/usr/lib/python2.7/dist-packages/cvs2svn_lib/checkout_internal.py is in cvs2svn 2.4.0-2.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 | # (Be in -*- python -*- mode.)
#
# ====================================================================
# Copyright (c) 2007-2009 CollabNet. All rights reserved.
#
# This software is licensed as described in the file COPYING, which
# you should have received as part of this distribution. The terms
# are also available at http://subversion.tigris.org/license-1.html.
# If newer versions of this license are posted there, you may use a
# newer version instead, at your option.
#
# This software consists of voluntary contributions made by many
# individuals. For exact contribution history, see the revision
# history and logs, available at http://cvs2svn.tigris.org/.
# ====================================================================
"""This module contains classes that implement the --use-internal-co option.
The idea is to patch up the revisions' contents incrementally, thus
avoiding the huge number of process spawns and the O(n^2) overhead of
using 'co' and 'cvs'.
InternalRevisionCollector saves the RCS deltas and RCS revision trees
to databases. Notably, deltas from the trunk need to be reversed, as
CVS stores them so they apply from HEAD backwards.
InternalRevisionReader produces the revisions' contents on demand. To
generate the text for a typical revision, we need the revision's delta
text plus the fulltext of the previous revision. Therefore, we
maintain a checkout database containing a copy of the fulltext of any
revision for which subsequent revisions still need to be retrieved.
It is crucial to remove text from this database as soon as it is no
longer needed, to prevent it from growing enormous.
There are two reasons that the text from a revision can be needed: (1)
because the revision itself still needs to be output to a dumpfile;
(2) because another revision needs it as the base of its delta. We
maintain a reference count for each revision, which includes *both*
possibilities. The first time a revision's text is needed, it is
generated by applying the revision's deltatext to the previous
revision's fulltext, and the resulting fulltext is stored in the
checkout database. Each time a revision's fulltext is retrieved, its
reference count is decremented. When the reference count goes to
zero, then the fulltext is deleted from the checkout database.
The administrative data for managing this consists of one TextRecord
entry for each revision. Each TextRecord has an id, which is the same
id as used for the corresponding CVSRevision instance. It also
maintains a count of the times it is expected to be retrieved.
TextRecords come in several varieties:
FullTextRecord -- Used for revisions whose fulltext is derived
directly from the RCS file by the InternalRevisionCollector (i.e.,
typically revision 1.1 of each file).
DeltaTextRecord -- Used for revisions that are defined via a delta
relative to some other TextRecord. These records record the id of
the TextRecord that holds the base text against which the delta is
defined. When the text for a DeltaTextRecord is retrieved, the
DeltaTextRecord instance is deleted and a CheckedOutTextRecord
instance is created to take its place.
CheckedOutTextRecord -- Used during OutputPass for a revision that
started out as a DeltaTextRecord, but has already been retrieved
(and therefore its fulltext is stored in the checkout database).
While a file is being processed during FilterSymbolsPass, the fulltext
and deltas are stored to the delta database, and TextRecord instances
are created to keep track of things. The reference counts are all
initialized: each record referred to by a delta has its refcount
incremented, and each record that corresponds to a non-delete
CVSRevision is incremented. After that, any records with refcount==0
are removed. When one record is removed, that can cause another
record's reference count to go to zero and be removed too,
recursively. When a TextRecord is deleted at this stage, its
deltatext is also deleted from the delta database."""
from cvs2svn_lib import config
from cvs2svn_lib.common import DB_OPEN_NEW
from cvs2svn_lib.common import DB_OPEN_READ
from cvs2svn_lib.common import warning_prefix
from cvs2svn_lib.common import FatalError
from cvs2svn_lib.common import InternalError
from cvs2svn_lib.common import canonicalize_eol
from cvs2svn_lib.common import is_trunk_revision
from cvs2svn_lib.context import Ctx
from cvs2svn_lib.log import logger
from cvs2svn_lib.artifact_manager import artifact_manager
from cvs2svn_lib.cvs_item import CVSRevisionModification
from cvs2svn_lib.indexed_database import IndexedDatabase
from cvs2svn_lib.rcs_stream import RCSStream
from cvs2svn_lib.rcs_stream import MalformedDeltaException
from cvs2svn_lib.keyword_expander import expand_keywords
from cvs2svn_lib.keyword_expander import collapse_keywords
from cvs2svn_lib.revision_manager import RevisionCollector
from cvs2svn_lib.revision_manager import RevisionReader
from cvs2svn_lib.serializer import MarshalSerializer
from cvs2svn_lib.serializer import CompressingSerializer
from cvs2svn_lib.serializer import PrimedPickleSerializer
from cvs2svn_lib.apple_single_filter import get_maybe_apple_single
from cvs2svn_lib.rcsparser import Sink
from cvs2svn_lib.rcsparser import parse
class TextRecord(object):
"""Bookkeeping data for the text of a single CVSRevision."""
__slots__ = ['id', 'refcount']
def __init__(self, id):
# The cvs_rev_id of the revision whose text this is.
self.id = id
# The number of times that the text of this revision will be
# retrieved.
self.refcount = 0
def __getstate__(self):
return (self.id, self.refcount,)
def __setstate__(self, state):
(self.id, self.refcount,) = state
def increment_dependency_refcounts(self, text_record_db):
"""Increment the refcounts of any records that this one depends on."""
pass
def decrement_refcount(self, text_record_db):
"""Decrement the number of times our text still has to be checked out.
If the reference count goes to zero, call discard()."""
self.refcount -= 1
if self.refcount == 0:
text_record_db.discard(self.id)
def checkout(self, text_record_db):
"""Workhorse of the checkout process.
Return the text for this revision, decrement our reference count,
and update the databases depending on whether there will be future
checkouts."""
raise NotImplementedError()
def free(self, text_record_db):
"""This instance will never again be checked out; free it.
Also free any associated resources and decrement the refcounts of
any other TextRecords that this one depends on."""
raise NotImplementedError()
class FullTextRecord(TextRecord):
"""A record whose revision's fulltext is stored in the delta_db.
These records are used for revisions whose fulltext was determined
by the InternalRevisionCollector during FilterSymbolsPass. The
fulltext for such a revision is is stored in the delta_db as a
single string."""
__slots__ = []
def __getstate__(self):
return (self.id, self.refcount,)
def __setstate__(self, state):
(self.id, self.refcount,) = state
def checkout(self, text_record_db):
text = text_record_db.delta_db[self.id]
self.decrement_refcount(text_record_db)
return text
def free(self, text_record_db):
del text_record_db.delta_db[self.id]
def __str__(self):
return 'FullTextRecord(%x, %d)' % (self.id, self.refcount,)
class DeltaTextRecord(TextRecord):
"""A record whose revision's delta is stored as an RCS delta.
The text of this revision must be derived by applying an RCS delta
to the text of the predecessor revision. The RCS delta is stored
in the delta_db."""
__slots__ = ['pred_id']
def __init__(self, id, pred_id):
TextRecord.__init__(self, id)
# The cvs_rev_id of the revision relative to which this delta is
# defined.
self.pred_id = pred_id
def __getstate__(self):
return (self.id, self.refcount, self.pred_id,)
def __setstate__(self, state):
(self.id, self.refcount, self.pred_id,) = state
def increment_dependency_refcounts(self, text_record_db):
text_record_db[self.pred_id].refcount += 1
def checkout(self, text_record_db):
base_text = text_record_db[self.pred_id].checkout(text_record_db)
rcs_stream = RCSStream(base_text)
delta_text = text_record_db.delta_db[self.id]
rcs_stream.apply_diff(delta_text)
text = rcs_stream.get_text()
del rcs_stream
self.refcount -= 1
if self.refcount == 0:
# This text will never be needed again; just delete ourselves
# without ever having stored the fulltext to the checkout
# database:
del text_record_db[self.id]
else:
# Store a new CheckedOutTextRecord in place of ourselves:
text_record_db.checkout_db['%x' % self.id] = text
new_text_record = CheckedOutTextRecord(self.id)
new_text_record.refcount = self.refcount
text_record_db.replace(new_text_record)
return text
def free(self, text_record_db):
del text_record_db.delta_db[self.id]
text_record_db[self.pred_id].decrement_refcount(text_record_db)
def __str__(self):
return 'DeltaTextRecord(%x -> %x, %d)' % (
self.pred_id, self.id, self.refcount,
)
class CheckedOutTextRecord(TextRecord):
"""A record whose revision's fulltext is stored in the text_record_db.
These records are used for revisions whose fulltext has been
computed already during OutputPass. The fulltext for such a
revision is stored in the text_record_db as a single string."""
__slots__ = []
def __getstate__(self):
return (self.id, self.refcount,)
def __setstate__(self, state):
(self.id, self.refcount,) = state
def checkout(self, text_record_db):
text = text_record_db.checkout_db['%x' % self.id]
self.decrement_refcount(text_record_db)
return text
def free(self, text_record_db):
del text_record_db.checkout_db['%x' % self.id]
def __str__(self):
return 'CheckedOutTextRecord(%x, %d)' % (self.id, self.refcount,)
class NullDatabase(object):
"""A do-nothing database that can be used with TextRecordDatabase.
Use this when you don't actually want to allow anything to be
deleted."""
def __delitem__(self, id):
pass
class TextRecordDatabase:
"""Holds the TextRecord instances that are currently live.
During FilterSymbolsPass, files are processed one by one and a new
TextRecordDatabase instance is used for each file. During
OutputPass, a single TextRecordDatabase instance is used for the
duration of OutputPass; individual records are added and removed
when they are active."""
def __init__(self, delta_db, checkout_db):
# A map { cvs_rev_id -> TextRecord }.
self.text_records = {}
# A database-like object using cvs_rev_ids as keys and containing
# fulltext/deltatext strings as values. Its __getitem__() method
# is used to retrieve deltas when they are needed, and its
# __delitem__() method is used to delete deltas when they can be
# freed. The modifiability of the delta database varies from pass
# to pass, so the object stored here varies as well:
#
# FilterSymbolsPass: a NullDatabase. The delta database cannot be
# modified during this pass, and we have no need to retrieve
# deltas, so we just use a dummy object here.
#
# OutputPass: a disabled IndexedDatabase. During this pass we
# need to retrieve deltas, but we are not allowed to modify
# the delta database. So we use an IndexedDatabase whose
# __del__() method has been disabled to do nothing.
self.delta_db = delta_db
# A database-like object using cvs_rev_ids as keys and containing
# fulltext strings as values. This database is only set during
# OutputPass.
self.checkout_db = checkout_db
# If this is set to a list, then the list holds the ids of
# text_records that have to be deleted; when discard() is called,
# it adds the requested id to the list but does not delete it. If
# this member is set to None, then text_records are deleted
# immediately when discard() is called.
self.deferred_deletes = None
def __getstate__(self):
return (self.text_records.values(),)
def __setstate__(self, state):
(text_records,) = state
self.text_records = {}
for text_record in text_records:
self.add(text_record)
self.delta_db = NullDatabase()
self.checkout_db = NullDatabase()
self.deferred_deletes = None
def add(self, text_record):
"""Add TEXT_RECORD to our database.
There must not already be a record with the same id."""
assert not self.text_records.has_key(text_record.id)
self.text_records[text_record.id] = text_record
def __getitem__(self, id):
return self.text_records[id]
def __delitem__(self, id):
"""Free the record with the specified ID."""
del self.text_records[id]
def replace(self, text_record):
"""Store TEXT_RECORD in place of the existing record with the same id.
Do not do anything with the old record."""
assert self.text_records.has_key(text_record.id)
self.text_records[text_record.id] = text_record
def discard(self, *ids):
"""The text records with IDS are no longer needed; discard them.
This involves calling their free() methods and also removing them
from SELF.
If SELF.deferred_deletes is not None, then the ids to be deleted
are added to the list instead of deleted immediately. This
mechanism is to prevent a stack overflow from the avalanche of
deletes that can result from deleting a long chain of revisions."""
if self.deferred_deletes is None:
# This is an outer-level delete.
self.deferred_deletes = list(ids)
while self.deferred_deletes:
id = self.deferred_deletes.pop()
text_record = self[id]
if text_record.refcount != 0:
raise InternalError(
'TextRecordDatabase.discard(%s) called with refcount = %d'
% (text_record, text_record.refcount,)
)
# This call might cause other text_record ids to be added to
# self.deferred_deletes:
text_record.free(self)
del self[id]
self.deferred_deletes = None
else:
self.deferred_deletes.extend(ids)
def itervalues(self):
return self.text_records.itervalues()
def recompute_refcounts(self, cvs_file_items):
"""Recompute the refcounts of the contained TextRecords.
Use CVS_FILE_ITEMS to determine which records will be needed by
cvs2svn."""
# First clear all of the refcounts:
for text_record in self.itervalues():
text_record.refcount = 0
# Now increment the reference count of records that are needed as
# the source of another record's deltas:
for text_record in self.itervalues():
text_record.increment_dependency_refcounts(self.text_records)
# Now increment the reference count of records that will be needed
# by cvs2svn:
for lod_items in cvs_file_items.iter_lods():
for cvs_rev in lod_items.cvs_revisions:
if isinstance(cvs_rev, CVSRevisionModification):
self[cvs_rev.id].refcount += 1
def free_unused(self):
"""Free any TextRecords whose reference counts are zero."""
# The deletion of some of these text records might cause others to
# be unused, in which case they will be deleted automatically.
# But since the initially-unused records are not referred to by
# any others, we don't have to be afraid that they will be deleted
# before we get to them. But it *is* crucial that we create the
# whole unused list before starting the loop.
unused = [
text_record.id
for text_record in self.itervalues()
if text_record.refcount == 0
]
self.discard(*unused)
def log_leftovers(self):
"""If any TextRecords still exist, log them."""
if self.text_records:
logger.warn(
"%s: internal problem: leftover revisions in the checkout cache:"
% warning_prefix)
for text_record in self.itervalues():
logger.warn(' %s' % (text_record,))
def __repr__(self):
"""Debugging output of the current contents of the TextRecordDatabase."""
retval = ['TextRecordDatabase:']
for text_record in self.itervalues():
retval.append(' %s' % (text_record,))
return '\n'.join(retval)
class _Sink(Sink):
def __init__(self, revision_collector, cvs_file_items):
self.revision_collector = revision_collector
self.cvs_file_items = cvs_file_items
# A map {rev : base_rev} indicating that the text for rev is
# stored in CVS as a delta relative to base_rev.
self.base_revisions = {}
# The revision that is stored with its fulltext in CVS (usually
# the oldest revision on trunk):
self.head_revision = None
# The first logical revision on trunk (usually '1.1'):
self.revision_1_1 = None
# Keep track of the revisions whose revision info has been seen so
# far (to avoid repeated revision info blocks):
self.revisions_seen = set()
def set_head_revision(self, revision):
self.head_revision = revision
def define_revision(
self, revision, timestamp, author, state, branches, next
):
if next:
self.base_revisions[next] = revision
else:
if is_trunk_revision(revision):
self.revision_1_1 = revision
for branch in branches:
self.base_revisions[branch] = revision
def set_revision_info(self, revision, log, text):
if revision in self.revisions_seen:
# One common form of CVS repository corruption is that the
# Deltatext block for revision 1.1 appears twice. CollectData
# has already warned about this problem; here we can just ignore
# it.
return
else:
self.revisions_seen.add(revision)
cvs_rev_id = self.cvs_file_items.original_ids[revision]
if is_trunk_revision(revision):
# On trunk, revisions are encountered in reverse order (1.<N>
# ... 1.1) and deltas are inverted. The first text that we see
# is the fulltext for the HEAD revision. After that, the text
# corresponding to revision 1.N is the delta (1.<N+1> ->
# 1.<N>)). We have to invert the deltas here so that we can
# read the revisions out in dependency order; that is, for
# revision 1.1 we want the fulltext, and for revision 1.<N> we
# want the delta (1.<N-1> -> 1.<N>). This means that we can't
# compute the delta for a revision until we see its logical
# parent. When we finally see revision 1.1 (which is recognized
# because it doesn't have a parent), we can record the diff (1.1
# -> 1.2) for revision 1.2, and also the fulltext for 1.1.
if revision == self.head_revision:
# This is HEAD, as fulltext. Initialize the RCSStream so
# that we can compute deltas backwards in time.
self._rcs_stream = RCSStream(text)
self._rcs_stream_revision = revision
else:
# Any other trunk revision is a backward delta. Apply the
# delta to the RCSStream to mutate it to the contents of this
# revision, and also to get the reverse delta, which we store
# as the forward delta of our child revision.
try:
text = self._rcs_stream.invert_diff(text)
except MalformedDeltaException, e:
logger.error(
'Malformed RCS delta in %s, revision %s: %s'
% (self.cvs_file_items.cvs_file.rcs_path, revision, e)
)
raise RuntimeError()
text_record = DeltaTextRecord(
self.cvs_file_items.original_ids[self._rcs_stream_revision],
cvs_rev_id
)
self.revision_collector._writeout(text_record, text)
self._rcs_stream_revision = revision
if revision == self.revision_1_1:
# This is revision 1.1. Write its fulltext:
text_record = FullTextRecord(cvs_rev_id)
self.revision_collector._writeout(
text_record, self._rcs_stream.get_text()
)
# There will be no more trunk revisions delivered, so free the
# RCSStream.
del self._rcs_stream
del self._rcs_stream_revision
else:
# On branches, revisions are encountered in logical order
# (<BRANCH>.1 ... <BRANCH>.<N>) and the text corresponding to
# revision <BRANCH>.<N> is the forward delta (<BRANCH>.<N-1> ->
# <BRANCH>.<N>). That's what we need, so just store it.
# FIXME: It would be nice to avoid writing out branch deltas
# when --trunk-only. (They will be deleted when finish_file()
# is called, but if the delta db is in an IndexedDatabase the
# deletions won't actually recover any disk space.)
text_record = DeltaTextRecord(
cvs_rev_id,
self.cvs_file_items.original_ids[self.base_revisions[revision]]
)
self.revision_collector._writeout(text_record, text)
return None
class InternalRevisionCollector(RevisionCollector):
"""The RevisionCollector used by InternalRevisionReader."""
def __init__(self, compress):
RevisionCollector.__init__(self)
self._compress = compress
def register_artifacts(self, which_pass):
artifact_manager.register_temp_file(
config.RCS_DELTAS_INDEX_TABLE, which_pass
)
artifact_manager.register_temp_file(config.RCS_DELTAS_STORE, which_pass)
artifact_manager.register_temp_file(
config.RCS_TREES_INDEX_TABLE, which_pass
)
artifact_manager.register_temp_file(config.RCS_TREES_STORE, which_pass)
def start(self):
serializer = MarshalSerializer()
if self._compress:
serializer = CompressingSerializer(serializer)
self._delta_db = IndexedDatabase(
artifact_manager.get_temp_file(config.RCS_DELTAS_STORE),
artifact_manager.get_temp_file(config.RCS_DELTAS_INDEX_TABLE),
DB_OPEN_NEW, serializer,
)
primer = (FullTextRecord, DeltaTextRecord)
self._rcs_trees = IndexedDatabase(
artifact_manager.get_temp_file(config.RCS_TREES_STORE),
artifact_manager.get_temp_file(config.RCS_TREES_INDEX_TABLE),
DB_OPEN_NEW, PrimedPickleSerializer(primer),
)
def _writeout(self, text_record, text):
self.text_record_db.add(text_record)
self._delta_db[text_record.id] = text
def process_file(self, cvs_file_items):
"""Read revision information for the file described by CVS_FILE_ITEMS.
Compute the text record refcounts, discard any records that are
unneeded, and store the text records for the file to the
_rcs_trees database."""
# A map from cvs_rev_id to TextRecord instance:
self.text_record_db = TextRecordDatabase(self._delta_db, NullDatabase())
parse(
open(cvs_file_items.cvs_file.rcs_path, 'rb'),
_Sink(self, cvs_file_items),
)
self.text_record_db.recompute_refcounts(cvs_file_items)
self.text_record_db.free_unused()
self._rcs_trees[cvs_file_items.cvs_file.id] = self.text_record_db
del self.text_record_db
def finish(self):
self._delta_db.close()
self._rcs_trees.close()
class InternalRevisionReader(RevisionReader):
"""A RevisionReader that reads the contents from an own delta store."""
def __init__(self, compress):
# Only import Database if an InternalRevisionReader is really
# instantiated, because the import fails if a decent dbm is not
# installed.
from cvs2svn_lib.database import Database
self._Database = Database
self._compress = compress
def register_artifacts(self, which_pass):
artifact_manager.register_temp_file(config.CVS_CHECKOUT_DB, which_pass)
artifact_manager.register_temp_file_needed(
config.RCS_DELTAS_STORE, which_pass
)
artifact_manager.register_temp_file_needed(
config.RCS_DELTAS_INDEX_TABLE, which_pass
)
artifact_manager.register_temp_file_needed(
config.RCS_TREES_STORE, which_pass
)
artifact_manager.register_temp_file_needed(
config.RCS_TREES_INDEX_TABLE, which_pass
)
def start(self):
self._delta_db = IndexedDatabase(
artifact_manager.get_temp_file(config.RCS_DELTAS_STORE),
artifact_manager.get_temp_file(config.RCS_DELTAS_INDEX_TABLE),
DB_OPEN_READ,
)
self._delta_db.__delitem__ = lambda id: None
self._tree_db = IndexedDatabase(
artifact_manager.get_temp_file(config.RCS_TREES_STORE),
artifact_manager.get_temp_file(config.RCS_TREES_INDEX_TABLE),
DB_OPEN_READ,
)
serializer = MarshalSerializer()
if self._compress:
serializer = CompressingSerializer(serializer)
self._co_db = self._Database(
artifact_manager.get_temp_file(config.CVS_CHECKOUT_DB),
DB_OPEN_NEW, serializer,
)
# The set of CVSFile instances whose TextRecords have already been
# read:
self._loaded_files = set()
# A map { CVSFILE : _FileTree } for files that currently have live
# revisions:
self._text_record_db = TextRecordDatabase(self._delta_db, self._co_db)
def _get_text_record(self, cvs_rev):
"""Return the TextRecord instance for CVS_REV.
If the TextRecords for CVS_REV.cvs_file haven't been loaded yet,
do so now."""
if cvs_rev.cvs_file not in self._loaded_files:
for text_record in self._tree_db[cvs_rev.cvs_file.id].itervalues():
self._text_record_db.add(text_record)
self._loaded_files.add(cvs_rev.cvs_file)
return self._text_record_db[cvs_rev.id]
def get_content(self, cvs_rev):
"""Check out the text for revision C_REV from the repository.
Return the text. If CVS_REV has a property _keyword_handling, use
it to determine how to handle RCS keywords in the output:
'collapsed' -- collapse keywords
'expanded' -- expand keywords
'untouched' -- output keywords in the form they are found in
the RCS file
Note that $Log$ never actually generates a log (which makes test
'requires_cvs()' fail).
Revisions may be requested in any order, but if they are not
requested in dependency order the checkout database will become
very large. Revisions may be skipped. Each revision may be
requested only once."""
try:
text = self._get_text_record(cvs_rev).checkout(self._text_record_db)
except MalformedDeltaException, (msg):
raise FatalError(
'Malformed RCS delta in %s, revision %s: %s'
% (cvs_rev.cvs_file.rcs_path, cvs_rev.rev, msg)
)
keyword_handling = cvs_rev.get_property('_keyword_handling')
if keyword_handling == 'untouched':
# Leave keywords in the form that they were checked in.
pass
elif keyword_handling == 'collapsed':
text = collapse_keywords(text)
elif keyword_handling == 'expanded':
text = expand_keywords(text, cvs_rev)
else:
raise FatalError(
'Undefined _keyword_handling property (%r) for %s'
% (keyword_handling, cvs_rev,)
)
if Ctx().decode_apple_single:
# Insert a filter to decode any files that are in AppleSingle
# format:
text = get_maybe_apple_single(text)
eol_fix = cvs_rev.get_property('_eol_fix')
if eol_fix:
text = canonicalize_eol(text, eol_fix)
return text
def finish(self):
self._text_record_db.log_leftovers()
del self._text_record_db
self._delta_db.close()
self._tree_db.close()
self._co_db.close()
|