1
.. -*- mode: indented-text; compile-command: "make -C doc" -*-
9
See also various low-level TODOs in the source code. Try looking in
10
the list archive or on gmane.org for previous discussion of these
13
These are classified by approximate size: an hour or less, a day or
14
less, and several days or more.
20
* Merging add of a new file clashing with an existing file doesn't
21
work; add gets an error that it's already versioned and the merge
24
* Merge should ignore the destination's working directory, otherwise
25
we get an error about the statcache when pulling from a remote
28
* Add of a file that was present in the base revision should put back
31
* Not sure I'm happy with needing to pass a root id to EmptyTree;
32
comparing anything against an EmptyTree with no root should have the
35
* Handle diff of files which do not have a trailing newline; probably
36
requires patching difflib to get it exactly right, or otherwise
37
calling out to GNU diff.
39
* Should be able to copy files between branches to preserve their
40
file-id (and perhaps eventually parentage.)
42
* -r option should take a revision-id as well as a revno.
44
* allow ``bzr st -r 300`` to show a summary of changes since then.
46
* ``bzr info`` should count only people with distinct email addresses as
47
different committers. (Or perhaps only distinct userids?)
49
* On Windows, command-line arguments should be `glob-expanded`__,
50
because the shell doesn't do this. However, there are probably some
51
commands where this shouldn't be done, such as 'bzr ignore', because
52
we want to accept globs.
54
* ``bzr ignore`` command that just adds a line to the ``.bzrignore`` file
55
and makes it versioned. Fix this to break symlinks.
57
* Any useful sanity checks in 'bzr ignore'? Perhaps give a warning if
58
they try to add a single file which is already versioned, or if they
59
add a pattern which already exists, or if it looks like they gave an
62
__ http://mail.python.org/pipermail/python-list/2001-April/037847.html
64
* Separate read and write version checks?
66
* ``bzr status DIR`` should give status on all files under that
69
* ``bzr log DIR`` should give changes to any files within DIR.
71
* ``bzr inventory -r REV`` and perhaps unify this with ``bzr ls``,
72
giving options to display ids, types, etc.
74
* Split BzrError into various more specific subclasses for different
75
errors people might want to catch.
77
* If the export destination ends in '.tar', '.tar.gz', etc then create
78
a tarball instead of a directory. (Need to actually make a
79
temporary directory and then tar that up.)
81
http://www.gelato.unsw.edu.au/archives/git/0504/2194.html
83
* RemoteBranch could maintain a cache either in memory or on disk. We
84
know more than an external cache might about which files are
85
immutable and which can vary. On the other hand, it's much simpler
86
to just use an external proxy cache.
88
Perhaps ~/.bzr/http-cache. Baz has a fairly simple cache under
89
~/.arch-cache, containing revision information encoded almost as a
90
bunch of archives. Perhaps we could simply store full paths.
92
* Maybe also store directories in the statcache so that we can quickly
93
identify that they still exist.
95
* Diff should show timestamps; for files from the working directory we
96
can use the file itself; for files from a revision we should use the
97
commit time of the revision.
99
* Perhaps split command infrastructure from the actual command
102
* Cleaner support for negative boolean options like --no-recurse.
104
* Statcache should possibly map all file paths to / separators
106
* quotefn doubles all backslashes on Windows; this is probably not the
107
best thing to do. What would be a better way to safely represent
108
filenames? Perhaps we could doublequote things containing spaces,
109
on the principle that filenames containing quotes are unlikely?
110
Nice for humans; less good for machine parsing.
112
* Patches should probably use only forward slashes, even on Windows,
113
otherwise Unix patch can't apply them. (?)
115
* Branch.update_revisions() inefficiently fetches revisions from the
116
remote server twice; once to find out what text and inventory they
117
need and then again to actually get the thing. This is a bit
120
One complicating factor here is that we don't really want to have
121
revisions present in the revision-store until all their constituent
122
parts are also stored.
124
The basic problem is that RemoteBranch.get_revision() and similar
125
methods return object, but what we really want is the raw XML, which
126
can be popped into our own store. That needs to be refactored.
128
* ``bzr status FOO`` where foo is ignored should say so.
130
* ``bzr mkdir A...`` should just create and add A.
132
* Guard against repeatedly merging any particular patch.
134
* More options for diff:
136
- diff two revisions of the same tree
138
- diff two different branches, optionally at different revisions
140
- diff a particular file in another tree against the corresponding
141
version in this tree (which should be the default if the second
142
parameter is a tree root)
144
- diff everything under a particular directory, in any of the above
147
- diff two files inside the same tree, even if they have different
150
- and, of course, tests for all this
152
* stat-cache update is too slow for some reason - why is Python making
153
a lot of futex calls?
155
* ``bzr add`` with no arguments should probably be the same as ``bzr add .``
162
* Merge revert patch.
164
* ``bzr mv`` that does either rename or move as in Unix.
166
* More efficient diff of only selected files. We should be able to
167
just get the id for the selected files, look up their location and
168
diff just those files. No need to traverse the entire inventories.
170
* ``bzr status DIR`` or ``bzr diff DIR`` should report on all changes
171
under that directory.
173
* Fix up Inventory objects to represent root object as an entry.
175
* Don't convert entire entry from ElementTree to an object when it is
176
read in, but rather wait until the program actually wants to know
179
* Extract changes from one revision to the next to a text form
180
suitable for transmission over email.
184
- Selected-file commit
186
- Impossible selected-file commit: adding things in non-versioned
187
directories, crossing renames, etc.
189
* Write a reproducible benchmark, perhaps importing various kernel versions.
191
* Directly import diffs! It seems a bit redundant to need to rescan
192
the directory to work out what files diff added/deleted/changed when
193
all the information is there in the diff in the first place.
194
Getting the exact behaviour for added/deleted subdirectories etc
197
At the very least we could run diffstat over the diff, or perhaps
198
read the status output from patch. Just knowing which files might
199
be modified would be enough to guide the add and commit.
201
Given this we might be able to import patches at 1/second or better.
203
* Get branch over http.
205
* Pull pure updates over http.
207
* revfile compression.
209
* Split inventory into per-directory files.
211
* Fix ignore file parsing:
213
- fnmatch is not the same as unix patterns
215
- perhaps add extended globs from rsh/rsync
217
- perhaps a pattern that matches only directories or non-directories
219
* Consider using Python logging library as well as/instead of
222
* Commands should give some progress indication by default.
224
- But quieten this with ``--silent``.
226
* Change to using gettext message localization.
228
* Make a clearer separation between internal and external bzrlib
229
interfaces. Make internal interfaces use protected names. Write at
230
least some documentation for those APIs, probably as docstrings.
232
Consider using ZopeInterface definitions for the external interface;
233
I think these are already used in PyBaz. They allow automatic
234
checking of the interface but may be unfamiliar to general Python
235
developers, so I'm not really keen.
237
* Commands to dump out all command help into a manpage or HTML file or
240
* Handle symlinks in the working directory; at the very least it
241
should be possible for them to be present and ignored/unknown
242
without causing assertion failures.
244
Eventually symlinks should be versioned.
246
* Allow init in a subdirectory to create a nested repository, but only
247
if the subdirectory is not already versioned. Perhaps also require
248
a ``--nested`` to protect against confusion.
252
* More test framework:
254
- Class that describes the state of a working tree so we can just
257
* There are too many methods on Branch() that really manipulate the
258
WorkingTree. They should be moved across.
260
Also there are some methods which are duplicated on Tree and
261
Inventory objects, and it should be made more clear which ones are
262
proxies and which ones behave differently, and how.
264
* Try using XSLT to add some formatting to REST-generated HTML. Or
265
maybe write a small Python program that specifies a header and foot
266
for the pages and calls into the docutils libraries.
268
* --format=xml for log, status and other commands.
270
* Attempting to explicitly add a file that's already added should give
271
a warning; however there should be no warning for directories (since
272
we scan for new children) or files encountered in a directory that's
275
* Better handling of possible collisions on case-losing filesystems;
276
make sure a single file does not get added twice under different
279
* Clean up XML inventory:
281
- Use nesting rather than parent_id pointers.
283
- Hold the ElementTree in memory in the Inventory object and work
284
directly on that, rather than converting into Python objects every
285
time it is read in. Probably still exposoe it through some kind of
286
object interface though, but perhaps that should just be a proxy
289
- Less special cases for the root directory.
291
* Perhaps inventories should remember the revision in which each file
292
was last changed, as well as its current state? This is a bit
293
redundant but might often be interested to know.
295
* stat cache should perhaps only stat files as necessary, rather than
296
doing them all up-front. On the other hand, that disallows the
297
opimization of stating them in inode order.
299
* It'd be nice to pipeline multiple HTTP requests. Often we can
300
predict what will be wanted in future: all revisions, or all texts
301
in a particular revision, etc.
303
urlgrabber's docs say they are working on batched downloads; we
304
could perhaps ride on that or just create a background thread (ew).
306
* Paranoid mode where we never trust SHA-1 matches.
308
* Don't commit if there are no changes unless forced.
310
* --dry-run mode for commit? (Or maybe just run with
311
check-command=false?)
313
* Generally, be a bit more verbose unless --silent is specified.
315
* Function that finds all changes to files under a given directory;
316
perhaps log should use this if a directory is given.
318
* XML attributes might have trouble with filenames containing \n and
319
\r. Do we really want to support this? I think perhaps not.
321
* Remember execute bits, so that exports will work OK.
323
* Unify smart_add and plain Branch.add(); perhaps smart_add should
324
just build a list of files to add and pass that to the regular add
327
* Function to list a directory, saying in which revision each file was
328
last modified. Useful for web and gui interfaces, and slow to
329
compute one file at a time.
331
* unittest is standard, but the results are kind of ugly; would be
332
nice to make it cleaner.
334
* Check locking is correct during merge-related operations.
336
* Perhaps attempts to get locks should timeout after some period of
337
time, or at least display a progress message.
339
* Split out upgrade functionality from check command into a separate
342
* Don't pass around command classes but rather pass objects. This'd
343
make it cleaner to construct objects wrapping external commands.
345
* Track all merged-in revisions in a versioned add-only metafile.
351
* Generate annotations from current file relative to previous
354
- Is it necessary to store any kind of annotation where data was
357
* Update revfile_ format and make it active:
359
- Texts should be identified by something keyed on the revision, not
360
an individual text-id. This is much more useful for annotate I
361
think; we want to map back to the revision that last changed it.
363
- Access revfile revisions through the Tree/Store classes.
365
- Check them from check commands.
369
.. _revfile: revfile.html
371
* Hooks for pre-commit, post-commit, etc.
373
Consider the security implications; probably should not enable hooks
374
for remotely-fetched branches by default.
376
* Pre-commit check. If this hook is defined, it needs to be handled
377
specially: create a temporary directory containing the tree as it
378
will be after the commit. This means excluding any ignored/unknown
379
files, and respecting selective commits. Run the pre-commit check
380
(e.g. compile and run test suite) in there.
382
Possibly this should be done by splitting the commit function into
383
several parts (under a single interface). It is already rather
384
large. Decomposition:
386
- find tree modifications and prepare in-memory inventory
388
- export that inventory to a temporary directory
390
- run the test in that temporary directory
392
- if that succeeded, continue to actually finish the commit
394
What should be done with the text of modified files while this is
395
underway? I don't think we want to count on holding them in memory
396
and we can't trust the working files to stay in one place so I
397
suppose we need to move them into the text store, or otherwise into
398
a temporary directory.
400
If the commit does not actually complete, we would rather the
401
content was not left behind in the stores.
405
* GUI (maybe in Python GTK+?)
407
* C library interface
409
* Expansion of $Id$ keywords within working files. Perhaps do this in
410
exports first as a simpler case because then we don't need to deal
411
with removing the tags on the way back in.
added
removed
TODO
