← Back to branch summary

~bzr-pqm/bzr/bzr.dev

« back to all changes in this revision

Viewing changes to bzrlib/repofmt/pack_repo.py

  • Committer: Jonathan Lange
  • Date: 2009-06-26 08:46:52 UTC
  • mto: (4484.1.1 bring-1.16.1-back)
  • mto: This revision was merged to the branch mainline in revision 4485.
  • Revision ID: jml@canonical.com-20090626084652-x7wn8yimd3fj0j0y
Tweak NEWS slightly based on mbp's feedback.

Show diffs side-by-side

added added

removed removed

Lines of Context:
bzrlib/repofmt/pack_repo.py
73
73
    MetaDirRepositoryFormat,
74
74
    RepositoryFormat,
75
75
    RootCommitBuilder,
76
 
    StreamSource,
77
76
    )
78
77
import bzrlib.revision as _mod_revision
79
78
from bzrlib.trace import (
313
312
 
314
313
    def finish(self):
315
314
        self._check_references()
 
315
        new_name = '../packs/' + self.file_name()
 
316
        self.upload_transport.rename(self.file_name(), new_name)
316
317
        index_types = ['revision', 'inventory', 'text', 'signature']
317
318
        if self.chk_index is not None:
318
319
            index_types.append('chk')
321
322
            new_name = '../indices/' + old_name
322
323
            self.upload_transport.rename(old_name, new_name)
323
324
            self._replace_index_with_readonly(index_type)
324
 
        new_name = '../packs/' + self.file_name()
325
 
        self.upload_transport.rename(self.file_name(), new_name)
326
325
        self._state = 'finished'
327
326
 
328
327
    def _get_external_refs(self, index):
2266
2265
            pb.finished()
2267
2266
        return result
2268
2267
 
2269
 
    def _get_source(self, to_format):
2270
 
        if to_format.network_name() == self._format.network_name():
2271
 
            return KnitPackStreamSource(self, to_format)
2272
 
        return super(KnitPackRepository, self)._get_source(to_format)
2273
 
 
2274
2268
    def _make_parents_provider(self):
2275
2269
        return graph.CachingParentsProvider(self)
2276
2270
 
2390
2384
                repo.unlock()
2391
2385
 
2392
2386
 
2393
 
class KnitPackStreamSource(StreamSource):
2394
 
    """A StreamSource used to transfer data between same-format KnitPack repos.
2395
 
 
2396
 
    This source assumes:
2397
 
        1) Same serialization format for all objects
2398
 
        2) Same root information
2399
 
        3) XML format inventories
2400
 
        4) Atomic inserts (so we can stream inventory texts before text
2401
 
           content)
2402
 
        5) No chk_bytes
2403
 
    """
2404
 
 
2405
 
    def __init__(self, from_repository, to_format):
2406
 
        super(KnitPackStreamSource, self).__init__(from_repository, to_format)
2407
 
        self._text_keys = None
2408
 
        self._text_fetch_order = 'unordered'
2409
 
 
2410
 
    def _get_filtered_inv_stream(self, revision_ids):
2411
 
        from_repo = self.from_repository
2412
 
        parent_ids = from_repo._find_parent_ids_of_revisions(revision_ids)
2413
 
        parent_keys = [(p,) for p in parent_ids]
2414
 
        find_text_keys = from_repo._find_text_key_references_from_xml_inventory_lines
2415
 
        parent_text_keys = set(find_text_keys(
2416
 
            from_repo._inventory_xml_lines_for_keys(parent_keys)))
2417
 
        content_text_keys = set()
2418
 
        knit = KnitVersionedFiles(None, None)
2419
 
        factory = KnitPlainFactory()
2420
 
        def find_text_keys_from_content(record):
2421
 
            if record.storage_kind not in ('knit-delta-gz', 'knit-ft-gz'):
2422
 
                raise ValueError("Unknown content storage kind for"
2423
 
                    " inventory text: %s" % (record.storage_kind,))
2424
 
            # It's a knit record, it has a _raw_record field (even if it was
2425
 
            # reconstituted from a network stream).
2426
 
            raw_data = record._raw_record
2427
 
            # read the entire thing
2428
 
            revision_id = record.key[-1]
2429
 
            content, _ = knit._parse_record(revision_id, raw_data)
2430
 
            if record.storage_kind == 'knit-delta-gz':
2431
 
                line_iterator = factory.get_linedelta_content(content)
2432
 
            elif record.storage_kind == 'knit-ft-gz':
2433
 
                line_iterator = factory.get_fulltext_content(content)
2434
 
            content_text_keys.update(find_text_keys(
2435
 
                [(line, revision_id) for line in line_iterator]))
2436
 
        revision_keys = [(r,) for r in revision_ids]
2437
 
        def _filtered_inv_stream():
2438
 
            source_vf = from_repo.inventories
2439
 
            stream = source_vf.get_record_stream(revision_keys,
2440
 
                                                 'unordered', False)
2441
 
            for record in stream:
2442
 
                if record.storage_kind == 'absent':
2443
 
                    raise errors.NoSuchRevision(from_repo, record.key)
2444
 
                find_text_keys_from_content(record)
2445
 
                yield record
2446
 
            self._text_keys = content_text_keys - parent_text_keys
2447
 
        return ('inventories', _filtered_inv_stream())
2448
 
 
2449
 
    def _get_text_stream(self):
2450
 
        # Note: We know we don't have to handle adding root keys, because both
2451
 
        # the source and target are the identical network name.
2452
 
        text_stream = self.from_repository.texts.get_record_stream(
2453
 
                        self._text_keys, self._text_fetch_order, False)
2454
 
        return ('texts', text_stream)
2455
 
 
2456
 
    def get_stream(self, search):
2457
 
        revision_ids = search.get_keys()
2458
 
        for stream_info in self._fetch_revision_texts(revision_ids):
2459
 
            yield stream_info
2460
 
        self._revision_keys = [(rev_id,) for rev_id in revision_ids]
2461
 
        yield self._get_filtered_inv_stream(revision_ids)
2462
 
        yield self._get_text_stream()
2463
 
 
2464
 
 
2465
 
 
2466
2387
class RepositoryFormatPack(MetaDirRepositoryFormat):
2467
2388
    """Format logic for pack structured repositories.
2468
2389