Show More
@@ -1,422 +1,422 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 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | import util |
|
8 | import util | |
9 | import heapq |
|
9 | import heapq | |
10 |
|
10 | |||
11 | def _nonoverlap(d1, d2, d3): |
|
11 | def _nonoverlap(d1, d2, d3): | |
12 | "Return list of elements in d1 not in d2 or d3" |
|
12 | "Return list of elements in d1 not in d2 or d3" | |
13 | return sorted([d for d in d1 if d not in d3 and d not in d2]) |
|
13 | return sorted([d for d in d1 if d not in d3 and d not in d2]) | |
14 |
|
14 | |||
15 | def _dirname(f): |
|
15 | def _dirname(f): | |
16 | s = f.rfind("/") |
|
16 | s = f.rfind("/") | |
17 | if s == -1: |
|
17 | if s == -1: | |
18 | return "" |
|
18 | return "" | |
19 | return f[:s] |
|
19 | return f[:s] | |
20 |
|
20 | |||
21 | def _findlimit(repo, a, b): |
|
21 | def _findlimit(repo, a, b): | |
22 | """Find the earliest revision that's an ancestor of a or b but not both, |
|
22 | """Find the earliest revision that's an ancestor of a or b but not both, | |
23 | None if no such revision exists. |
|
23 | None if no such revision exists. | |
24 | """ |
|
24 | """ | |
25 | # basic idea: |
|
25 | # basic idea: | |
26 | # - mark a and b with different sides |
|
26 | # - mark a and b with different sides | |
27 | # - if a parent's children are all on the same side, the parent is |
|
27 | # - if a parent's children are all on the same side, the parent is | |
28 | # on that side, otherwise it is on no side |
|
28 | # on that side, otherwise it is on no side | |
29 | # - walk the graph in topological order with the help of a heap; |
|
29 | # - walk the graph in topological order with the help of a heap; | |
30 | # - add unseen parents to side map |
|
30 | # - add unseen parents to side map | |
31 | # - clear side of any parent that has children on different sides |
|
31 | # - clear side of any parent that has children on different sides | |
32 | # - track number of interesting revs that might still be on a side |
|
32 | # - track number of interesting revs that might still be on a side | |
33 | # - track the lowest interesting rev seen |
|
33 | # - track the lowest interesting rev seen | |
34 | # - quit when interesting revs is zero |
|
34 | # - quit when interesting revs is zero | |
35 |
|
35 | |||
36 | cl = repo.changelog |
|
36 | cl = repo.changelog | |
37 | working = len(cl) # pseudo rev for the working directory |
|
37 | working = len(cl) # pseudo rev for the working directory | |
38 | if a is None: |
|
38 | if a is None: | |
39 | a = working |
|
39 | a = working | |
40 | if b is None: |
|
40 | if b is None: | |
41 | b = working |
|
41 | b = working | |
42 |
|
42 | |||
43 | side = {a: -1, b: 1} |
|
43 | side = {a: -1, b: 1} | |
44 | visit = [-a, -b] |
|
44 | visit = [-a, -b] | |
45 | heapq.heapify(visit) |
|
45 | heapq.heapify(visit) | |
46 | interesting = len(visit) |
|
46 | interesting = len(visit) | |
47 | hascommonancestor = False |
|
47 | hascommonancestor = False | |
48 | limit = working |
|
48 | limit = working | |
49 |
|
49 | |||
50 | while interesting: |
|
50 | while interesting: | |
51 | r = -heapq.heappop(visit) |
|
51 | r = -heapq.heappop(visit) | |
52 | if r == working: |
|
52 | if r == working: | |
53 | parents = [cl.rev(p) for p in repo.dirstate.parents()] |
|
53 | parents = [cl.rev(p) for p in repo.dirstate.parents()] | |
54 | else: |
|
54 | else: | |
55 | parents = cl.parentrevs(r) |
|
55 | parents = cl.parentrevs(r) | |
56 | for p in parents: |
|
56 | for p in parents: | |
57 | if p < 0: |
|
57 | if p < 0: | |
58 | continue |
|
58 | continue | |
59 | if p not in side: |
|
59 | if p not in side: | |
60 | # first time we see p; add it to visit |
|
60 | # first time we see p; add it to visit | |
61 | side[p] = side[r] |
|
61 | side[p] = side[r] | |
62 | if side[p]: |
|
62 | if side[p]: | |
63 | interesting += 1 |
|
63 | interesting += 1 | |
64 | heapq.heappush(visit, -p) |
|
64 | heapq.heappush(visit, -p) | |
65 | elif side[p] and side[p] != side[r]: |
|
65 | elif side[p] and side[p] != side[r]: | |
66 | # p was interesting but now we know better |
|
66 | # p was interesting but now we know better | |
67 | side[p] = 0 |
|
67 | side[p] = 0 | |
68 | interesting -= 1 |
|
68 | interesting -= 1 | |
69 | hascommonancestor = True |
|
69 | hascommonancestor = True | |
70 | if side[r]: |
|
70 | if side[r]: | |
71 | limit = r # lowest rev visited |
|
71 | limit = r # lowest rev visited | |
72 | interesting -= 1 |
|
72 | interesting -= 1 | |
73 |
|
73 | |||
74 | if not hascommonancestor: |
|
74 | if not hascommonancestor: | |
75 | return None |
|
75 | return None | |
76 | return limit |
|
76 | return limit | |
77 |
|
77 | |||
78 | def _chain(src, dst, a, b): |
|
78 | def _chain(src, dst, a, b): | |
79 | '''chain two sets of copies a->b''' |
|
79 | '''chain two sets of copies a->b''' | |
80 | t = a.copy() |
|
80 | t = a.copy() | |
81 | for k, v in b.iteritems(): |
|
81 | for k, v in b.iteritems(): | |
82 | if v in t: |
|
82 | if v in t: | |
83 | # found a chain |
|
83 | # found a chain | |
84 | if t[v] != k: |
|
84 | if t[v] != k: | |
85 | # file wasn't renamed back to itself |
|
85 | # file wasn't renamed back to itself | |
86 | t[k] = t[v] |
|
86 | t[k] = t[v] | |
87 | if v not in dst: |
|
87 | if v not in dst: | |
88 | # chain was a rename, not a copy |
|
88 | # chain was a rename, not a copy | |
89 | del t[v] |
|
89 | del t[v] | |
90 | if v in src: |
|
90 | if v in src: | |
91 | # file is a copy of an existing file |
|
91 | # file is a copy of an existing file | |
92 | t[k] = v |
|
92 | t[k] = v | |
93 |
|
93 | |||
94 | # remove criss-crossed copies |
|
94 | # remove criss-crossed copies | |
95 | for k, v in t.items(): |
|
95 | for k, v in t.items(): | |
96 | if k in src and v in dst: |
|
96 | if k in src and v in dst: | |
97 | del t[k] |
|
97 | del t[k] | |
98 |
|
98 | |||
99 | return t |
|
99 | return t | |
100 |
|
100 | |||
101 | def _tracefile(fctx, am, limit=-1): |
|
101 | def _tracefile(fctx, am, limit=-1): | |
102 | '''return file context that is the ancestor of fctx present in ancestor |
|
102 | '''return file context that is the ancestor of fctx present in ancestor | |
103 | manifest am, stopping after the first ancestor lower than limit''' |
|
103 | manifest am, stopping after the first ancestor lower than limit''' | |
104 |
|
104 | |||
105 | for f in fctx.ancestors(): |
|
105 | for f in fctx.ancestors(): | |
106 | if am.get(f.path(), None) == f.filenode(): |
|
106 | if am.get(f.path(), None) == f.filenode(): | |
107 | return f |
|
107 | return f | |
108 | if f.rev() < limit: |
|
108 | if f.rev() < limit: | |
109 | return None |
|
109 | return None | |
110 |
|
110 | |||
111 | def _dirstatecopies(d): |
|
111 | def _dirstatecopies(d): | |
112 | ds = d._repo.dirstate |
|
112 | ds = d._repo.dirstate | |
113 | c = ds.copies().copy() |
|
113 | c = ds.copies().copy() | |
114 | for k in c.keys(): |
|
114 | for k in c.keys(): | |
115 | if ds[k] not in 'anm': |
|
115 | if ds[k] not in 'anm': | |
116 | del c[k] |
|
116 | del c[k] | |
117 | return c |
|
117 | return c | |
118 |
|
118 | |||
119 | def _forwardcopies(a, b): |
|
119 | def _forwardcopies(a, b): | |
120 | '''find {dst@b: src@a} copy mapping where a is an ancestor of b''' |
|
120 | '''find {dst@b: src@a} copy mapping where a is an ancestor of b''' | |
121 |
|
121 | |||
122 | # check for working copy |
|
122 | # check for working copy | |
123 | w = None |
|
123 | w = None | |
124 | if b.rev() is None: |
|
124 | if b.rev() is None: | |
125 | w = b |
|
125 | w = b | |
126 | b = w.p1() |
|
126 | b = w.p1() | |
127 | if a == b: |
|
127 | if a == b: | |
128 | # short-circuit to avoid issues with merge states |
|
128 | # short-circuit to avoid issues with merge states | |
129 | return _dirstatecopies(w) |
|
129 | return _dirstatecopies(w) | |
130 |
|
130 | |||
131 | # files might have to be traced back to the fctx parent of the last |
|
131 | # files might have to be traced back to the fctx parent of the last | |
132 | # one-side-only changeset, but not further back than that |
|
132 | # one-side-only changeset, but not further back than that | |
133 | limit = _findlimit(a._repo, a.rev(), b.rev()) |
|
133 | limit = _findlimit(a._repo, a.rev(), b.rev()) | |
134 | if limit is None: |
|
134 | if limit is None: | |
135 | limit = -1 |
|
135 | limit = -1 | |
136 | am = a.manifest() |
|
136 | am = a.manifest() | |
137 |
|
137 | |||
138 | # find where new files came from |
|
138 | # find where new files came from | |
139 | # we currently don't try to find where old files went, too expensive |
|
139 | # we currently don't try to find where old files went, too expensive | |
140 | # this means we can miss a case like 'hg rm b; hg cp a b' |
|
140 | # this means we can miss a case like 'hg rm b; hg cp a b' | |
141 | cm = {} |
|
141 | cm = {} | |
142 | missing = set(b.manifest().iterkeys()) |
|
142 | missing = set(b.manifest().iterkeys()) | |
143 | missing.difference_update(a.manifest().iterkeys()) |
|
143 | missing.difference_update(a.manifest().iterkeys()) | |
144 |
|
144 | |||
145 | for f in missing: |
|
145 | for f in missing: | |
146 | ofctx = _tracefile(b[f], am, limit) |
|
146 | ofctx = _tracefile(b[f], am, limit) | |
147 | if ofctx: |
|
147 | if ofctx: | |
148 | cm[f] = ofctx.path() |
|
148 | cm[f] = ofctx.path() | |
149 |
|
149 | |||
150 | # combine copies from dirstate if necessary |
|
150 | # combine copies from dirstate if necessary | |
151 | if w is not None: |
|
151 | if w is not None: | |
152 | cm = _chain(a, w, cm, _dirstatecopies(w)) |
|
152 | cm = _chain(a, w, cm, _dirstatecopies(w)) | |
153 |
|
153 | |||
154 | return cm |
|
154 | return cm | |
155 |
|
155 | |||
156 | def _backwardrenames(a, b): |
|
156 | def _backwardrenames(a, b): | |
157 | # Even though we're not taking copies into account, 1:n rename situations |
|
157 | # Even though we're not taking copies into account, 1:n rename situations | |
158 | # can still exist (e.g. hg cp a b; hg mv a c). In those cases we |
|
158 | # can still exist (e.g. hg cp a b; hg mv a c). In those cases we | |
159 | # arbitrarily pick one of the renames. |
|
159 | # arbitrarily pick one of the renames. | |
160 | f = _forwardcopies(b, a) |
|
160 | f = _forwardcopies(b, a) | |
161 | r = {} |
|
161 | r = {} | |
162 | for k, v in sorted(f.iteritems()): |
|
162 | for k, v in sorted(f.iteritems()): | |
163 | # remove copies |
|
163 | # remove copies | |
164 | if v in a: |
|
164 | if v in a: | |
165 | continue |
|
165 | continue | |
166 | r[v] = k |
|
166 | r[v] = k | |
167 | return r |
|
167 | return r | |
168 |
|
168 | |||
169 | def pathcopies(x, y): |
|
169 | def pathcopies(x, y): | |
170 | '''find {dst@y: src@x} copy mapping for directed compare''' |
|
170 | '''find {dst@y: src@x} copy mapping for directed compare''' | |
171 | if x == y or not x or not y: |
|
171 | if x == y or not x or not y: | |
172 | return {} |
|
172 | return {} | |
173 | a = y.ancestor(x) |
|
173 | a = y.ancestor(x) | |
174 | if a == x: |
|
174 | if a == x: | |
175 | return _forwardcopies(x, y) |
|
175 | return _forwardcopies(x, y) | |
176 | if a == y: |
|
176 | if a == y: | |
177 | return _backwardrenames(x, y) |
|
177 | return _backwardrenames(x, y) | |
178 | return _chain(x, y, _backwardrenames(x, a), _forwardcopies(a, y)) |
|
178 | return _chain(x, y, _backwardrenames(x, a), _forwardcopies(a, y)) | |
179 |
|
179 | |||
180 | def mergecopies(repo, c1, c2, ca): |
|
180 | def mergecopies(repo, c1, c2, ca): | |
181 | """ |
|
181 | """ | |
182 | Find moves and copies between context c1 and c2 that are relevant |
|
182 | Find moves and copies between context c1 and c2 that are relevant | |
183 | for merging. |
|
183 | for merging. | |
184 |
|
184 | |||
185 | Returns four dicts: "copy", "movewithdir", "diverge", and |
|
185 | Returns four dicts: "copy", "movewithdir", "diverge", and | |
186 | "renamedelete". |
|
186 | "renamedelete". | |
187 |
|
187 | |||
188 | "copy" is a mapping from destination name -> source name, |
|
188 | "copy" is a mapping from destination name -> source name, | |
189 | where source is in c1 and destination is in c2 or vice-versa. |
|
189 | where source is in c1 and destination is in c2 or vice-versa. | |
190 |
|
190 | |||
191 | "movewithdir" is a mapping from source name -> destination name, |
|
191 | "movewithdir" is a mapping from source name -> destination name, | |
192 | where the file at source present in one context but not the other |
|
192 | where the file at source present in one context but not the other | |
193 | needs to be moved to destination by the merge process, because the |
|
193 | needs to be moved to destination by the merge process, because the | |
194 | other context moved the directory it is in. |
|
194 | other context moved the directory it is in. | |
195 |
|
195 | |||
196 | "diverge" is a mapping of source name -> list of destination names |
|
196 | "diverge" is a mapping of source name -> list of destination names | |
197 | for divergent renames. |
|
197 | for divergent renames. | |
198 |
|
198 | |||
199 | "renamedelete" is a mapping of source name -> list of destination |
|
199 | "renamedelete" is a mapping of source name -> list of destination | |
200 | names for files deleted in c1 that were renamed in c2 or vice-versa. |
|
200 | names for files deleted in c1 that were renamed in c2 or vice-versa. | |
201 | """ |
|
201 | """ | |
202 | # avoid silly behavior for update from empty dir |
|
202 | # avoid silly behavior for update from empty dir | |
203 | if not c1 or not c2 or c1 == c2: |
|
203 | if not c1 or not c2 or c1 == c2: | |
204 | return {}, {}, {}, {} |
|
204 | return {}, {}, {}, {} | |
205 |
|
205 | |||
206 | # avoid silly behavior for parent -> working dir |
|
206 | # avoid silly behavior for parent -> working dir | |
207 | if c2.node() is None and c1.node() == repo.dirstate.p1(): |
|
207 | if c2.node() is None and c1.node() == repo.dirstate.p1(): | |
208 | return repo.dirstate.copies(), {}, {}, {} |
|
208 | return repo.dirstate.copies(), {}, {}, {} | |
209 |
|
209 | |||
210 | limit = _findlimit(repo, c1.rev(), c2.rev()) |
|
210 | limit = _findlimit(repo, c1.rev(), c2.rev()) | |
211 | if limit is None: |
|
211 | if limit is None: | |
212 | # no common ancestor, no copies |
|
212 | # no common ancestor, no copies | |
213 | return {}, {}, {}, {} |
|
213 | return {}, {}, {}, {} | |
214 | m1 = c1.manifest() |
|
214 | m1 = c1.manifest() | |
215 | m2 = c2.manifest() |
|
215 | m2 = c2.manifest() | |
216 | ma = ca.manifest() |
|
216 | ma = ca.manifest() | |
217 |
|
217 | |||
218 | def makectx(f, n): |
|
218 | def makectx(f, n): | |
219 | if len(n) != 20: # in a working context? |
|
219 | if len(n) != 20: # in a working context? | |
220 | if c1.rev() is None: |
|
220 | if c1.rev() is None: | |
221 | return c1.filectx(f) |
|
221 | return c1.filectx(f) | |
222 | return c2.filectx(f) |
|
222 | return c2.filectx(f) | |
223 | return repo.filectx(f, fileid=n) |
|
223 | return repo.filectx(f, fileid=n) | |
224 |
|
224 | |||
225 | ctx = util.lrucachefunc(makectx) |
|
225 | ctx = util.lrucachefunc(makectx) | |
226 | copy = {} |
|
226 | copy = {} | |
227 | movewithdir = {} |
|
227 | movewithdir = {} | |
228 | fullcopy = {} |
|
228 | fullcopy = {} | |
229 | diverge = {} |
|
229 | diverge = {} | |
230 |
|
230 | |||
231 | repo.ui.debug(" searching for copies back to rev %d\n" % limit) |
|
231 | repo.ui.debug(" searching for copies back to rev %d\n" % limit) | |
232 |
|
232 | |||
233 | u1 = _nonoverlap(m1, m2, ma) |
|
233 | u1 = _nonoverlap(m1, m2, ma) | |
234 | u2 = _nonoverlap(m2, m1, ma) |
|
234 | u2 = _nonoverlap(m2, m1, ma) | |
235 |
|
235 | |||
236 | if u1: |
|
236 | if u1: | |
237 | repo.ui.debug(" unmatched files in local:\n %s\n" |
|
237 | repo.ui.debug(" unmatched files in local:\n %s\n" | |
238 | % "\n ".join(u1)) |
|
238 | % "\n ".join(u1)) | |
239 | if u2: |
|
239 | if u2: | |
240 | repo.ui.debug(" unmatched files in other:\n %s\n" |
|
240 | repo.ui.debug(" unmatched files in other:\n %s\n" | |
241 | % "\n ".join(u2)) |
|
241 | % "\n ".join(u2)) | |
242 |
|
242 | |||
243 | for f in u1: |
|
243 | for f in u1: | |
244 | checkcopies(ctx, f, m1, m2, ca, limit, diverge, copy, fullcopy) |
|
244 | checkcopies(ctx, f, m1, m2, ca, limit, diverge, copy, fullcopy) | |
245 |
|
245 | |||
246 | for f in u2: |
|
246 | for f in u2: | |
247 | checkcopies(ctx, f, m2, m1, ca, limit, diverge, copy, fullcopy) |
|
247 | checkcopies(ctx, f, m2, m1, ca, limit, diverge, copy, fullcopy) | |
248 |
|
248 | |||
249 | renamedelete = {} |
|
249 | renamedelete = {} | |
250 | renamedelete2 = set() |
|
250 | renamedelete2 = set() | |
251 | diverge2 = set() |
|
251 | diverge2 = set() | |
252 | for of, fl in diverge.items(): |
|
252 | for of, fl in diverge.items(): | |
253 | if len(fl) == 1 or of in c1 or of in c2: |
|
253 | if len(fl) == 1 or of in c1 or of in c2: | |
254 | del diverge[of] # not actually divergent, or not a rename |
|
254 | del diverge[of] # not actually divergent, or not a rename | |
255 | if of not in c1 and of not in c2: |
|
255 | if of not in c1 and of not in c2: | |
256 | # renamed on one side, deleted on the other side, but filter |
|
256 | # renamed on one side, deleted on the other side, but filter | |
257 | # out files that have been renamed and then deleted |
|
257 | # out files that have been renamed and then deleted | |
258 | renamedelete[of] = [f for f in fl if f in c1 or f in c2] |
|
258 | renamedelete[of] = [f for f in fl if f in c1 or f in c2] | |
259 | renamedelete2.update(fl) # reverse map for below |
|
259 | renamedelete2.update(fl) # reverse map for below | |
260 | else: |
|
260 | else: | |
261 | diverge2.update(fl) # reverse map for below |
|
261 | diverge2.update(fl) # reverse map for below | |
262 |
|
262 | |||
263 | bothnew = sorted([d for d in m1 if d in m2 and d not in ma]) |
|
263 | bothnew = sorted([d for d in m1 if d in m2 and d not in ma]) | |
264 | if bothnew: |
|
264 | if bothnew: | |
265 | repo.ui.debug(" unmatched files new in both:\n %s\n" |
|
265 | repo.ui.debug(" unmatched files new in both:\n %s\n" | |
266 | % "\n ".join(bothnew)) |
|
266 | % "\n ".join(bothnew)) | |
267 | bothdiverge, _copy, _fullcopy = {}, {}, {} |
|
267 | bothdiverge, _copy, _fullcopy = {}, {}, {} | |
268 | for f in bothnew: |
|
268 | for f in bothnew: | |
269 | checkcopies(ctx, f, m1, m2, ca, limit, bothdiverge, _copy, _fullcopy) |
|
269 | checkcopies(ctx, f, m1, m2, ca, limit, bothdiverge, _copy, _fullcopy) | |
270 | checkcopies(ctx, f, m2, m1, ca, limit, bothdiverge, _copy, _fullcopy) |
|
270 | checkcopies(ctx, f, m2, m1, ca, limit, bothdiverge, _copy, _fullcopy) | |
271 | for of, fl in bothdiverge.items(): |
|
271 | for of, fl in bothdiverge.items(): | |
272 | if len(fl) == 2 and fl[0] == fl[1]: |
|
272 | if len(fl) == 2 and fl[0] == fl[1]: | |
273 | copy[fl[0]] = of # not actually divergent, just matching renames |
|
273 | copy[fl[0]] = of # not actually divergent, just matching renames | |
274 |
|
274 | |||
275 | if fullcopy: |
|
275 | if fullcopy and repo.ui.debugflag: | |
276 | repo.ui.debug(" all copies found (* = to merge, ! = divergent, " |
|
276 | repo.ui.debug(" all copies found (* = to merge, ! = divergent, " | |
277 | "% = renamed and deleted):\n") |
|
277 | "% = renamed and deleted):\n") | |
278 | for f in sorted(fullcopy): |
|
278 | for f in sorted(fullcopy): | |
279 | note = "" |
|
279 | note = "" | |
280 | if f in copy: |
|
280 | if f in copy: | |
281 | note += "*" |
|
281 | note += "*" | |
282 | if f in diverge2: |
|
282 | if f in diverge2: | |
283 | note += "!" |
|
283 | note += "!" | |
284 | if f in renamedelete2: |
|
284 | if f in renamedelete2: | |
285 | note += "%" |
|
285 | note += "%" | |
286 | repo.ui.debug(" src: '%s' -> dst: '%s' %s\n" % (fullcopy[f], f, |
|
286 | repo.ui.debug(" src: '%s' -> dst: '%s' %s\n" % (fullcopy[f], f, | |
287 | note)) |
|
287 | note)) | |
288 | del diverge2 |
|
288 | del diverge2 | |
289 |
|
289 | |||
290 | if not fullcopy: |
|
290 | if not fullcopy: | |
291 | return copy, movewithdir, diverge, renamedelete |
|
291 | return copy, movewithdir, diverge, renamedelete | |
292 |
|
292 | |||
293 | repo.ui.debug(" checking for directory renames\n") |
|
293 | repo.ui.debug(" checking for directory renames\n") | |
294 |
|
294 | |||
295 | # generate a directory move map |
|
295 | # generate a directory move map | |
296 | d1, d2 = c1.dirs(), c2.dirs() |
|
296 | d1, d2 = c1.dirs(), c2.dirs() | |
297 | d1.addpath('/') |
|
297 | d1.addpath('/') | |
298 | d2.addpath('/') |
|
298 | d2.addpath('/') | |
299 | invalid = set() |
|
299 | invalid = set() | |
300 | dirmove = {} |
|
300 | dirmove = {} | |
301 |
|
301 | |||
302 | # examine each file copy for a potential directory move, which is |
|
302 | # examine each file copy for a potential directory move, which is | |
303 | # when all the files in a directory are moved to a new directory |
|
303 | # when all the files in a directory are moved to a new directory | |
304 | for dst, src in fullcopy.iteritems(): |
|
304 | for dst, src in fullcopy.iteritems(): | |
305 | dsrc, ddst = _dirname(src), _dirname(dst) |
|
305 | dsrc, ddst = _dirname(src), _dirname(dst) | |
306 | if dsrc in invalid: |
|
306 | if dsrc in invalid: | |
307 | # already seen to be uninteresting |
|
307 | # already seen to be uninteresting | |
308 | continue |
|
308 | continue | |
309 | elif dsrc in d1 and ddst in d1: |
|
309 | elif dsrc in d1 and ddst in d1: | |
310 | # directory wasn't entirely moved locally |
|
310 | # directory wasn't entirely moved locally | |
311 | invalid.add(dsrc) |
|
311 | invalid.add(dsrc) | |
312 | elif dsrc in d2 and ddst in d2: |
|
312 | elif dsrc in d2 and ddst in d2: | |
313 | # directory wasn't entirely moved remotely |
|
313 | # directory wasn't entirely moved remotely | |
314 | invalid.add(dsrc) |
|
314 | invalid.add(dsrc) | |
315 | elif dsrc in dirmove and dirmove[dsrc] != ddst: |
|
315 | elif dsrc in dirmove and dirmove[dsrc] != ddst: | |
316 | # files from the same directory moved to two different places |
|
316 | # files from the same directory moved to two different places | |
317 | invalid.add(dsrc) |
|
317 | invalid.add(dsrc) | |
318 | else: |
|
318 | else: | |
319 | # looks good so far |
|
319 | # looks good so far | |
320 | dirmove[dsrc + "/"] = ddst + "/" |
|
320 | dirmove[dsrc + "/"] = ddst + "/" | |
321 |
|
321 | |||
322 | for i in invalid: |
|
322 | for i in invalid: | |
323 | if i in dirmove: |
|
323 | if i in dirmove: | |
324 | del dirmove[i] |
|
324 | del dirmove[i] | |
325 | del d1, d2, invalid |
|
325 | del d1, d2, invalid | |
326 |
|
326 | |||
327 | if not dirmove: |
|
327 | if not dirmove: | |
328 | return copy, movewithdir, diverge, renamedelete |
|
328 | return copy, movewithdir, diverge, renamedelete | |
329 |
|
329 | |||
330 | for d in dirmove: |
|
330 | for d in dirmove: | |
331 | repo.ui.debug(" discovered dir src: '%s' -> dst: '%s'\n" % |
|
331 | repo.ui.debug(" discovered dir src: '%s' -> dst: '%s'\n" % | |
332 | (d, dirmove[d])) |
|
332 | (d, dirmove[d])) | |
333 |
|
333 | |||
334 | # check unaccounted nonoverlapping files against directory moves |
|
334 | # check unaccounted nonoverlapping files against directory moves | |
335 | for f in u1 + u2: |
|
335 | for f in u1 + u2: | |
336 | if f not in fullcopy: |
|
336 | if f not in fullcopy: | |
337 | for d in dirmove: |
|
337 | for d in dirmove: | |
338 | if f.startswith(d): |
|
338 | if f.startswith(d): | |
339 | # new file added in a directory that was moved, move it |
|
339 | # new file added in a directory that was moved, move it | |
340 | df = dirmove[d] + f[len(d):] |
|
340 | df = dirmove[d] + f[len(d):] | |
341 | if df not in copy: |
|
341 | if df not in copy: | |
342 | movewithdir[f] = df |
|
342 | movewithdir[f] = df | |
343 | repo.ui.debug((" pending file src: '%s' -> " |
|
343 | repo.ui.debug((" pending file src: '%s' -> " | |
344 | "dst: '%s'\n") % (f, df)) |
|
344 | "dst: '%s'\n") % (f, df)) | |
345 | break |
|
345 | break | |
346 |
|
346 | |||
347 | return copy, movewithdir, diverge, renamedelete |
|
347 | return copy, movewithdir, diverge, renamedelete | |
348 |
|
348 | |||
349 | def checkcopies(ctx, f, m1, m2, ca, limit, diverge, copy, fullcopy): |
|
349 | def checkcopies(ctx, f, m1, m2, ca, limit, diverge, copy, fullcopy): | |
350 | """ |
|
350 | """ | |
351 | check possible copies of f from m1 to m2 |
|
351 | check possible copies of f from m1 to m2 | |
352 |
|
352 | |||
353 | ctx = function accepting (filename, node) that returns a filectx. |
|
353 | ctx = function accepting (filename, node) that returns a filectx. | |
354 | f = the filename to check |
|
354 | f = the filename to check | |
355 | m1 = the source manifest |
|
355 | m1 = the source manifest | |
356 | m2 = the destination manifest |
|
356 | m2 = the destination manifest | |
357 | ca = the changectx of the common ancestor |
|
357 | ca = the changectx of the common ancestor | |
358 | limit = the rev number to not search beyond |
|
358 | limit = the rev number to not search beyond | |
359 | diverge = record all diverges in this dict |
|
359 | diverge = record all diverges in this dict | |
360 | copy = record all non-divergent copies in this dict |
|
360 | copy = record all non-divergent copies in this dict | |
361 | fullcopy = record all copies in this dict |
|
361 | fullcopy = record all copies in this dict | |
362 | """ |
|
362 | """ | |
363 |
|
363 | |||
364 | ma = ca.manifest() |
|
364 | ma = ca.manifest() | |
365 |
|
365 | |||
366 | def _related(f1, f2, limit): |
|
366 | def _related(f1, f2, limit): | |
367 | # Walk back to common ancestor to see if the two files originate |
|
367 | # Walk back to common ancestor to see if the two files originate | |
368 | # from the same file. Since workingfilectx's rev() is None it messes |
|
368 | # from the same file. Since workingfilectx's rev() is None it messes | |
369 | # up the integer comparison logic, hence the pre-step check for |
|
369 | # up the integer comparison logic, hence the pre-step check for | |
370 | # None (f1 and f2 can only be workingfilectx's initially). |
|
370 | # None (f1 and f2 can only be workingfilectx's initially). | |
371 |
|
371 | |||
372 | if f1 == f2: |
|
372 | if f1 == f2: | |
373 | return f1 # a match |
|
373 | return f1 # a match | |
374 |
|
374 | |||
375 | g1, g2 = f1.ancestors(), f2.ancestors() |
|
375 | g1, g2 = f1.ancestors(), f2.ancestors() | |
376 | try: |
|
376 | try: | |
377 | f1r, f2r = f1.rev(), f2.rev() |
|
377 | f1r, f2r = f1.rev(), f2.rev() | |
378 |
|
378 | |||
379 | if f1r is None: |
|
379 | if f1r is None: | |
380 | f1 = g1.next() |
|
380 | f1 = g1.next() | |
381 | if f2r is None: |
|
381 | if f2r is None: | |
382 | f2 = g2.next() |
|
382 | f2 = g2.next() | |
383 |
|
383 | |||
384 | while True: |
|
384 | while True: | |
385 | f1r, f2r = f1.rev(), f2.rev() |
|
385 | f1r, f2r = f1.rev(), f2.rev() | |
386 | if f1r > f2r: |
|
386 | if f1r > f2r: | |
387 | f1 = g1.next() |
|
387 | f1 = g1.next() | |
388 | elif f2r > f1r: |
|
388 | elif f2r > f1r: | |
389 | f2 = g2.next() |
|
389 | f2 = g2.next() | |
390 | elif f1 == f2: |
|
390 | elif f1 == f2: | |
391 | return f1 # a match |
|
391 | return f1 # a match | |
392 | elif f1r == f2r or f1r < limit or f2r < limit: |
|
392 | elif f1r == f2r or f1r < limit or f2r < limit: | |
393 | return False # copy no longer relevant |
|
393 | return False # copy no longer relevant | |
394 | except StopIteration: |
|
394 | except StopIteration: | |
395 | return False |
|
395 | return False | |
396 |
|
396 | |||
397 | of = None |
|
397 | of = None | |
398 | seen = set([f]) |
|
398 | seen = set([f]) | |
399 | for oc in ctx(f, m1[f]).ancestors(): |
|
399 | for oc in ctx(f, m1[f]).ancestors(): | |
400 | ocr = oc.rev() |
|
400 | ocr = oc.rev() | |
401 | of = oc.path() |
|
401 | of = oc.path() | |
402 | if of in seen: |
|
402 | if of in seen: | |
403 | # check limit late - grab last rename before |
|
403 | # check limit late - grab last rename before | |
404 | if ocr < limit: |
|
404 | if ocr < limit: | |
405 | break |
|
405 | break | |
406 | continue |
|
406 | continue | |
407 | seen.add(of) |
|
407 | seen.add(of) | |
408 |
|
408 | |||
409 | fullcopy[f] = of # remember for dir rename detection |
|
409 | fullcopy[f] = of # remember for dir rename detection | |
410 | if of not in m2: |
|
410 | if of not in m2: | |
411 | continue # no match, keep looking |
|
411 | continue # no match, keep looking | |
412 | if m2[of] == ma.get(of): |
|
412 | if m2[of] == ma.get(of): | |
413 | break # no merge needed, quit early |
|
413 | break # no merge needed, quit early | |
414 | c2 = ctx(of, m2[of]) |
|
414 | c2 = ctx(of, m2[of]) | |
415 | cr = _related(oc, c2, ca.rev()) |
|
415 | cr = _related(oc, c2, ca.rev()) | |
416 | if cr and (of == f or of == c2.path()): # non-divergent |
|
416 | if cr and (of == f or of == c2.path()): # non-divergent | |
417 | copy[f] = of |
|
417 | copy[f] = of | |
418 | of = None |
|
418 | of = None | |
419 | break |
|
419 | break | |
420 |
|
420 | |||
421 | if of in ma: |
|
421 | if of in ma: | |
422 | diverge.setdefault(of, []).append(f) |
|
422 | diverge.setdefault(of, []).append(f) |
General Comments 0
You need to be logged in to leave comments.
Login now