Show More
@@ -1,380 +1,380 b'' | |||
|
1 | 1 | #!/usr/bin/env python |
|
2 | 2 | from __future__ import absolute_import, print_function |
|
3 | 3 | |
|
4 | 4 | import hashlib |
|
5 | 5 | import os |
|
6 | 6 | import random |
|
7 | 7 | import shutil |
|
8 | 8 | import stat |
|
9 | 9 | import struct |
|
10 | 10 | import sys |
|
11 | 11 | import tempfile |
|
12 | 12 | import time |
|
13 | 13 | import unittest |
|
14 | 14 | |
|
15 | 15 | import silenttestrunner |
|
16 | 16 | |
|
17 | 17 | # Load the local remotefilelog, not the system one |
|
18 | 18 | sys.path[0:0] = [os.path.join(os.path.dirname(__file__), '..')] |
|
19 | 19 | from mercurial.node import nullid |
|
20 | 20 | from mercurial import ( |
|
21 | 21 | pycompat, |
|
22 | 22 | ui as uimod, |
|
23 | 23 | ) |
|
24 | 24 | from hgext.remotefilelog import ( |
|
25 | 25 | basepack, |
|
26 | 26 | constants, |
|
27 | 27 | datapack, |
|
28 | 28 | ) |
|
29 | 29 | |
|
30 | 30 | class datapacktestsbase(object): |
|
31 | 31 | def __init__(self, datapackreader, paramsavailable): |
|
32 | 32 | self.datapackreader = datapackreader |
|
33 | 33 | self.paramsavailable = paramsavailable |
|
34 | 34 | |
|
35 | 35 | def setUp(self): |
|
36 | 36 | self.tempdirs = [] |
|
37 | 37 | |
|
38 | 38 | def tearDown(self): |
|
39 | 39 | for d in self.tempdirs: |
|
40 | 40 | shutil.rmtree(d) |
|
41 | 41 | |
|
42 | 42 | def makeTempDir(self): |
|
43 | 43 | tempdir = pycompat.bytestr(tempfile.mkdtemp()) |
|
44 | 44 | self.tempdirs.append(tempdir) |
|
45 | 45 | return tempdir |
|
46 | 46 | |
|
47 | 47 | def getHash(self, content): |
|
48 | 48 | return hashlib.sha1(content).digest() |
|
49 | 49 | |
|
50 | 50 | def getFakeHash(self): |
|
51 | 51 | return b''.join(pycompat.bytechr(random.randint(0, 255)) |
|
52 | 52 | for _ in range(20)) |
|
53 | 53 | |
|
54 | 54 | def createPack(self, revisions=None, packdir=None): |
|
55 | 55 | if revisions is None: |
|
56 | 56 | revisions = [(b"filename", self.getFakeHash(), nullid, b"content")] |
|
57 | 57 | |
|
58 | 58 | if packdir is None: |
|
59 | 59 | packdir = self.makeTempDir() |
|
60 | 60 | |
|
61 | 61 | packer = datapack.mutabledatapack(uimod.ui(), packdir, version=2) |
|
62 | 62 | |
|
63 | 63 | for args in revisions: |
|
64 | 64 | filename, node, base, content = args[0:4] |
|
65 | 65 | # meta is optional |
|
66 | 66 | meta = None |
|
67 | 67 | if len(args) > 4: |
|
68 | 68 | meta = args[4] |
|
69 | 69 | packer.add(filename, node, base, content, metadata=meta) |
|
70 | 70 | |
|
71 | 71 | path = packer.close() |
|
72 | 72 | return self.datapackreader(path) |
|
73 | 73 | |
|
74 | 74 | def _testAddSingle(self, content): |
|
75 | 75 | """Test putting a simple blob into a pack and reading it out. |
|
76 | 76 | """ |
|
77 | 77 | filename = b"foo" |
|
78 | 78 | node = self.getHash(content) |
|
79 | 79 | |
|
80 | 80 | revisions = [(filename, node, nullid, content)] |
|
81 | 81 | pack = self.createPack(revisions) |
|
82 | 82 | if self.paramsavailable: |
|
83 | 83 | self.assertEqual(pack.params.fanoutprefix, |
|
84 | 84 | basepack.SMALLFANOUTPREFIX) |
|
85 | 85 | |
|
86 | 86 | chain = pack.getdeltachain(filename, node) |
|
87 | 87 | self.assertEqual(content, chain[0][4]) |
|
88 | 88 | |
|
89 | 89 | def testAddSingle(self): |
|
90 | 90 | self._testAddSingle(b'') |
|
91 | 91 | |
|
92 | 92 | def testAddSingleEmpty(self): |
|
93 | 93 | self._testAddSingle(b'abcdef') |
|
94 | 94 | |
|
95 | 95 | def testAddMultiple(self): |
|
96 | 96 | """Test putting multiple unrelated blobs into a pack and reading them |
|
97 | 97 | out. |
|
98 | 98 | """ |
|
99 | 99 | revisions = [] |
|
100 | 100 | for i in range(10): |
|
101 | 101 | filename = b"foo%d" % i |
|
102 | 102 | content = b"abcdef%d" % i |
|
103 | 103 | node = self.getHash(content) |
|
104 | 104 | revisions.append((filename, node, self.getFakeHash(), content)) |
|
105 | 105 | |
|
106 | 106 | pack = self.createPack(revisions) |
|
107 | 107 | |
|
108 | 108 | for filename, node, base, content in revisions: |
|
109 | 109 | entry = pack.getdelta(filename, node) |
|
110 | 110 | self.assertEqual((content, filename, base, {}), entry) |
|
111 | 111 | |
|
112 | 112 | chain = pack.getdeltachain(filename, node) |
|
113 | 113 | self.assertEqual(content, chain[0][4]) |
|
114 | 114 | |
|
115 | 115 | def testAddDeltas(self): |
|
116 | 116 | """Test putting multiple delta blobs into a pack and read the chain. |
|
117 | 117 | """ |
|
118 | 118 | revisions = [] |
|
119 | 119 | filename = b"foo" |
|
120 | 120 | lastnode = nullid |
|
121 | 121 | for i in range(10): |
|
122 | 122 | content = b"abcdef%d" % i |
|
123 | 123 | node = self.getHash(content) |
|
124 | 124 | revisions.append((filename, node, lastnode, content)) |
|
125 | 125 | lastnode = node |
|
126 | 126 | |
|
127 | 127 | pack = self.createPack(revisions) |
|
128 | 128 | |
|
129 | 129 | entry = pack.getdelta(filename, revisions[0][1]) |
|
130 | 130 | realvalue = (revisions[0][3], filename, revisions[0][2], {}) |
|
131 | 131 | self.assertEqual(entry, realvalue) |
|
132 | 132 | |
|
133 | 133 | # Test that the chain for the final entry has all the others |
|
134 | 134 | chain = pack.getdeltachain(filename, node) |
|
135 | 135 | for i in range(10): |
|
136 | 136 | content = b"abcdef%d" % i |
|
137 | 137 | self.assertEqual(content, chain[-i - 1][4]) |
|
138 | 138 | |
|
139 | 139 | def testPackMany(self): |
|
140 | 140 | """Pack many related and unrelated objects. |
|
141 | 141 | """ |
|
142 | 142 | # Build a random pack file |
|
143 | 143 | revisions = [] |
|
144 | 144 | blobs = {} |
|
145 | 145 | random.seed(0) |
|
146 | 146 | for i in range(100): |
|
147 | 147 | filename = b"filename-%d" % i |
|
148 | 148 | filerevs = [] |
|
149 | 149 | for j in range(random.randint(1, 100)): |
|
150 | 150 | content = b"content-%d" % j |
|
151 | 151 | node = self.getHash(content) |
|
152 | 152 | lastnode = nullid |
|
153 | 153 | if len(filerevs) > 0: |
|
154 | 154 | lastnode = filerevs[random.randint(0, len(filerevs) - 1)] |
|
155 | 155 | filerevs.append(node) |
|
156 | 156 | blobs[(filename, node, lastnode)] = content |
|
157 | 157 | revisions.append((filename, node, lastnode, content)) |
|
158 | 158 | |
|
159 | 159 | pack = self.createPack(revisions) |
|
160 | 160 | |
|
161 | 161 | # Verify the pack contents |
|
162 | 162 | for (filename, node, lastnode), content in sorted(blobs.items()): |
|
163 | 163 | chain = pack.getdeltachain(filename, node) |
|
164 | 164 | for entry in chain: |
|
165 | 165 | expectedcontent = blobs[(entry[0], entry[1], entry[3])] |
|
166 | 166 | self.assertEqual(entry[4], expectedcontent) |
|
167 | 167 | |
|
168 | 168 | def testPackMetadata(self): |
|
169 | 169 | revisions = [] |
|
170 | 170 | for i in range(100): |
|
171 | 171 | filename = b'%d.txt' % i |
|
172 | 172 | content = b'put-something-here \n' * i |
|
173 | 173 | node = self.getHash(content) |
|
174 | 174 | meta = {constants.METAKEYFLAG: i ** 4, |
|
175 | 175 | constants.METAKEYSIZE: len(content), |
|
176 | 176 | b'Z': b'random_string', |
|
177 | 177 | b'_': b'\0' * i} |
|
178 | 178 | revisions.append((filename, node, nullid, content, meta)) |
|
179 | 179 | pack = self.createPack(revisions) |
|
180 | 180 | for name, node, x, content, origmeta in revisions: |
|
181 | 181 | parsedmeta = pack.getmeta(name, node) |
|
182 | 182 | # flag == 0 should be optimized out |
|
183 | 183 | if origmeta[constants.METAKEYFLAG] == 0: |
|
184 | 184 | del origmeta[constants.METAKEYFLAG] |
|
185 | 185 | self.assertEqual(parsedmeta, origmeta) |
|
186 | 186 | |
|
187 | 187 | def testGetMissing(self): |
|
188 | 188 | """Test the getmissing() api. |
|
189 | 189 | """ |
|
190 | 190 | revisions = [] |
|
191 | 191 | filename = b"foo" |
|
192 | 192 | lastnode = nullid |
|
193 | 193 | for i in range(10): |
|
194 | 194 | content = b"abcdef%d" % i |
|
195 | 195 | node = self.getHash(content) |
|
196 | 196 | revisions.append((filename, node, lastnode, content)) |
|
197 | 197 | lastnode = node |
|
198 | 198 | |
|
199 | 199 | pack = self.createPack(revisions) |
|
200 | 200 | |
|
201 | 201 | missing = pack.getmissing([(b"foo", revisions[0][1])]) |
|
202 | 202 | self.assertFalse(missing) |
|
203 | 203 | |
|
204 | 204 | missing = pack.getmissing([(b"foo", revisions[0][1]), |
|
205 | 205 | (b"foo", revisions[1][1])]) |
|
206 | 206 | self.assertFalse(missing) |
|
207 | 207 | |
|
208 | 208 | fakenode = self.getFakeHash() |
|
209 | 209 | missing = pack.getmissing([(b"foo", revisions[0][1]), |
|
210 | 210 | (b"foo", fakenode)]) |
|
211 | 211 | self.assertEqual(missing, [(b"foo", fakenode)]) |
|
212 | 212 | |
|
213 | 213 | def testAddThrows(self): |
|
214 | 214 | pack = self.createPack() |
|
215 | 215 | |
|
216 | 216 | try: |
|
217 | 217 | pack.add(b'filename', nullid, b'contents') |
|
218 | 218 | self.assertTrue(False, "datapack.add should throw") |
|
219 | 219 | except RuntimeError: |
|
220 | 220 | pass |
|
221 | 221 | |
|
222 | 222 | def testBadVersionThrows(self): |
|
223 | 223 | pack = self.createPack() |
|
224 | 224 | path = pack.path + b'.datapack' |
|
225 | 225 | with open(path, 'rb') as f: |
|
226 | 226 | raw = f.read() |
|
227 | 227 | raw = struct.pack('!B', 255) + raw[1:] |
|
228 | 228 | os.chmod(path, os.stat(path).st_mode | stat.S_IWRITE) |
|
229 | 229 | with open(path, 'wb+') as f: |
|
230 | 230 | f.write(raw) |
|
231 | 231 | |
|
232 | 232 | try: |
|
233 | 233 | pack = self.datapackreader(pack.path) |
|
234 | 234 | self.assertTrue(False, "bad version number should have thrown") |
|
235 | 235 | except RuntimeError: |
|
236 | 236 | pass |
|
237 | 237 | |
|
238 | 238 | def testMissingDeltabase(self): |
|
239 | 239 | fakenode = self.getFakeHash() |
|
240 | 240 | revisions = [(b"filename", fakenode, self.getFakeHash(), b"content")] |
|
241 | 241 | pack = self.createPack(revisions) |
|
242 | 242 | chain = pack.getdeltachain(b"filename", fakenode) |
|
243 | 243 | self.assertEqual(len(chain), 1) |
|
244 | 244 | |
|
245 | 245 | def testLargePack(self): |
|
246 | 246 | """Test creating and reading from a large pack with over X entries. |
|
247 | 247 | This causes it to use a 2^16 fanout table instead.""" |
|
248 | 248 | revisions = [] |
|
249 | 249 | blobs = {} |
|
250 | 250 | total = basepack.SMALLFANOUTCUTOFF + 1 |
|
251 | 251 | for i in pycompat.xrange(total): |
|
252 | 252 | filename = b"filename-%d" % i |
|
253 | 253 | content = filename |
|
254 | 254 | node = self.getHash(content) |
|
255 | 255 | blobs[(filename, node)] = content |
|
256 | 256 | revisions.append((filename, node, nullid, content)) |
|
257 | 257 | |
|
258 | 258 | pack = self.createPack(revisions) |
|
259 | 259 | if self.paramsavailable: |
|
260 | 260 | self.assertEqual(pack.params.fanoutprefix, |
|
261 | 261 | basepack.LARGEFANOUTPREFIX) |
|
262 | 262 | |
|
263 | 263 | for (filename, node), content in blobs.items(): |
|
264 | 264 | actualcontent = pack.getdeltachain(filename, node)[0][4] |
|
265 | 265 | self.assertEqual(actualcontent, content) |
|
266 | 266 | |
|
267 | 267 | def testPacksCache(self): |
|
268 | 268 | """Test that we remember the most recent packs while fetching the delta |
|
269 | 269 | chain.""" |
|
270 | 270 | |
|
271 | 271 | packdir = self.makeTempDir() |
|
272 | 272 | deltachains = [] |
|
273 | 273 | |
|
274 | 274 | numpacks = 10 |
|
275 | 275 | revisionsperpack = 100 |
|
276 | 276 | |
|
277 | 277 | for i in range(numpacks): |
|
278 | 278 | chain = [] |
|
279 | 279 | revision = (b'%d' % i, self.getFakeHash(), nullid, b"content") |
|
280 | 280 | |
|
281 | 281 | for _ in range(revisionsperpack): |
|
282 | 282 | chain.append(revision) |
|
283 | 283 | revision = ( |
|
284 | 284 | b'%d' % i, |
|
285 | 285 | self.getFakeHash(), |
|
286 | 286 | revision[1], |
|
287 | 287 | self.getFakeHash() |
|
288 | 288 | ) |
|
289 | 289 | |
|
290 | 290 | self.createPack(chain, packdir) |
|
291 | 291 | deltachains.append(chain) |
|
292 | 292 | |
|
293 | 293 | class testdatapackstore(datapack.datapackstore): |
|
294 | 294 | # Ensures that we are not keeping everything in the cache. |
|
295 | DEFAULTCACHESIZE = numpacks / 2 | |
|
295 | DEFAULTCACHESIZE = int(numpacks / 2) | |
|
296 | 296 | |
|
297 | 297 | store = testdatapackstore(uimod.ui(), packdir) |
|
298 | 298 | |
|
299 | 299 | random.shuffle(deltachains) |
|
300 | 300 | for randomchain in deltachains: |
|
301 | 301 | revision = random.choice(randomchain) |
|
302 | 302 | chain = store.getdeltachain(revision[0], revision[1]) |
|
303 | 303 | |
|
304 | 304 | mostrecentpack = next(iter(store.packs), None) |
|
305 | 305 | self.assertEqual( |
|
306 | 306 | mostrecentpack.getdeltachain(revision[0], revision[1]), |
|
307 | 307 | chain |
|
308 | 308 | ) |
|
309 | 309 | |
|
310 | 310 | self.assertEqual(randomchain.index(revision) + 1, len(chain)) |
|
311 | 311 | |
|
312 | 312 | # perf test off by default since it's slow |
|
313 | 313 | def _testIndexPerf(self): |
|
314 | 314 | random.seed(0) |
|
315 | 315 | print("Multi-get perf test") |
|
316 | 316 | packsizes = [ |
|
317 | 317 | 100, |
|
318 | 318 | 10000, |
|
319 | 319 | 100000, |
|
320 | 320 | 500000, |
|
321 | 321 | 1000000, |
|
322 | 322 | 3000000, |
|
323 | 323 | ] |
|
324 | 324 | lookupsizes = [ |
|
325 | 325 | 10, |
|
326 | 326 | 100, |
|
327 | 327 | 1000, |
|
328 | 328 | 10000, |
|
329 | 329 | 100000, |
|
330 | 330 | 1000000, |
|
331 | 331 | ] |
|
332 | 332 | for packsize in packsizes: |
|
333 | 333 | revisions = [] |
|
334 | 334 | for i in pycompat.xrange(packsize): |
|
335 | 335 | filename = b"filename-%d" % i |
|
336 | 336 | content = b"content-%d" % i |
|
337 | 337 | node = self.getHash(content) |
|
338 | 338 | revisions.append((filename, node, nullid, content)) |
|
339 | 339 | |
|
340 | 340 | path = self.createPack(revisions).path |
|
341 | 341 | |
|
342 | 342 | # Perf of large multi-get |
|
343 | 343 | import gc |
|
344 | 344 | gc.disable() |
|
345 | 345 | pack = self.datapackreader(path) |
|
346 | 346 | for lookupsize in lookupsizes: |
|
347 | 347 | if lookupsize > packsize: |
|
348 | 348 | continue |
|
349 | 349 | random.shuffle(revisions) |
|
350 | 350 | findnodes = [(rev[0], rev[1]) for rev in revisions] |
|
351 | 351 | |
|
352 | 352 | start = time.time() |
|
353 | 353 | pack.getmissing(findnodes[:lookupsize]) |
|
354 | 354 | elapsed = time.time() - start |
|
355 | 355 | print ("%s pack %d lookups = %0.04f" % |
|
356 | 356 | (('%d' % packsize).rjust(7), |
|
357 | 357 | ('%d' % lookupsize).rjust(7), |
|
358 | 358 | elapsed)) |
|
359 | 359 | |
|
360 | 360 | print("") |
|
361 | 361 | gc.enable() |
|
362 | 362 | |
|
363 | 363 | # The perf test is meant to produce output, so we always fail the test |
|
364 | 364 | # so the user sees the output. |
|
365 | 365 | raise RuntimeError("perf test always fails") |
|
366 | 366 | |
|
367 | 367 | class datapacktests(datapacktestsbase, unittest.TestCase): |
|
368 | 368 | def __init__(self, *args, **kwargs): |
|
369 | 369 | datapacktestsbase.__init__(self, datapack.datapack, True) |
|
370 | 370 | unittest.TestCase.__init__(self, *args, **kwargs) |
|
371 | 371 | |
|
372 | 372 | # TODO: |
|
373 | 373 | # datapack store: |
|
374 | 374 | # - getmissing |
|
375 | 375 | # - GC two packs into one |
|
376 | 376 | |
|
377 | 377 | if __name__ == '__main__': |
|
378 | 378 | if pycompat.iswindows: |
|
379 | 379 | sys.exit(80) # Skip on Windows |
|
380 | 380 | silenttestrunner.main(__name__) |
General Comments 0
You need to be logged in to leave comments.
Login now