1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
|
# -*- ruby encoding: utf-8 -*-
module Diff::LCS::Internals # :nodoc:
end
class << Diff::LCS::Internals
# Compute the longest common subsequence between the sequenced
# Enumerables +a+ and +b+. The result is an array whose contents is such
# that
#
# result = Diff::LCS::Internals.lcs(a, b)
# result.each_with_index do |e, i|
# assert_equal(a[i], b[e]) unless e.nil?
# end
def lcs(a, b)
a_start = b_start = 0
a_finish = a.size - 1
b_finish = b.size - 1
vector = []
# Prune off any common elements at the beginning...
while (a_start <= a_finish) and
(b_start <= b_finish) and
(a[a_start] == b[b_start])
vector[a_start] = b_start
a_start += 1
b_start += 1
end
# Now the end...
while (a_start <= a_finish) and
(b_start <= b_finish) and
(a[a_finish] == b[b_finish])
vector[a_finish] = b_finish
a_finish -= 1
b_finish -= 1
end
# Now, compute the equivalence classes of positions of elements.
b_matches = position_hash(b, b_start..b_finish)
thresh = []
links = []
(a_start .. a_finish).each do |i|
ai = a.kind_of?(String) ? a[i, 1] : a[i]
bm = b_matches[ai]
k = nil
bm.reverse_each do |j|
if k and (thresh[k] > j) and (thresh[k - 1] < j)
thresh[k] = j
else
k = replace_next_larger(thresh, j, k)
end
links[k] = [ (k > 0) ? links[k - 1] : nil, i, j ] unless k.nil?
end
end
unless thresh.empty?
link = links[thresh.size - 1]
while not link.nil?
vector[link[1]] = link[2]
link = link[0]
end
end
vector
end
# This method will analyze the provided patchset to provide a
# single-pass normalization (conversion of the array form of
# Diff::LCS::Change objects to the object form of same) and detection of
# whether the patchset represents changes to be made.
def analyze_patchset(patchset, depth = 0)
raise "Patchset too complex" if depth > 1
has_changes = false
# Format:
# [ # patchset
# # hunk (change)
# [ # hunk
# # change
# ]
# ]
patchset = patchset.map do |hunk|
case hunk
when Diff::LCS::Change
has_changes ||= !hunk.unchanged?
hunk
when Array
# Detect if the 'hunk' is actually an array-format
# Change object.
if Diff::LCS::Change.valid_action? hunk[0]
hunk = Diff::LCS::Change.from_a(hunk)
has_changes ||= !hunk.unchanged?
hunk
else
with_changes, hunk = analyze_patchset(hunk, depth + 1)
has_changes ||= with_changes
hunk.flatten
end
else
raise ArgumentError, "Cannot normalise a hunk of class #{hunk.class}."
end
end
[ has_changes, patchset.flatten(1) ]
end
class DiffDirectionCounter
attr_reader :count, :left_match, :left_miss, :right_match, :right_miss
def initialize
@count = @left_match = @left_miss = @right_match = @right_miss = 0
end
def count!
@count += 1
end
def left_match!
@left_match += 1
end
def left_miss!
@left_miss += 1
end
def right_match!
@right_match += 1
end
def right_miss!
@right_miss += 1
end
def inspect
l = "L#{left_match}:#{left_miss} (#{(left_match == 0) && left_miss > 0})"
r = "R#{right_match}:#{right_miss} (#{(right_match == 0) && right_miss > 0})"
"#{l} #{r}"
end
end
# Examine the patchset and the source to see in which direction the
# patch should be applied.
#
# WARNING: By default, this examines the whole patch, so this could take
# some time. This also works better with Diff::LCS::ContextChange or
# Diff::LCS::Change as its source, as an array will cause the creation
# of one of the above.
#
# Note: This will be deprecated as a public function in a future release.
def diff_direction(src, patchset, limit = nil)
fwd_ddc = DiffDirectionCounter.new
string = src.kind_of?(String)
count = 0
patchset.each do |change|
count += 1
case change
when Diff::LCS::ContextChange
le = string ? src[change.old_position, 1] : src[change.old_position]
re = string ? src[change.new_position, 1] : src[change.new_position]
case change.action
when '-' # Remove details from the old string
if le == change.old_element
fwd_ddc.left_match!
else
fwd_ddc.left_miss!
end
if re == change.old_element
else
end
when '+'
if re == change.new_element
fwd_ddc.right_match!
else
fwd_ddc.right_miss!
end
when '='
fwd_ddc.left_miss! if le != change.old_element
fwd_ddc.right_miss! if re != change.new_element
when '!'
if le == change.old_element
fwd_ddc.left_match!
else
if re == change.new_element
fwd_ddc.right_match!
else
fwd_ddc.left_miss!
fwd_ddc.right_miss!
end
end
end
when Diff::LCS::Change
# With a simplistic change, we can't tell the difference between
# the left and right on '!' actions, so we ignore those. On '='
# actions, if there's a miss, we miss both left and right.
element = string ? src[change.position, 1] : src[change.position]
case change.action
when '-'
if element == change.element
fwd_ddc.left_match!
else
fwd_ddc.left_miss!
end
when '+'
if element == change.element
fwd_ddc.right_match!
else
fwd_ddc.right_miss!
end
when '='
if element != change.element
fwd_ddc.left_miss!
fwd_ddc.right_miss!
end
end
end
break if (not limit.nil?) && (count > limit)
end
fwd_no_left = (fwd_ddc.left_match == 0) && (fwd_ddc.left_miss > 0)
fwd_no_right = (fwd_ddc.right_match == 0) && (fwd_ddc.right_miss > 0)
case [fwd_no_left, fwd_no_right]
when [false, true]
:patch
when [true, false]
:unpatch
else
case fwd_ddc.left_match <=> fwd_ddc.right_match
when 1
:patch
when -1
:unpatch
else
raise "The provided patchset does not appear to apply to the provided value as either source or destination value."
end
end
end
# Find the place at which +value+ would normally be inserted into the
# Enumerable. If that place is already occupied by +value+, do nothing
# and return +nil+. If the place does not exist (i.e., it is off the end
# of the Enumerable), add it to the end. Otherwise, replace the element
# at that point with +value+. It is assumed that the Enumerable's values
# are numeric.
#
# This operation preserves the sort order.
def replace_next_larger(enum, value, last_index = nil)
# Off the end?
if enum.empty? or (value > enum[-1])
enum << value
return enum.size - 1
end
# Binary search for the insertion point
last_index ||= enum.size
first_index = 0
while (first_index <= last_index)
i = (first_index + last_index) >> 1
found = enum[i]
if value == found
return nil
elsif value > found
first_index = i + 1
else
last_index = i - 1
end
end
# The insertion point is in first_index; overwrite the next larger
# value.
enum[first_index] = value
return first_index
end
private :replace_next_larger
# If +vector+ maps the matching elements of another collection onto this
# Enumerable, compute the inverse of +vector+ that maps this Enumerable
# onto the collection. (Currently unused.)
def inverse_vector(a, vector)
inverse = a.dup
(0...vector.size).each do |i|
inverse[vector[i]] = i unless vector[i].nil?
end
inverse
end
private :inverse_vector
# Returns a hash mapping each element of an Enumerable to the set of
# positions it occupies in the Enumerable, optionally restricted to the
# elements specified in the range of indexes specified by +interval+.
def position_hash(enum, interval)
hash = Hash.new { |h, k| h[k] = [] }
interval.each do |i|
k = enum.kind_of?(String) ? enum[i, 1] : enum[i]
hash[k] << i
end
hash
end
private :position_hash
end
|