Show More
@@ -1,233 +1,233 b'' | |||||
1 | # copies.py - copy detection for Mercurial |
|
1 | # copies.py - copy detection for Mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2008 Matt Mackall <mpm@selenic.com> |
|
3 | # Copyright 2008 Matt Mackall <mpm@selenic.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2, incorporated herein by reference. |
|
6 | # GNU General Public License version 2, incorporated herein by reference. | |
7 |
|
7 | |||
8 | from i18n import _ |
|
8 | from i18n import _ | |
9 | import util |
|
9 | import util | |
10 | import heapq |
|
10 | import heapq | |
11 |
|
11 | |||
12 | def _nonoverlap(d1, d2, d3): |
|
12 | def _nonoverlap(d1, d2, d3): | |
13 | "Return list of elements in d1 not in d2 or d3" |
|
13 | "Return list of elements in d1 not in d2 or d3" | |
14 | return sorted([d for d in d1 if d not in d3 and d not in d2]) |
|
14 | return sorted([d for d in d1 if d not in d3 and d not in d2]) | |
15 |
|
15 | |||
16 | def _dirname(f): |
|
16 | def _dirname(f): | |
17 | s = f.rfind("/") |
|
17 | s = f.rfind("/") | |
18 | if s == -1: |
|
18 | if s == -1: | |
19 | return "" |
|
19 | return "" | |
20 | return f[:s] |
|
20 | return f[:s] | |
21 |
|
21 | |||
22 | def _dirs(files): |
|
22 | def _dirs(files): | |
23 |
d = |
|
23 | d = set() | |
24 | for f in files: |
|
24 | for f in files: | |
25 | f = _dirname(f) |
|
25 | f = _dirname(f) | |
26 | while f not in d: |
|
26 | while f not in d: | |
27 |
d |
|
27 | d.add(f) | |
28 | f = _dirname(f) |
|
28 | f = _dirname(f) | |
29 | return d |
|
29 | return d | |
30 |
|
30 | |||
31 | def _findoldnames(fctx, limit): |
|
31 | def _findoldnames(fctx, limit): | |
32 | "find files that path was copied from, back to linkrev limit" |
|
32 | "find files that path was copied from, back to linkrev limit" | |
33 | old = {} |
|
33 | old = {} | |
34 |
seen = |
|
34 | seen = set() | |
35 | orig = fctx.path() |
|
35 | orig = fctx.path() | |
36 | visit = [(fctx, 0)] |
|
36 | visit = [(fctx, 0)] | |
37 | while visit: |
|
37 | while visit: | |
38 | fc, depth = visit.pop() |
|
38 | fc, depth = visit.pop() | |
39 | s = str(fc) |
|
39 | s = str(fc) | |
40 | if s in seen: |
|
40 | if s in seen: | |
41 | continue |
|
41 | continue | |
42 |
seen |
|
42 | seen.add(s) | |
43 | if fc.path() != orig and fc.path() not in old: |
|
43 | if fc.path() != orig and fc.path() not in old: | |
44 | old[fc.path()] = (depth, fc.path()) # remember depth |
|
44 | old[fc.path()] = (depth, fc.path()) # remember depth | |
45 | if fc.rev() < limit and fc.rev() is not None: |
|
45 | if fc.rev() < limit and fc.rev() is not None: | |
46 | continue |
|
46 | continue | |
47 | visit += [(p, depth - 1) for p in fc.parents()] |
|
47 | visit += [(p, depth - 1) for p in fc.parents()] | |
48 |
|
48 | |||
49 | # return old names sorted by depth |
|
49 | # return old names sorted by depth | |
50 | return [o[1] for o in sorted(old.values())] |
|
50 | return [o[1] for o in sorted(old.values())] | |
51 |
|
51 | |||
52 | def _findlimit(repo, a, b): |
|
52 | def _findlimit(repo, a, b): | |
53 | "find the earliest revision that's an ancestor of a or b but not both" |
|
53 | "find the earliest revision that's an ancestor of a or b but not both" | |
54 | # basic idea: |
|
54 | # basic idea: | |
55 | # - mark a and b with different sides |
|
55 | # - mark a and b with different sides | |
56 | # - if a parent's children are all on the same side, the parent is |
|
56 | # - if a parent's children are all on the same side, the parent is | |
57 | # on that side, otherwise it is on no side |
|
57 | # on that side, otherwise it is on no side | |
58 | # - walk the graph in topological order with the help of a heap; |
|
58 | # - walk the graph in topological order with the help of a heap; | |
59 | # - add unseen parents to side map |
|
59 | # - add unseen parents to side map | |
60 | # - clear side of any parent that has children on different sides |
|
60 | # - clear side of any parent that has children on different sides | |
61 | # - track number of interesting revs that might still be on a side |
|
61 | # - track number of interesting revs that might still be on a side | |
62 | # - track the lowest interesting rev seen |
|
62 | # - track the lowest interesting rev seen | |
63 | # - quit when interesting revs is zero |
|
63 | # - quit when interesting revs is zero | |
64 |
|
64 | |||
65 | cl = repo.changelog |
|
65 | cl = repo.changelog | |
66 | working = len(cl) # pseudo rev for the working directory |
|
66 | working = len(cl) # pseudo rev for the working directory | |
67 | if a is None: |
|
67 | if a is None: | |
68 | a = working |
|
68 | a = working | |
69 | if b is None: |
|
69 | if b is None: | |
70 | b = working |
|
70 | b = working | |
71 |
|
71 | |||
72 | side = {a: -1, b: 1} |
|
72 | side = {a: -1, b: 1} | |
73 | visit = [-a, -b] |
|
73 | visit = [-a, -b] | |
74 | heapq.heapify(visit) |
|
74 | heapq.heapify(visit) | |
75 | interesting = len(visit) |
|
75 | interesting = len(visit) | |
76 | limit = working |
|
76 | limit = working | |
77 |
|
77 | |||
78 | while interesting: |
|
78 | while interesting: | |
79 | r = -heapq.heappop(visit) |
|
79 | r = -heapq.heappop(visit) | |
80 | if r == working: |
|
80 | if r == working: | |
81 | parents = [cl.rev(p) for p in repo.dirstate.parents()] |
|
81 | parents = [cl.rev(p) for p in repo.dirstate.parents()] | |
82 | else: |
|
82 | else: | |
83 | parents = cl.parentrevs(r) |
|
83 | parents = cl.parentrevs(r) | |
84 | for p in parents: |
|
84 | for p in parents: | |
85 | if p not in side: |
|
85 | if p not in side: | |
86 | # first time we see p; add it to visit |
|
86 | # first time we see p; add it to visit | |
87 | side[p] = side[r] |
|
87 | side[p] = side[r] | |
88 | if side[p]: |
|
88 | if side[p]: | |
89 | interesting += 1 |
|
89 | interesting += 1 | |
90 | heapq.heappush(visit, -p) |
|
90 | heapq.heappush(visit, -p) | |
91 | elif side[p] and side[p] != side[r]: |
|
91 | elif side[p] and side[p] != side[r]: | |
92 | # p was interesting but now we know better |
|
92 | # p was interesting but now we know better | |
93 | side[p] = 0 |
|
93 | side[p] = 0 | |
94 | interesting -= 1 |
|
94 | interesting -= 1 | |
95 | if side[r]: |
|
95 | if side[r]: | |
96 | limit = r # lowest rev visited |
|
96 | limit = r # lowest rev visited | |
97 | interesting -= 1 |
|
97 | interesting -= 1 | |
98 | return limit |
|
98 | return limit | |
99 |
|
99 | |||
100 | def copies(repo, c1, c2, ca, checkdirs=False): |
|
100 | def copies(repo, c1, c2, ca, checkdirs=False): | |
101 | """ |
|
101 | """ | |
102 | Find moves and copies between context c1 and c2 |
|
102 | Find moves and copies between context c1 and c2 | |
103 | """ |
|
103 | """ | |
104 | # avoid silly behavior for update from empty dir |
|
104 | # avoid silly behavior for update from empty dir | |
105 | if not c1 or not c2 or c1 == c2: |
|
105 | if not c1 or not c2 or c1 == c2: | |
106 | return {}, {} |
|
106 | return {}, {} | |
107 |
|
107 | |||
108 | # avoid silly behavior for parent -> working dir |
|
108 | # avoid silly behavior for parent -> working dir | |
109 | if c2.node() == None and c1.node() == repo.dirstate.parents()[0]: |
|
109 | if c2.node() == None and c1.node() == repo.dirstate.parents()[0]: | |
110 | return repo.dirstate.copies(), {} |
|
110 | return repo.dirstate.copies(), {} | |
111 |
|
111 | |||
112 | limit = _findlimit(repo, c1.rev(), c2.rev()) |
|
112 | limit = _findlimit(repo, c1.rev(), c2.rev()) | |
113 | m1 = c1.manifest() |
|
113 | m1 = c1.manifest() | |
114 | m2 = c2.manifest() |
|
114 | m2 = c2.manifest() | |
115 | ma = ca.manifest() |
|
115 | ma = ca.manifest() | |
116 |
|
116 | |||
117 | def makectx(f, n): |
|
117 | def makectx(f, n): | |
118 | if len(n) != 20: # in a working context? |
|
118 | if len(n) != 20: # in a working context? | |
119 | if c1.rev() is None: |
|
119 | if c1.rev() is None: | |
120 | return c1.filectx(f) |
|
120 | return c1.filectx(f) | |
121 | return c2.filectx(f) |
|
121 | return c2.filectx(f) | |
122 | return repo.filectx(f, fileid=n) |
|
122 | return repo.filectx(f, fileid=n) | |
123 | ctx = util.cachefunc(makectx) |
|
123 | ctx = util.cachefunc(makectx) | |
124 |
|
124 | |||
125 | copy = {} |
|
125 | copy = {} | |
126 | fullcopy = {} |
|
126 | fullcopy = {} | |
127 | diverge = {} |
|
127 | diverge = {} | |
128 |
|
128 | |||
129 | def checkcopies(f, m1, m2): |
|
129 | def checkcopies(f, m1, m2): | |
130 | '''check possible copies of f from m1 to m2''' |
|
130 | '''check possible copies of f from m1 to m2''' | |
131 | c1 = ctx(f, m1[f]) |
|
131 | c1 = ctx(f, m1[f]) | |
132 | for of in _findoldnames(c1, limit): |
|
132 | for of in _findoldnames(c1, limit): | |
133 | fullcopy[f] = of # remember for dir rename detection |
|
133 | fullcopy[f] = of # remember for dir rename detection | |
134 | if of in m2: # original file not in other manifest? |
|
134 | if of in m2: # original file not in other manifest? | |
135 | # if the original file is unchanged on the other branch, |
|
135 | # if the original file is unchanged on the other branch, | |
136 | # no merge needed |
|
136 | # no merge needed | |
137 | if m2[of] != ma.get(of): |
|
137 | if m2[of] != ma.get(of): | |
138 | c2 = ctx(of, m2[of]) |
|
138 | c2 = ctx(of, m2[of]) | |
139 | ca = c1.ancestor(c2) |
|
139 | ca = c1.ancestor(c2) | |
140 | # related and named changed on only one side? |
|
140 | # related and named changed on only one side? | |
141 | if ca and (ca.path() == f or ca.path() == c2.path()): |
|
141 | if ca and (ca.path() == f or ca.path() == c2.path()): | |
142 | if c1 != ca or c2 != ca: # merge needed? |
|
142 | if c1 != ca or c2 != ca: # merge needed? | |
143 | copy[f] = of |
|
143 | copy[f] = of | |
144 | elif of in ma: |
|
144 | elif of in ma: | |
145 | diverge.setdefault(of, []).append(f) |
|
145 | diverge.setdefault(of, []).append(f) | |
146 |
|
146 | |||
147 | repo.ui.debug(_(" searching for copies back to rev %d\n") % limit) |
|
147 | repo.ui.debug(_(" searching for copies back to rev %d\n") % limit) | |
148 |
|
148 | |||
149 | u1 = _nonoverlap(m1, m2, ma) |
|
149 | u1 = _nonoverlap(m1, m2, ma) | |
150 | u2 = _nonoverlap(m2, m1, ma) |
|
150 | u2 = _nonoverlap(m2, m1, ma) | |
151 |
|
151 | |||
152 | if u1: |
|
152 | if u1: | |
153 | repo.ui.debug(_(" unmatched files in local:\n %s\n") |
|
153 | repo.ui.debug(_(" unmatched files in local:\n %s\n") | |
154 | % "\n ".join(u1)) |
|
154 | % "\n ".join(u1)) | |
155 | if u2: |
|
155 | if u2: | |
156 | repo.ui.debug(_(" unmatched files in other:\n %s\n") |
|
156 | repo.ui.debug(_(" unmatched files in other:\n %s\n") | |
157 | % "\n ".join(u2)) |
|
157 | % "\n ".join(u2)) | |
158 |
|
158 | |||
159 | for f in u1: |
|
159 | for f in u1: | |
160 | checkcopies(f, m1, m2) |
|
160 | checkcopies(f, m1, m2) | |
161 | for f in u2: |
|
161 | for f in u2: | |
162 | checkcopies(f, m2, m1) |
|
162 | checkcopies(f, m2, m1) | |
163 |
|
163 | |||
164 | diverge2 = set() |
|
164 | diverge2 = set() | |
165 | for of, fl in diverge.items(): |
|
165 | for of, fl in diverge.items(): | |
166 | if len(fl) == 1: |
|
166 | if len(fl) == 1: | |
167 | del diverge[of] # not actually divergent |
|
167 | del diverge[of] # not actually divergent | |
168 | else: |
|
168 | else: | |
169 | diverge2.update(fl) # reverse map for below |
|
169 | diverge2.update(fl) # reverse map for below | |
170 |
|
170 | |||
171 | if fullcopy: |
|
171 | if fullcopy: | |
172 | repo.ui.debug(_(" all copies found (* = to merge, ! = divergent):\n")) |
|
172 | repo.ui.debug(_(" all copies found (* = to merge, ! = divergent):\n")) | |
173 | for f in fullcopy: |
|
173 | for f in fullcopy: | |
174 | note = "" |
|
174 | note = "" | |
175 | if f in copy: note += "*" |
|
175 | if f in copy: note += "*" | |
176 | if f in diverge2: note += "!" |
|
176 | if f in diverge2: note += "!" | |
177 | repo.ui.debug(" %s -> %s %s\n" % (f, fullcopy[f], note)) |
|
177 | repo.ui.debug(" %s -> %s %s\n" % (f, fullcopy[f], note)) | |
178 | del diverge2 |
|
178 | del diverge2 | |
179 |
|
179 | |||
180 | if not fullcopy or not checkdirs: |
|
180 | if not fullcopy or not checkdirs: | |
181 | return copy, diverge |
|
181 | return copy, diverge | |
182 |
|
182 | |||
183 | repo.ui.debug(_(" checking for directory renames\n")) |
|
183 | repo.ui.debug(_(" checking for directory renames\n")) | |
184 |
|
184 | |||
185 | # generate a directory move map |
|
185 | # generate a directory move map | |
186 | d1, d2 = _dirs(m1), _dirs(m2) |
|
186 | d1, d2 = _dirs(m1), _dirs(m2) | |
187 |
invalid = |
|
187 | invalid = set() | |
188 | dirmove = {} |
|
188 | dirmove = {} | |
189 |
|
189 | |||
190 | # examine each file copy for a potential directory move, which is |
|
190 | # examine each file copy for a potential directory move, which is | |
191 | # when all the files in a directory are moved to a new directory |
|
191 | # when all the files in a directory are moved to a new directory | |
192 | for dst, src in fullcopy.iteritems(): |
|
192 | for dst, src in fullcopy.iteritems(): | |
193 | dsrc, ddst = _dirname(src), _dirname(dst) |
|
193 | dsrc, ddst = _dirname(src), _dirname(dst) | |
194 | if dsrc in invalid: |
|
194 | if dsrc in invalid: | |
195 | # already seen to be uninteresting |
|
195 | # already seen to be uninteresting | |
196 | continue |
|
196 | continue | |
197 | elif dsrc in d1 and ddst in d1: |
|
197 | elif dsrc in d1 and ddst in d1: | |
198 | # directory wasn't entirely moved locally |
|
198 | # directory wasn't entirely moved locally | |
199 |
invalid |
|
199 | invalid.add(dsrc) | |
200 | elif dsrc in d2 and ddst in d2: |
|
200 | elif dsrc in d2 and ddst in d2: | |
201 | # directory wasn't entirely moved remotely |
|
201 | # directory wasn't entirely moved remotely | |
202 |
invalid |
|
202 | invalid.add(dsrc) | |
203 | elif dsrc in dirmove and dirmove[dsrc] != ddst: |
|
203 | elif dsrc in dirmove and dirmove[dsrc] != ddst: | |
204 | # files from the same directory moved to two different places |
|
204 | # files from the same directory moved to two different places | |
205 |
invalid |
|
205 | invalid.add(dsrc) | |
206 | else: |
|
206 | else: | |
207 | # looks good so far |
|
207 | # looks good so far | |
208 | dirmove[dsrc + "/"] = ddst + "/" |
|
208 | dirmove[dsrc + "/"] = ddst + "/" | |
209 |
|
209 | |||
210 | for i in invalid: |
|
210 | for i in invalid: | |
211 | if i in dirmove: |
|
211 | if i in dirmove: | |
212 | del dirmove[i] |
|
212 | del dirmove[i] | |
213 | del d1, d2, invalid |
|
213 | del d1, d2, invalid | |
214 |
|
214 | |||
215 | if not dirmove: |
|
215 | if not dirmove: | |
216 | return copy, diverge |
|
216 | return copy, diverge | |
217 |
|
217 | |||
218 | for d in dirmove: |
|
218 | for d in dirmove: | |
219 | repo.ui.debug(_(" dir %s -> %s\n") % (d, dirmove[d])) |
|
219 | repo.ui.debug(_(" dir %s -> %s\n") % (d, dirmove[d])) | |
220 |
|
220 | |||
221 | # check unaccounted nonoverlapping files against directory moves |
|
221 | # check unaccounted nonoverlapping files against directory moves | |
222 | for f in u1 + u2: |
|
222 | for f in u1 + u2: | |
223 | if f not in fullcopy: |
|
223 | if f not in fullcopy: | |
224 | for d in dirmove: |
|
224 | for d in dirmove: | |
225 | if f.startswith(d): |
|
225 | if f.startswith(d): | |
226 | # new file added in a directory that was moved, move it |
|
226 | # new file added in a directory that was moved, move it | |
227 | df = dirmove[d] + f[len(d):] |
|
227 | df = dirmove[d] + f[len(d):] | |
228 | if df not in copy: |
|
228 | if df not in copy: | |
229 | copy[f] = df |
|
229 | copy[f] = df | |
230 | repo.ui.debug(_(" file %s -> %s\n") % (f, copy[f])) |
|
230 | repo.ui.debug(_(" file %s -> %s\n") % (f, copy[f])) | |
231 | break |
|
231 | break | |
232 |
|
232 | |||
233 | return copy, diverge |
|
233 | return copy, diverge |
General Comments 0
You need to be logged in to leave comments.
Login now