one input file and merge the data into the tables.
# Argument is an open file.
#
def readinput(fp):
while 1:
s = fp.readline()
if not s:
break
# If you get any output from this line,
# it is probably caused by an unexpected input line:
if matcher.search(s) < 0: s; continue # Shouldn't happen
(ra, rb), (r1a, r1b), (r2a, r2b), (r3a, r3b) = matcher.regs[:4]
fn, name, type = s[r1a:r1b], s[r3a:r3b], s[r2a:r2b]
if type in definitions:
store(def2file, name, fn)
store(file2def, fn, name)
elif type in externals:
store(file2undef, fn, name)
store(undef2file, name, fn)
elif not type in ignore:
print fn + ':' + name + ': unknown type ' + type
# Print all names that were undefined in some module and where they are
# defined.
#
def printcallee():
flist = file2undef.keys()
flist.sort()
for filename in flist:
print filename + ':'
elist = file2undef[filename]
elist.sort()
for ext in elist:
if len(ext) >= 8:
tabs = '\t'
else:
tabs = '\t\t'
if not def2file.has_key(ext):
print '\t' + ext + tabs + ' *undefined'
else:
print '\t' + ext + tabs + flat(def2file[ext])
# Print for each module the names of the other modules that use it.
#
def printcaller():
files = file2def.keys()
files.sort()
for filename in files:
callers = []
for label in file2def[filename]:
if undef2file.has_key(label):
callers = callers + undef2file[label]
if callers:
callers.sort()
print filename + ':'
lastfn = ''
for fn in callers:
if fn <> lastfn:
print '\t' + fn
lastfn = fn
else:
print filename + ': unused'
# Print undefined names and where they are used.
#
def printundef():
undefs = {}
for filename in file2undef.keys():
for ext in file2undef[filename]:
if not def2file.has_key(ext):
store(undefs, ext, filename)
elist = undefs.keys()
elist.sort()
for ext in elist:
print ext + ':'
flist = undefs[ext]
flist.sort()
for filename in flist:
print '\t' + filename
# Print warning messages about names defined in more than one file.
#
def warndups():
savestdout = sys.stdout
sys.stdout = sys.stderr
names = def2file.keys()
names.sort()
for name in names:
if len(def2file[name]) > 1:
print 'warning:', name, 'multiply defined:',
print flat(def2file[name])
sys.stdout = savestdout
# Main program
#
def main():
try:
optlist, args = getopt.getopt(sys.argv[1:], 'cdu')
except getopt.error:
sys.stdout = sys.stderr
print 'Usage:', os.path.basename(sys.argv[0]),
print '[-cdu] [file] ...'
print '-c: print callers per objectfile'
print '-d: print callees per objectfile'
print '-u: print usage of undefined symbols'
print 'If none of -cdu is specified, all are assumed.'
print 'Use "nm -o" to generate the input (on IRIX: "nm -Bo"),'
print 'e.g.: nm -o /lib/libc.a | objgraph'
return 1
optu = optc = optd = 0
for opt, void in optlist:
if opt == '-u':
optu = 1
elif opt == '-c':
optc = 1
elif opt == '-d':
optd = 1
if optu == optc == optd == 0:
optu = optc = optd = 1
if not args:
args = ['-']
for filename in args:
if filename == '-':
readinput(sys.stdin)
else:
readinput(open(filename, 'r'))
#
warndups()
#
more = (optu + optc + optd > 1)
if optd:
if more:
print '---------------All callees------------------'
printcallee()
if optu:
if more:
print '---------------Undefined callees------------'
printundef()
if optc:
if more:
print '---------------All Callers------------------'
printcaller()
return 0
# Call the main program.
# Use its return value as exit status.
# Catch interrupts to avoid stack trace.
#
if __name__ == '__main__':
try:
sys.exit(main())
except KeyboardInterrupt:
sys.exit(1)