1
=============================
2
Bazaar Architectural Overview
3
=============================
5
This document describes the key classes and concepts within Bazaar. It is
6
intended to be useful to people working on the Bazaar codebase, or to
7
people writing plugins.
9
If you have any questions, or if something seems to be incorrect, unclear
10
or missing, please talk to us in ``irc://irc.freenode.net/#bzr``, or write
11
to the Bazaar mailing list. To propose a correction or addition to this
12
document, send a merge request or new text to the mailing list.
14
The current version of this document is available in the file
15
``doc/developers/overview.txt`` in the source tree, and available online
16
within the developer documentation, <http://doc.bazaar-vcs.org/developers/>.
19
Essential Domain Classes
20
########################
22
The core domain objects within the bazaar model are:
32
Transports are explained below. See http://bazaar-vcs.org/Classes/
33
for an introduction to the other key classes.
38
The ``Transport`` layer handles access to local or remote directories.
39
Each Transport object acts as a logical connection to a particular
40
directory, and it allows various operations on files within it. You can
41
*clone* a transport to get a new Transport connected to a subdirectory or
44
Transports are not used for access to the working tree. At present
45
working trees are always local and they are accessed through the regular
46
Python file I/O mechanisms.
51
Transports work in terms of URLs. Take note that URLs are by definition
52
only ASCII - the decision of how to encode a Unicode string into a URL
53
must be taken at a higher level, typically in the Store. (Note that
54
Stores also escape filenames which cannot be safely stored on all
55
filesystems, but this is a different level.)
57
The main reason for this is that it's not possible to safely roundtrip a
58
URL into Unicode and then back into the same URL. The URL standard
59
gives a way to represent non-ASCII bytes in ASCII (as %-escapes), but
60
doesn't say how those bytes represent non-ASCII characters. (They're not
61
guaranteed to be UTF-8 -- that is common but doesn't happen everywhere.)
63
For example, if the user enters the URL ``http://example/%e0``, there's no
64
way to tell whether that character represents "latin small letter a with
65
grave" in iso-8859-1, or "latin small letter r with acute" in iso-8859-2,
66
or malformed UTF-8. So we can't convert the URL to Unicode reliably.
68
Equally problematic is if we're given a URL-like string containing
69
(unescaped) non-ASCII characters (such as the accented a). We can't be
70
sure how to convert that to a valid (i.e. ASCII-only) URL, because we
71
don't know what encoding the server expects for those characters.
72
(Although it is not totally reliable, we might still accept these and
73
assume that they should be put into UTF-8.)
75
A similar edge case is that the URL ``http://foo/sweet%2Fsour`` contains
76
one directory component whose name is "sweet/sour". The escaped slash is
77
not a directory separator, but if we try to convert the URL to a regular
78
Unicode path, this information will be lost.
80
This implies that Transports must natively deal with URLs. For simplicity
81
they *only* deal with URLs; conversion of other strings to URLs is done
82
elsewhere. Information that Transports return, such as from ``list_dir``,
83
is also in the form of URL components.
89
Repositories store committed history: file texts, revisions, inventories,
90
and graph relationships between them.
95
A repository can be configured to refer to a list of "fallback"
96
repositories. If a particular revision is not present in the original
97
repository, it refers the query to the fallbacks.
99
Compression deltas don't span physical repository boundaries. So the
100
first commit to a new, empty repository with fallback repositories will
101
store a full text of the inventory, and of every new file text.
103
At runtime, repository stacking is actually configured by the branch, not
104
the repository. So doing ``a_bzrdir.open_repository()``
105
gets you just the single physical repository, while
106
``a_bzrdir.open_branch().repository`` gets one configured with a stacking.
107
Therefore, to permanently change the fallback repository stored on disk,
108
you must use ``Branch.set_stacked_on_url``.
110
Changing away from an existing stacked-on URL will copy across any
111
necessary history so that the repository remains usable.
113
A repository opened from an HPSS server is never stacked on the server
114
side, because this could cause complexity or security problems with the
115
server acting as a proxy for the client. Instead, the branch on the
116
server exposes the stacked-on URL and the client can open that.