annotate contrib/hgdiff @ 2474:1e32e2fe8a67

Fix cold cache diff performance cold cache diff performance has regressed in two ways. localrepo.changes has optimizations for diffing against the working dir parent that expect node1 to be None. commands.revpair() usage means that commands.dodiff() never sends node1 == None. This is fixed in localrepo.changes by checking against the dirstate parents. In the non-dirstate parents case, localrepo.changes does a loop comparing files without first sorting the file names, leading to random access across the disk.
author Chris Mason <mason@suse.com>
date Wed, 21 Jun 2006 09:28:48 -0700
parents e7e6504c4989
children 0f308690bda8
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
1636
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
1 #!/usr/bin/env python
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
2
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
3 import os, sys, struct, stat
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
4 import difflib
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
5 import re
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
6 from optparse import OptionParser
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
7 from mercurial.bdiff import bdiff, blocks
1644
e7e6504c4989 Remove duplicate bunidiff code from hgdiff, importing from mdiff.py instead
mason@suse.com
parents: 1636
diff changeset
8 from mercurial.mdiff import bunidiff
1636
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
9
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
10 VERSION="0.2"
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
11 usage = "usage: %prog [options] file1 file2"
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
12 parser = OptionParser(usage=usage)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
13
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
14 parser.add_option("-d", "--difflib", action="store_true", default=False)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
15 parser.add_option('-x', '--count', default=1)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
16 parser.add_option('-c', '--context', type="int", default=3)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
17 parser.add_option('-p', '--show-c-function', action="store_true", default=False)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
18 parser.add_option('-w', '--ignore-all-space', action="store_true",
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
19 default=False)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
20
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
21 (options, args) = parser.parse_args()
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
22
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
23 if not args:
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
24 parser.print_help()
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
25 sys.exit(1)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
26
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
27 # simple utility function to put all the
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
28 # files from a directory tree into a dict
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
29 def buildlist(names, top):
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
30 tlen = len(top)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
31 for root, dirs, files in os.walk(top):
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
32 l = root[tlen + 1:]
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
33 for x in files:
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
34 p = os.path.join(root, x)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
35 st = os.lstat(p)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
36 if stat.S_ISREG(st.st_mode):
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
37 names[os.path.join(l, x)] = (st.st_dev, st.st_ino)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
38
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
39 def diff_files(file1, file2):
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
40 if file1 == None:
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
41 b = file(file2).read().splitlines(1)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
42 l1 = "--- %s\n" % (file2)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
43 l2 = "+++ %s\n" % (file2)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
44 l3 = "@@ -0,0 +1,%d @@\n" % len(b)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
45 l = [l1, l2, l3] + ["+" + e for e in b]
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
46 elif file2 == None:
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
47 a = file(file1).read().splitlines(1)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
48 l1 = "--- %s\n" % (file1)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
49 l2 = "+++ %s\n" % (file1)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
50 l3 = "@@ -1,%d +0,0 @@\n" % len(a)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
51 l = [l1, l2, l3] + ["-" + e for e in a]
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
52 else:
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
53 t1 = file(file1).read()
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
54 t2 = file(file2).read()
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
55 l1 = t1.splitlines(1)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
56 l2 = t2.splitlines(1)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
57 if options.difflib:
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
58 l = difflib.unified_diff(l1, l2, file1, file2)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
59 else:
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
60 l = bunidiff(t1, t2, l1, l2, file1, file2, context=options.context,
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
61 showfunc=options.show_c_function,
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
62 ignorews=options.ignore_all_space)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
63 for x in l:
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
64 if x[-1] != '\n':
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
65 x += "\n\ No newline at end of file\n"
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
66 print x,
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
67
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
68 file1 = args[0]
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
69 file2 = args[1]
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
70
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
71 if os.path.isfile(file1) and os.path.isfile(file2):
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
72 diff_files(file1, file2)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
73 elif os.path.isdir(file1):
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
74 if not os.path.isdir(file2):
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
75 sys.stderr.write("file types don't match\n")
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
76 sys.exit(1)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
77
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
78 d1 = {}
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
79 d2 = {}
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
80
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
81 buildlist(d1, file1)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
82 buildlist(d2, file2)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
83 keys = d1.keys()
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
84 keys.sort()
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
85 for x in keys:
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
86 if x not in d2:
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
87 f2 = None
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
88 else:
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
89 f2 = os.path.join(file2, x)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
90 st1 = d1[x]
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
91 st2 = d2[x]
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
92 del d2[x]
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
93 if st1[0] == st2[0] and st1[1] == st2[1]:
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
94 sys.stderr.write("%s is a hard link\n" % x)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
95 continue
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
96 x = os.path.join(file1, x)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
97 diff_files(x, f2)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
98 keys = d2.keys()
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
99 keys.sort()
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
100 for x in keys:
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
101 f1 = None
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
102 x = os.path.join(file2, x)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
103 diff_files(f1, x)
7da32bb3d1d3 contrib: add Chris Mason's stand-alone diff tool
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
104