← Back to branch summary

~bzr-pqm/bzr/bzr.dev

« back to all changes in this revision

Viewing changes to bzrlib/groupcompress.py

  • Committer: John Arbash Meinel
  • Date: 2009-09-01 21:41:27 UTC
  • mto: This revision was merged to the branch mainline in revision 4672.
  • Revision ID: john@arbash-meinel.com-20090901214127-gs4xwl6t5q81k0hh
Refactor the check_rebuild code a bit, so that we can potentially
not do the actual rebuilding, but just check that we want to.

Show diffs side-by-side

added added

removed removed

Lines of Context:
bzrlib/groupcompress.py
457
457
                # There are code paths that first extract as fulltext, and then
458
458
                # extract as storage_kind (smart fetch). So we don't break the
459
459
                # refcycle here, but instead in manager.get_record_stream()
460
 
                # self._manager = None
461
460
            if storage_kind == 'fulltext':
462
461
                return self._bytes
463
462
            else:
546
545
        # time (self._block._content) is a little expensive.
547
546
        self._block._ensure_content(self._last_byte)
548
547
 
549
 
    def _check_rebuild_block(self):
 
548
    def _check_rebuild_action(self):
550
549
        """Check to see if our block should be repacked."""
551
550
        total_bytes_used = 0
552
551
        last_byte_used = 0
553
552
        for factory in self._factories:
554
553
            total_bytes_used += factory._end - factory._start
555
 
            last_byte_used = max(last_byte_used, factory._end)
556
 
        # If we are using most of the bytes from the block, we have nothing
557
 
        # else to check (currently more than 1/2)
 
554
            if last_byte_used < factory._end:
 
555
                last_byte_used = factory._end
 
556
        # If we are using more than half of the bytes from the block, we have
 
557
        # nothing else to check
558
558
        if total_bytes_used * 2 >= self._block._content_length:
559
 
            return
560
 
        # Can we just strip off the trailing bytes? If we are going to be
561
 
        # transmitting more than 50% of the front of the content, go ahead
 
559
            return None, last_byte_used
 
560
        # We are using less than 50% of the content. Is the content we are
 
561
        # using at the beginning of the block? If so, we can just trim the
 
562
        # tail, rather than rebuilding from scratch.
562
563
        if total_bytes_used * 2 > last_byte_used:
563
 
            self._trim_block(last_byte_used)
564
 
            return
 
564
            return 'trim', last_byte_used
565
565
 
566
566
        # We are using a small amount of the data, and it isn't just packed
567
567
        # nicely at the front, so rebuild the content.
574
574
        #       expanding many deltas into fulltexts, as well.
575
575
        #       If we build a cheap enough 'strip', then we could try a strip,
576
576
        #       if that expands the content, we then rebuild.
577
 
        self._rebuild_block()
 
577
        return 'rebuild', last_byte_used
 
578
 
 
579
    def _check_rebuild_block(self):
 
580
        action, last_byte_used = self._check_rebuild_action()
 
581
        if action is None:
 
582
            return
 
583
        if action == 'trim':
 
584
            self._trim_block(last_byte_used)
 
585
        elif action == 'rebuild':
 
586
            self._rebuild_block()
 
587
        else:
 
588
            raise ValueError('unknown rebuild action: %r' % (action,))
578
589
 
579
590
    def _wire_bytes(self):
580
591
        """Return a byte stream suitable for transmitting over the wire."""
1587
1598
                if record.storage_kind == 'groupcompress-block':
1588
1599
                    # Check to see if we really want to re-use this block
1589
1600
                    insert_manager = record._manager
 
1601
                    insert_manager._check_rebuild_block()
1590
1602
                    if len(insert_manager._factories) == 1:
1591
1603
                        # This block only has a single record in it
1592
1604
                        # Mark this block to be rebuilt
1606
1618
                                           'groupcompress-block-ref'):
1607
1619
                    if insert_manager is None:
1608
1620
                        raise AssertionError('No insert_manager set')
 
1621
                    if insert_manager is not record._manager:
 
1622
                        raise AssertionError('insert_manager does not match'
 
1623
                            ' the current record, we cannot be positive'
 
1624
                            ' that the appropriate content was inserted.'
 
1625
                            )
1609
1626
                    value = "%d %d %d %d" % (block_start, block_length,
1610
1627
                                             record._start, record._end)
1611
1628
                    nodes = [(record.key, value, (record.parents,))]