1
Case Insensitive File Systems
2
=============================
4
Bazaar must be portable across operating-systems and file-systems. While the
5
primary file-system for an operating-system might have some particular
6
characteristics, it's not necessary that *all* file-systems for that
7
operating-system will have the same characteristics.
9
For example, the FAT32 file-system is most commonly found on Windows operating
10
systems, and has the characteristics usually associated with a Windows
11
file-system. However, USB devices means FAT32 file-systems are often used
12
with Linux, so the current operating system doesn't necessarily reflect the
13
capabilities of the file-system.
15
Bazaar supports 3 kinds of file-systems, each to different degrees.
17
* Case-sensitive file-systems: This is the file-system generally used on
18
Linux - 2 files can differ only by case, and the exact case must be used
21
* Case-insensitive, case-preserving (cicp) file-systems: This is the
22
file-system generally used on Windows; FAT32 is an example of such a
23
file-system. Although existing files can be opened using any case, the
24
exact case used to create the file is preserved and available for programs
25
to query. Two files that differ only by case is not allowed.
27
* Case-insensitive: This is the file-system used by very old Windows versions
28
and is rarely encountered "in the wild". Two files that differ only by
29
case is not allowed and the case used to create a file is not preserved.
31
As can be implied by the above descriptions, only the first two are considered
32
relevant to a modern Bazaar.
34
For more details, including use cases, please see
35
http://bazaar-vcs.org/CasePreservingWorkingTreeUseCases
37
Handling these file-systems
38
---------------------------
40
The fundamental problem handling these file-systems is that the user may
41
specify a file name or inventory item with an "incorrect" case - where
42
"incorrect" simply means different than what is stored - from the user's
43
point-of-view, the filename is still correct, as it can be used to open, edit
46
The approach Bazaar takes is to "fixup" each of the command-line arguments
47
which refer to a filename or an inventory item - where "fixup" means to
48
adjust the case specified by the user so it exactly matches an existing item.
50
There are two places this match can be performed against - the file-system
51
and the Bazaar inventory. When looking at a case-insensitive file-system, it
52
is impossible to have 2 names that differ only by case, so there is no
53
ambiguity. The inventory doesn't have the same rules, but it is expected that
54
projects which wish to work with Windows would, by convention, avoid filenames
55
that differ only by case.
57
The rules for such fixups turn out to be quite simple:
59
* If an argument refers to an existing inventory item, we fixup the argument
60
using the inventory. This is, basically, all commands that take a filename
61
or directory argument *other* than 'add' and in some cases 'mv'
63
* If an argument refers to an existing filename for the creation of an
64
inventory item (eg, add), then the case of the existing file on the disk
65
will be used. However, Bazaar must still check the inventory to prevent
66
accidentally creating 2 inventory items that differ only by case.
68
* If an argument results in the creation of a *new* filename (eg, a move
69
destination), the argument will be used as specified. Bzr will create
70
a file and inventory item that exactly matches the case specified (although
71
as above, care must be taken to avoid creating two inventory items that
74
Implementation of support for these file-systems
75
------------------------------------------------
77
From the description above, it can be seen the implementation is fairly
78
simple and need not intrude on the internals of Bazaar too much; most of
79
the time it is simply converting a string specified by the user to the
80
"canonical" form as stored in either the inventory or filesystem. These
81
boil down to the following new API functions:
83
* osutils.canonical_relpath() - like osutils.relpath() but adjust the case
84
of the result to match any existing items.
86
* Tree.get_canonical_inventory_path - somewhat like Tree.get_symlink_target(),
87
Tree.get_file_by_path() etc; returns a name with the case adjusted to match
88
existing inventory items.
90
* osutils.canonical_relpaths() and Tree.get_canonical_inventory_paths() - like
91
the 'singular' versions above, but accept and return sequences and therefore
92
offer more optimization opportunities when working with multiple names.
94
The only complication is the requirement that Bazaar not allow the creation
95
of items that differ only by case on such file-systems. For this requirement,
96
case-insensitive and cicp file-systems can be treated the same. The
97
'case_sensitive' attribute on a MutableTree is used to control this behaviour.
added
removed
doc/developers/case-insensitive-file-systems.txt
