mercurial/mdiff.py
author mpm@selenic.com
Mon, 13 Jun 2005 13:43:48 -0800
changeset 325 ad87e19854a6
parent 318 2819f63b16bf
child 326 235443668bea
permissions -rw-r--r--
mdiff: reinstate new algorithm -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 mdiff: reinstate new algorithm This unreverts the new algorithm with a fix from Chris (s/len/lena) and adds some comments on what it's doing. manifest hash: 75fc1acee1926e57d495f67a44cd88d9555f2356 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.0 (GNU/Linux) iD8DBQFCrf4UywK+sNU5EO8RAoRzAKCA2vpUAGNqTkDeba3YHo6XXht7VgCfXQK0 /j5yv5cucnsYezCdclpftOA= =FNMD -----END PGP SIGNATURE-----
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
239
75840796e8e2 mdiff.py: kill #! line, add copyright notice
mpm@selenic.com
parents: 184
diff changeset
     1
# mdiff.py - diff and patch routines for mercurial
75840796e8e2 mdiff.py: kill #! line, add copyright notice
mpm@selenic.com
parents: 184
diff changeset
     2
#
75840796e8e2 mdiff.py: kill #! line, add copyright notice
mpm@selenic.com
parents: 184
diff changeset
     3
# Copyright 2005 Matt Mackall <mpm@selenic.com>
75840796e8e2 mdiff.py: kill #! line, add copyright notice
mpm@selenic.com
parents: 184
diff changeset
     4
#
75840796e8e2 mdiff.py: kill #! line, add copyright notice
mpm@selenic.com
parents: 184
diff changeset
     5
# This software may be used and distributed according to the terms
75840796e8e2 mdiff.py: kill #! line, add copyright notice
mpm@selenic.com
parents: 184
diff changeset
     6
# of the GNU General Public License, incorporated herein by reference.
75840796e8e2 mdiff.py: kill #! line, add copyright notice
mpm@selenic.com
parents: 184
diff changeset
     7
249
619e775aa7f9 import and startup cleanups
mpm@selenic.com
parents: 241
diff changeset
     8
import difflib, struct
127
44538462d3c8 Fix braindamaged import in mdiff.
mpm@selenic.com
parents: 125
diff changeset
     9
from mercurial.mpatch import *
0
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    10
64
b3e2ddff0159 Diff in subdirectories from Jake Edge
mpm@selenic.com
parents: 35
diff changeset
    11
def unidiff(a, ad, b, bd, fn):
35
9197c26a414b unidiff: punt on comparing empty files
mpm@selenic.com
parents: 0
diff changeset
    12
    if not a and not b: return ""
264
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    13
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    14
    if a == None:
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    15
        b = b.splitlines(1)
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    16
        l1 = "--- %s\t%s\n" % ("/dev/null", ad)
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    17
        l2 = "+++ %s\t%s\n" % ("b/" + fn, bd)
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    18
        l3 = "@@ -0,0 +1,%d @@\n" % len(b)
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    19
        l = [l1, l2, l3] + ["+" + e for e in b]
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    20
    elif b == None:
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    21
        a = a.splitlines(1)
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    22
        l1 = "--- %s\t%s\n" % ("a/" + fn, ad)
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    23
        l2 = "+++ %s\t%s\n" % ("/dev/null", bd)
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    24
        l3 = "@@ -1,%d +0,0 @@\n" % len(a)
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    25
        l = [l1, l2, l3] + ["-" + e for e in a]
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    26
    else:
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    27
        a = a.splitlines(1)
4c1d7072d5cd Attempt to make diff deal with null sources properly
mpm@selenic.com
parents: 249
diff changeset
    28
        b = b.splitlines(1)
272
467cea2bf2d8 diff: use tab to separate date from filename
mpm@selenic.com
parents: 264
diff changeset
    29
        l = list(difflib.unified_diff(a, b, "a/" + fn, "b/" + fn))
278
777e388c06d6 unidiff: handle empty diffs more gracefully
mpm@selenic.com
parents: 272
diff changeset
    30
        if not l: return ""
272
467cea2bf2d8 diff: use tab to separate date from filename
mpm@selenic.com
parents: 264
diff changeset
    31
        # difflib uses a space, rather than a tab
467cea2bf2d8 diff: use tab to separate date from filename
mpm@selenic.com
parents: 264
diff changeset
    32
        l[0] = l[0][:-2] + "\t" + ad + "\n"
467cea2bf2d8 diff: use tab to separate date from filename
mpm@selenic.com
parents: 264
diff changeset
    33
        l[1] = l[1][:-2] + "\t" + bd + "\n"
170
e6c621a825f2 hg diff: fix missing final newline bug
mpm@selenic.com
parents: 127
diff changeset
    34
e6c621a825f2 hg diff: fix missing final newline bug
mpm@selenic.com
parents: 127
diff changeset
    35
    for ln in xrange(len(l)):
e6c621a825f2 hg diff: fix missing final newline bug
mpm@selenic.com
parents: 127
diff changeset
    36
        if l[ln][-1] != '\n':
e6c621a825f2 hg diff: fix missing final newline bug
mpm@selenic.com
parents: 127
diff changeset
    37
            l[ln] += "\n\ No newline at end of file\n"
e6c621a825f2 hg diff: fix missing final newline bug
mpm@selenic.com
parents: 127
diff changeset
    38
0
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    39
    return "".join(l)
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    40
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    41
def textdiff(a, b):
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    42
    return diff(a.splitlines(1), b.splitlines(1))
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    43
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    44
def sortdiff(a, b):
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    45
    la = lb = 0
325
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    46
    lena = len(a)
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    47
    lenb = len(b)
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    48
    
0
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    49
    while 1:
325
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    50
        am, bm, = la, lb
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    51
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    52
        # walk over matching lines
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    53
        while lb < lenb and la < lenb and a[la] == b[lb] :
0
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    54
            la += 1
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    55
            lb += 1
318
2819f63b16bf mdiff: revert grouping optimization for the time being
mpm@selenic.com
parents: 317
diff changeset
    56
325
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    57
        if la > am:
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    58
            yield (am, bm, la - am) # return a match
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    59
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    60
        # skip mismatched lines from b
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    61
        while lb < lenb and b[lb] < a[la]:
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    62
            lb += 1
318
2819f63b16bf mdiff: revert grouping optimization for the time being
mpm@selenic.com
parents: 317
diff changeset
    63
325
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    64
        if lb >= lenb:
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    65
            break
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    66
        
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    67
        # skip mismatched lines from a
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    68
        while la < lena and b[lb] > a[la]:
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    69
            la += 1
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    70
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    71
        if la >= lena:
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    72
            break
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    73
        
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    74
    yield (lena, lenb, 0)
0
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    75
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    76
def diff(a, b, sorted=0):
325
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    77
    if not a:
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    78
        s = "".join(b)
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    79
        return s and (struct.pack(">lll", 0, 0, len(s)) + s)
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    80
0
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    81
    bin = []
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    82
    p = [0]
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    83
    for i in a: p.append(p[-1] + len(i))
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    84
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    85
    if sorted:
317
b18ce742566a hg commit: user and date options
mpm@selenic.com
parents: 278
diff changeset
    86
        try:
b18ce742566a hg commit: user and date options
mpm@selenic.com
parents: 278
diff changeset
    87
            d = sortdiff(a, b)
b18ce742566a hg commit: user and date options
mpm@selenic.com
parents: 278
diff changeset
    88
        except:
b18ce742566a hg commit: user and date options
mpm@selenic.com
parents: 278
diff changeset
    89
            print a, b
b18ce742566a hg commit: user and date options
mpm@selenic.com
parents: 278
diff changeset
    90
            raise
0
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
    91
    else:
325
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    92
        d = difflib.SequenceMatcher(None, a, b).get_matching_blocks()
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    93
    la = 0
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    94
    lb = 0
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    95
    for am, bm, size in d:
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    96
        s = "".join(b[lb:bm])
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    97
        if am > la or s:
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    98
            bin.append(struct.pack(">lll", p[la], p[am], len(s)) + s)
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
    99
        la = am + size
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
   100
        lb = bm + size
ad87e19854a6 mdiff: reinstate new algorithm
mpm@selenic.com
parents: 318
diff changeset
   101
    
0
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
   102
    return "".join(bin)
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
   103
120
bae6f0328f63 Add a function to return the new text from a binary diff
mpm@selenic.com
parents: 75
diff changeset
   104
def patchtext(bin):
bae6f0328f63 Add a function to return the new text from a binary diff
mpm@selenic.com
parents: 75
diff changeset
   105
    pos = 0
bae6f0328f63 Add a function to return the new text from a binary diff
mpm@selenic.com
parents: 75
diff changeset
   106
    t = []
bae6f0328f63 Add a function to return the new text from a binary diff
mpm@selenic.com
parents: 75
diff changeset
   107
    while pos < len(bin):
bae6f0328f63 Add a function to return the new text from a binary diff
mpm@selenic.com
parents: 75
diff changeset
   108
        p1, p2, l = struct.unpack(">lll", bin[pos:pos + 12])
bae6f0328f63 Add a function to return the new text from a binary diff
mpm@selenic.com
parents: 75
diff changeset
   109
        pos += 12
bae6f0328f63 Add a function to return the new text from a binary diff
mpm@selenic.com
parents: 75
diff changeset
   110
        t.append(bin[pos:pos + l])
bae6f0328f63 Add a function to return the new text from a binary diff
mpm@selenic.com
parents: 75
diff changeset
   111
        pos += l
bae6f0328f63 Add a function to return the new text from a binary diff
mpm@selenic.com
parents: 75
diff changeset
   112
    return "".join(t)
bae6f0328f63 Add a function to return the new text from a binary diff
mpm@selenic.com
parents: 75
diff changeset
   113
0
9117c6561b0b Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff changeset
   114
def patch(a, bin):
72
4a6ab4d80dc4 Add an O(m + nlog n) patching extension
mpm@selenic.com
parents: 71
diff changeset
   115
    return patches(a, [bin])