Show More
@@ -235,6 +235,23 b' static inline int dirstate_item_c_v1_mti' | |||||
235 | } |
|
235 | } | |
236 | } |
|
236 | } | |
237 |
|
237 | |||
|
238 | static PyObject *dirstate_item_v2_data(dirstateItemObject *self) | |||
|
239 | { | |||
|
240 | unsigned char flags = self->flags; | |||
|
241 | int mode = dirstate_item_c_v1_mode(self); | |||
|
242 | if ((mode & S_IXUSR) != 0) { | |||
|
243 | flags |= dirstate_flag_mode_exec_perm; | |||
|
244 | } else { | |||
|
245 | flags &= ~dirstate_flag_mode_exec_perm; | |||
|
246 | } | |||
|
247 | if (S_ISLNK(mode)) { | |||
|
248 | flags |= dirstate_flag_mode_is_symlink; | |||
|
249 | } else { | |||
|
250 | flags &= ~dirstate_flag_mode_is_symlink; | |||
|
251 | } | |||
|
252 | return Py_BuildValue("Bii", flags, self->size, self->mtime); | |||
|
253 | }; | |||
|
254 | ||||
238 | static PyObject *dirstate_item_v1_state(dirstateItemObject *self) |
|
255 | static PyObject *dirstate_item_v1_state(dirstateItemObject *self) | |
239 | { |
|
256 | { | |
240 | char state = dirstate_item_c_v1_state(self); |
|
257 | char state = dirstate_item_c_v1_state(self); | |
@@ -428,6 +445,8 b' static PyObject *dirstate_item_drop_merg' | |||||
428 | Py_RETURN_NONE; |
|
445 | Py_RETURN_NONE; | |
429 | } |
|
446 | } | |
430 | static PyMethodDef dirstate_item_methods[] = { |
|
447 | static PyMethodDef dirstate_item_methods[] = { | |
|
448 | {"v2_data", (PyCFunction)dirstate_item_v2_data, METH_NOARGS, | |||
|
449 | "return data suitable for v2 serialization"}, | |||
431 | {"v1_state", (PyCFunction)dirstate_item_v1_state, METH_NOARGS, |
|
450 | {"v1_state", (PyCFunction)dirstate_item_v1_state, METH_NOARGS, | |
432 | "return a \"state\" suitable for v1 serialization"}, |
|
451 | "return a \"state\" suitable for v1 serialization"}, | |
433 | {"v1_mode", (PyCFunction)dirstate_item_v1_mode, METH_NOARGS, |
|
452 | {"v1_mode", (PyCFunction)dirstate_item_v1_mode, METH_NOARGS, |
@@ -9,7 +9,8 b' from __future__ import absolute_import' | |||||
9 |
|
9 | |||
10 | import struct |
|
10 | import struct | |
11 |
|
11 | |||
12 |
from .. import |
|
12 | from ..thirdparty import attr | |
|
13 | from .. import error, policy | |||
13 |
|
14 | |||
14 | parsers = policy.importmod('parsers') |
|
15 | parsers = policy.importmod('parsers') | |
15 |
|
16 | |||
@@ -116,3 +117,295 b' def parse_nodes(map, copy_map, data, sta' | |||||
116 |
|
117 | |||
117 | def slice_with_len(data, start, len): |
|
118 | def slice_with_len(data, start, len): | |
118 | return data[start : start + len] |
|
119 | return data[start : start + len] | |
|
120 | ||||
|
121 | ||||
|
122 | @attr.s | |||
|
123 | class Node(object): | |||
|
124 | path = attr.ib() | |||
|
125 | entry = attr.ib() | |||
|
126 | parent = attr.ib(default=None) | |||
|
127 | children_count = attr.ib(default=0) | |||
|
128 | children_offset = attr.ib(default=0) | |||
|
129 | descendants_with_entry = attr.ib(default=0) | |||
|
130 | tracked_descendants = attr.ib(default=0) | |||
|
131 | ||||
|
132 | def pack(self, copy_map, paths_offset): | |||
|
133 | path = self.path | |||
|
134 | copy = copy_map.get(path) | |||
|
135 | entry = self.entry | |||
|
136 | ||||
|
137 | path_start = paths_offset | |||
|
138 | path_len = len(path) | |||
|
139 | basename_start = path.rfind(b'/') + 1 # 0 if rfind returns -1 | |||
|
140 | if copy is not None: | |||
|
141 | copy_source_start = paths_offset + len(path) | |||
|
142 | copy_source_len = len(copy) | |||
|
143 | else: | |||
|
144 | copy_source_start = 0 | |||
|
145 | copy_source_len = 0 | |||
|
146 | if entry is not None: | |||
|
147 | flags, size, mtime_s = entry.v2_data() | |||
|
148 | mtime_ns = 0 | |||
|
149 | else: | |||
|
150 | # There are no mtime-cached directories in the Python implementation | |||
|
151 | flags = 0 | |||
|
152 | mode = 0 | |||
|
153 | size = 0 | |||
|
154 | mtime_s = 0 | |||
|
155 | mtime_ns = 0 | |||
|
156 | return NODE.pack( | |||
|
157 | path_start, | |||
|
158 | path_len, | |||
|
159 | basename_start, | |||
|
160 | copy_source_start, | |||
|
161 | copy_source_len, | |||
|
162 | self.children_offset, | |||
|
163 | self.children_count, | |||
|
164 | self.descendants_with_entry, | |||
|
165 | self.tracked_descendants, | |||
|
166 | flags, | |||
|
167 | size, | |||
|
168 | mtime_s, | |||
|
169 | mtime_ns, | |||
|
170 | ) | |||
|
171 | ||||
|
172 | ||||
|
173 | def pack_dirstate(map, copy_map, now): | |||
|
174 | """ | |||
|
175 | Pack `map` and `copy_map` into the dirstate v2 binary format and return | |||
|
176 | the bytearray. | |||
|
177 | `now` is a timestamp of the current filesystem time used to detect race | |||
|
178 | conditions in writing the dirstate to disk, see inline comment. | |||
|
179 | ||||
|
180 | The on-disk format expects a tree-like structure where the leaves are | |||
|
181 | written first (and sorted per-directory), going up levels until the root | |||
|
182 | node and writing that one to the docket. See more details on the on-disk | |||
|
183 | format in `mercurial/helptext/internals/dirstate-v2`. | |||
|
184 | ||||
|
185 | Since both `map` and `copy_map` are flat dicts we need to figure out the | |||
|
186 | hierarchy. This algorithm does so without having to build the entire tree | |||
|
187 | in-memory: it only keeps the minimum number of nodes around to satisfy the | |||
|
188 | format. | |||
|
189 | ||||
|
190 | # Algorithm explanation | |||
|
191 | ||||
|
192 | This explanation does not talk about the different counters for tracked | |||
|
193 | descendents and storing the copies, but that work is pretty simple once this | |||
|
194 | algorithm is in place. | |||
|
195 | ||||
|
196 | ## Building a subtree | |||
|
197 | ||||
|
198 | First, sort `map`: this makes it so the leaves of the tree are contiguous | |||
|
199 | per directory (i.e. a/b/c and a/b/d will be next to each other in the list), | |||
|
200 | and enables us to use the ordering of folders to have a "cursor" of the | |||
|
201 | current folder we're in without ever going twice in the same branch of the | |||
|
202 | tree. The cursor is a node that remembers its parent and any information | |||
|
203 | relevant to the format (see the `Node` class), building the relevant part | |||
|
204 | of the tree lazily. | |||
|
205 | Then, for each file in `map`, move the cursor into the tree to the | |||
|
206 | corresponding folder of the file: for example, if the very first file | |||
|
207 | is "a/b/c", we start from `Node[""]`, create `Node["a"]` which points to | |||
|
208 | its parent `Node[""]`, then create `Node["a/b"]`, which points to its parent | |||
|
209 | `Node["a"]`. These nodes are kept around in a stack. | |||
|
210 | If the next file in `map` is in the same subtree ("a/b/d" or "a/b/e/f"), we | |||
|
211 | add it to the stack and keep looping with the same logic of creating the | |||
|
212 | tree nodes as needed. If however the next file in `map` is *not* in the same | |||
|
213 | subtree ("a/other", if we're still in the "a/b" folder), then we know that | |||
|
214 | the subtree we're in is complete. | |||
|
215 | ||||
|
216 | ## Writing the subtree | |||
|
217 | ||||
|
218 | We have the entire subtree in the stack, so we start writing it to disk | |||
|
219 | folder by folder. The way we write a folder is to pop the stack into a list | |||
|
220 | until the folder changes, revert this list of direct children (to satisfy | |||
|
221 | the format requirement that children be sorted). This process repeats until | |||
|
222 | we hit the "other" subtree. | |||
|
223 | ||||
|
224 | An example: | |||
|
225 | a | |||
|
226 | dir1/b | |||
|
227 | dir1/c | |||
|
228 | dir2/dir3/d | |||
|
229 | dir2/dir3/e | |||
|
230 | dir2/f | |||
|
231 | ||||
|
232 | Would have us: | |||
|
233 | - add to the stack until "dir2/dir3/e" | |||
|
234 | - realize that "dir2/f" is in a different subtree | |||
|
235 | - pop "dir2/dir3/e", "dir2/dir3/d", reverse them so they're sorted and | |||
|
236 | pack them since the next entry is "dir2/dir3" | |||
|
237 | - go back up to "dir2" | |||
|
238 | - add "dir2/f" to the stack | |||
|
239 | - realize we're done with the map | |||
|
240 | - pop "dir2/f", "dir2/dir3" from the stack, reverse and pack them | |||
|
241 | - go up to the root node, do the same to write "a", "dir1" and "dir2" in | |||
|
242 | that order | |||
|
243 | ||||
|
244 | ## Special case for the root node | |||
|
245 | ||||
|
246 | The root node is not serialized in the format, but its information is | |||
|
247 | written to the docket. Again, see more details on the on-disk format in | |||
|
248 | `mercurial/helptext/internals/dirstate-v2`. | |||
|
249 | """ | |||
|
250 | now = int(now) | |||
|
251 | data = bytearray() | |||
|
252 | root_nodes_start = 0 | |||
|
253 | root_nodes_len = 0 | |||
|
254 | nodes_with_entry_count = 0 | |||
|
255 | nodes_with_copy_source_count = 0 | |||
|
256 | # Will always be 0 since this implementation always re-writes everything | |||
|
257 | # to disk | |||
|
258 | unreachable_bytes = 0 | |||
|
259 | unused = b'\x00' * 4 | |||
|
260 | # This is an optimization that's only useful for the Rust implementation | |||
|
261 | ignore_patterns_hash = b'\x00' * 20 | |||
|
262 | ||||
|
263 | if len(map) == 0: | |||
|
264 | tree_metadata = TREE_METADATA.pack( | |||
|
265 | root_nodes_start, | |||
|
266 | root_nodes_len, | |||
|
267 | nodes_with_entry_count, | |||
|
268 | nodes_with_copy_source_count, | |||
|
269 | unreachable_bytes, | |||
|
270 | unused, | |||
|
271 | ignore_patterns_hash, | |||
|
272 | ) | |||
|
273 | return data, tree_metadata | |||
|
274 | ||||
|
275 | sorted_map = sorted(map.items(), key=lambda x: x[0]) | |||
|
276 | ||||
|
277 | # Use a stack to not have to only remember the nodes we currently need | |||
|
278 | # instead of building the entire tree in memory | |||
|
279 | stack = [] | |||
|
280 | current_node = Node(b"", None) | |||
|
281 | stack.append(current_node) | |||
|
282 | ||||
|
283 | for index, (path, entry) in enumerate(sorted_map, 1): | |||
|
284 | if entry.need_delay(now): | |||
|
285 | # The file was last modified "simultaneously" with the current | |||
|
286 | # write to dirstate (i.e. within the same second for file- | |||
|
287 | # systems with a granularity of 1 sec). This commonly happens | |||
|
288 | # for at least a couple of files on 'update'. | |||
|
289 | # The user could change the file without changing its size | |||
|
290 | # within the same second. Invalidate the file's mtime in | |||
|
291 | # dirstate, forcing future 'status' calls to compare the | |||
|
292 | # contents of the file if the size is the same. This prevents | |||
|
293 | # mistakenly treating such files as clean. | |||
|
294 | entry.set_possibly_dirty() | |||
|
295 | nodes_with_entry_count += 1 | |||
|
296 | if path in copy_map: | |||
|
297 | nodes_with_copy_source_count += 1 | |||
|
298 | current_folder = get_folder(path) | |||
|
299 | current_node = move_to_correct_node_in_tree( | |||
|
300 | current_folder, current_node, stack | |||
|
301 | ) | |||
|
302 | ||||
|
303 | current_node.children_count += 1 | |||
|
304 | # Entries from `map` are never `None` | |||
|
305 | if entry.tracked: | |||
|
306 | current_node.tracked_descendants += 1 | |||
|
307 | current_node.descendants_with_entry += 1 | |||
|
308 | stack.append(Node(path, entry, current_node)) | |||
|
309 | ||||
|
310 | should_pack = True | |||
|
311 | next_path = None | |||
|
312 | if index < len(sorted_map): | |||
|
313 | # Determine if the next entry is in the same sub-tree, if so don't | |||
|
314 | # pack yet | |||
|
315 | next_path = sorted_map[index][0] | |||
|
316 | should_pack = not get_folder(next_path).startswith(current_folder) | |||
|
317 | if should_pack: | |||
|
318 | pack_directory_children(current_node, copy_map, data, stack) | |||
|
319 | while stack and current_node.path != b"": | |||
|
320 | # Go up the tree and write until we reach the folder of the next | |||
|
321 | # entry (if any, otherwise the root) | |||
|
322 | parent = current_node.parent | |||
|
323 | in_parent_folder_of_next_entry = next_path is not None and ( | |||
|
324 | get_folder(next_path).startswith(get_folder(stack[-1].path)) | |||
|
325 | ) | |||
|
326 | if parent is None or in_parent_folder_of_next_entry: | |||
|
327 | break | |||
|
328 | pack_directory_children(parent, copy_map, data, stack) | |||
|
329 | current_node = parent | |||
|
330 | ||||
|
331 | # Special case for the root node since we don't write it to disk, only its | |||
|
332 | # children to the docket | |||
|
333 | current_node = stack.pop() | |||
|
334 | assert current_node.path == b"", current_node.path | |||
|
335 | assert len(stack) == 0, len(stack) | |||
|
336 | ||||
|
337 | tree_metadata = TREE_METADATA.pack( | |||
|
338 | current_node.children_offset, | |||
|
339 | current_node.children_count, | |||
|
340 | nodes_with_entry_count, | |||
|
341 | nodes_with_copy_source_count, | |||
|
342 | unreachable_bytes, | |||
|
343 | unused, | |||
|
344 | ignore_patterns_hash, | |||
|
345 | ) | |||
|
346 | ||||
|
347 | return data, tree_metadata | |||
|
348 | ||||
|
349 | ||||
|
350 | def get_folder(path): | |||
|
351 | """ | |||
|
352 | Return the folder of the path that's given, an empty string for root paths. | |||
|
353 | """ | |||
|
354 | return path.rsplit(b'/', 1)[0] if b'/' in path else b'' | |||
|
355 | ||||
|
356 | ||||
|
357 | def move_to_correct_node_in_tree(target_folder, current_node, stack): | |||
|
358 | """ | |||
|
359 | Move inside the dirstate node tree to the node corresponding to | |||
|
360 | `target_folder`, creating the missing nodes along the way if needed. | |||
|
361 | """ | |||
|
362 | while target_folder != current_node.path: | |||
|
363 | if target_folder.startswith(current_node.path): | |||
|
364 | # We need to go down a folder | |||
|
365 | prefix = target_folder[len(current_node.path) :].lstrip(b'/') | |||
|
366 | subfolder_name = prefix.split(b'/', 1)[0] | |||
|
367 | if current_node.path: | |||
|
368 | subfolder_path = current_node.path + b'/' + subfolder_name | |||
|
369 | else: | |||
|
370 | subfolder_path = subfolder_name | |||
|
371 | next_node = stack[-1] | |||
|
372 | if next_node.path == target_folder: | |||
|
373 | # This folder is now a file and only contains removed entries | |||
|
374 | # merge with the last node | |||
|
375 | current_node = next_node | |||
|
376 | else: | |||
|
377 | current_node.children_count += 1 | |||
|
378 | current_node = Node(subfolder_path, None, current_node) | |||
|
379 | stack.append(current_node) | |||
|
380 | else: | |||
|
381 | # We need to go up a folder | |||
|
382 | current_node = current_node.parent | |||
|
383 | return current_node | |||
|
384 | ||||
|
385 | ||||
|
386 | def pack_directory_children(node, copy_map, data, stack): | |||
|
387 | """ | |||
|
388 | Write the binary representation of the direct sorted children of `node` to | |||
|
389 | `data` | |||
|
390 | """ | |||
|
391 | direct_children = [] | |||
|
392 | ||||
|
393 | while stack[-1].path != b"" and get_folder(stack[-1].path) == node.path: | |||
|
394 | direct_children.append(stack.pop()) | |||
|
395 | if not direct_children: | |||
|
396 | raise error.ProgrammingError(b"no direct children for %r" % node.path) | |||
|
397 | ||||
|
398 | # Reverse the stack to get the correct sorted order | |||
|
399 | direct_children.reverse() | |||
|
400 | packed_children = bytearray() | |||
|
401 | # Write the paths to `data`. Pack child nodes but don't write them yet | |||
|
402 | for child in direct_children: | |||
|
403 | packed = child.pack(copy_map=copy_map, paths_offset=len(data)) | |||
|
404 | packed_children.extend(packed) | |||
|
405 | data.extend(child.path) | |||
|
406 | data.extend(copy_map.get(child.path, b"")) | |||
|
407 | node.tracked_descendants += child.tracked_descendants | |||
|
408 | node.descendants_with_entry += child.descendants_with_entry | |||
|
409 | # Write the fixed-size child nodes all together | |||
|
410 | node.children_offset = len(data) | |||
|
411 | data.extend(packed_children) |
@@ -313,6 +313,25 b' class DirstateItem(object):' | |||||
313 | """True if the file has been removed""" |
|
313 | """True if the file has been removed""" | |
314 | return not self._wc_tracked and (self._p1_tracked or self._p2_info) |
|
314 | return not self._wc_tracked and (self._p1_tracked or self._p2_info) | |
315 |
|
315 | |||
|
316 | def v2_data(self): | |||
|
317 | """Returns (flags, mode, size, mtime) for v2 serialization""" | |||
|
318 | flags = 0 | |||
|
319 | if self._wc_tracked: | |||
|
320 | flags |= DIRSTATE_V2_WDIR_TRACKED | |||
|
321 | if self._p1_tracked: | |||
|
322 | flags |= DIRSTATE_V2_P1_TRACKED | |||
|
323 | if self._p2_info: | |||
|
324 | flags |= DIRSTATE_V2_P2_INFO | |||
|
325 | if self.mode is not None and self.size is not None: | |||
|
326 | flags |= DIRSTATE_V2_HAS_MODE_AND_SIZE | |||
|
327 | if self.mode & stat.S_IXUSR: | |||
|
328 | flags |= DIRSTATE_V2_MODE_EXEC_PERM | |||
|
329 | if stat.S_ISLNK(self.mode): | |||
|
330 | flags |= DIRSTATE_V2_MODE_IS_SYMLINK | |||
|
331 | if self.mtime is not None: | |||
|
332 | flags |= DIRSTATE_V2_HAS_MTIME | |||
|
333 | return (flags, self.size or 0, self.mtime or 0) | |||
|
334 | ||||
316 | def v1_state(self): |
|
335 | def v1_state(self): | |
317 | """return a "state" suitable for v1 serialization""" |
|
336 | """return a "state" suitable for v1 serialization""" | |
318 | if not self.any_tracked: |
|
337 | if not self.any_tracked: |
General Comments 0
You need to be logged in to leave comments.
Login now