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
* ``bzr status`` should show some description of patches that are merged but
21
not yet committed. Preferably in a very compact format so that they
22
just fit on a single line; this could also be used for say ``log
25
2005-11-05 mbp Flowers for stephane
27
* Fix tests so that import errors caused by modules don't produce false reports
28
that the tests themselves don't exist.
30
* Fix tests so that one test failure doesn't prevent other tests from running
32
* print a message at the end of running the tests telling them that the
33
test log and output exists but can be removed
35
* tests for running the commit editor, and fix problem of not passing in
38
* Merging add of a new file clashing with an existing file doesn't
39
work; add gets an error that it's already versioned and the merge
42
* Merge should ignore the destination's working directory, otherwise
43
we get an error about the statcache when pulling from a remote
46
* Add of a file that was present in the base revision should put back
49
* Not sure I'm happy with needing to pass a root id to EmptyTree;
50
comparing anything against an EmptyTree with no root should have the
53
* Handle diff of files which do not have a trailing newline; probably
54
requires patching difflib to get it exactly right, or otherwise
55
calling out to GNU diff.
57
* Should be able to copy files between branches to preserve their
58
file-id (and perhaps eventually parentage.)
60
* -r option should take a revision-id as well as a revno.
62
* allow ``bzr st -r 300`` to show a summary of changes since then.
64
* ``bzr info`` should count only people with distinct email addresses as
65
different committers. (Or perhaps only distinct userids?)
67
* On Windows, command-line arguments should be `glob-expanded`__,
68
because the shell doesn't do this. However, there are probably some
69
commands where this shouldn't be done, such as 'bzr ignore', because
70
we want to accept globs.
72
* ``bzr ignore`` command that just adds a line to the ``.bzrignore`` file
73
and makes it versioned. Fix this to break symlinks.
75
* Any useful sanity checks in 'bzr ignore'? Perhaps give a warning if
76
they try to add a single file which is already versioned, or if they
77
add a pattern which already exists, or if it looks like they gave an
80
__ http://mail.python.org/pipermail/python-list/2001-April/037847.html
82
* Separate read and write version checks?
84
* ``bzr status DIR`` should give status on all files under that
87
* ``bzr log DIR`` should give changes to any files within DIR; at the
88
moment it only lists things which modify the specific named file
89
(and not its contents)
91
* ``bzr inventory -r REV`` and perhaps unify this with ``bzr ls``,
92
giving options to display ids, types, etc.
94
* Split BzrError into various more specific subclasses for different
95
errors people might want to catch.
97
* If the export destination ends in '.tar', '.tar.gz', etc then create
98
a tarball instead of a directory. (Need to actually make a
99
temporary directory and then tar that up.)
101
http://www.gelato.unsw.edu.au/archives/git/0504/2194.html
103
* RemoteBranch could maintain a cache either in memory or on disk. We
104
know more than an external cache might about which files are
105
immutable and which can vary. On the other hand, it's much simpler
106
to just use an external proxy cache.
108
Perhaps ~/.bzr/http-cache. Baz has a fairly simple cache under
109
~/.arch-cache, containing revision information encoded almost as a
110
bunch of archives. Perhaps we could simply store full paths.
112
* Maybe also store directories in the statcache so that we can quickly
113
identify that they still exist.
115
* Diff should show timestamps; for files from the working directory we
116
can use the file itself; for files from a revision we should use the
117
commit time of the revision.
119
* Perhaps split command infrastructure from the actual command
122
* Cleaner support for negative boolean options like --no-recurse.
124
* Statcache should possibly map all file paths to / separators
126
* quotefn doubles all backslashes on Windows; this is probably not the
127
best thing to do. What would be a better way to safely represent
128
filenames? Perhaps we could doublequote things containing spaces,
129
on the principle that filenames containing quotes are unlikely?
130
Nice for humans; less good for machine parsing.
132
* Patches should probably use only forward slashes, even on Windows,
133
otherwise Unix patch can't apply them. (?)
135
* Branch.update_revisions() inefficiently fetches revisions from the
136
remote server twice; once to find out what text and inventory they
137
need and then again to actually get the thing. This is a bit
140
One complicating factor here is that we don't really want to have
141
revisions present in the revision-store until all their constituent
142
parts are also stored.
144
The basic problem is that RemoteBranch.get_revision() and similar
145
methods return object, but what we really want is the raw XML, which
146
can be popped into our own store. That needs to be refactored.
148
* ``bzr status FOO`` where foo is ignored should say so.
150
* ``bzr mkdir A...`` should just create and add A.
152
* Guard against repeatedly merging any particular patch.
154
* More options for diff:
156
- diff two revisions of the same tree
158
- diff two different branches, optionally at different revisions
160
- diff a particular file in another tree against the corresponding
161
version in this tree (which should be the default if the second
162
parameter is a tree root)
164
- diff everything under a particular directory, in any of the above
167
- diff two files inside the same tree, even if they have different
170
- and, of course, tests for all this
172
* ``bzr add`` with no arguments should probably be the same as ``bzr add .``
174
* Reproducible performance benchmark to measure whether performance is
175
getting better or worse.
177
* ``bzr log -m foo`` should perhaps error if nothing matches?
179
* ``bzr diff -r 30 -r 40 foo.c`` or ``bzr diff -r30..40 foo.c``
181
If diffing between two branches then we probably want two -r
182
options, since the revisions don't form a range that can be
183
evaluated on either one.
185
* bzr diff shouldn't diff binary files
187
* setup.py install when run from a bzr tree should freeze the tree
188
revision-id into the installed bzr.
190
* bzr script should trap ImportError and perhaps give a better error
193
* revert after a merge should possibly remove all the BASE/THIS/OTHER
194
files to get you back to where you were.
196
* files that are added and then deleted are still reported as added
198
* stores should raise KeyError, not IndexError
200
* merging from a remote branch seems to sometimes raise errors not
203
* should be possible to give a related branch when pulling from a
204
remote branch to make things faster
206
* sometimes gives "conflicting add" even when the contents are in fact
209
* BZRDIR should be in branch.py not __init__.py.
211
* ``status `` should accept a -r option to show changes relative to a revision,
217
* ~/.bzr.log is not written anymore for some reason.
219
* merge should add all revision and inventory XML to the local store.
221
* check should give a warning for revisions that are named in the
222
chain but not actually present in the store.
224
* remove anything outside of the branch implementation that directly
227
* More efficient diff of only selected files. We should be able to
228
just get the id for the selected files, look up their location and
229
diff just those files. No need to traverse the entire inventories.
231
* ``bzr status DIR`` or ``bzr diff DIR`` should report on all changes
232
under that directory.
234
* Fix up Inventory objects to represent root object as an entry.
236
* Don't convert entire entry from ElementTree to an object when it is
237
read in, but rather wait until the program actually wants to know
240
* Extract changes from one revision to the next to a text form
241
suitable for transmission over email.
245
- ``missing`` command
247
- Selected-file commit
249
- Impossible selected-file commit: adding things in non-versioned
250
directories, crossing renames, etc.
252
* Write a reproducible benchmark, perhaps importing various kernel versions.
254
* Directly import diffs! It seems a bit redundant to need to rescan
255
the directory to work out what files diff added/deleted/changed when
256
all the information is there in the diff in the first place.
257
Getting the exact behaviour for added/deleted subdirectories etc
260
At the very least we could run diffstat over the diff, or perhaps
261
read the status output from patch. Just knowing which files might
262
be modified would be enough to guide the add and commit.
264
Given this we might be able to import patches at 1/second or better.
266
* revfile compression.
268
* Split inventory into per-directory files?
270
* Fix ignore file parsing:
272
- fnmatch is not the same as unix patterns
274
- perhaps add extended globs from rsh/rsync
276
- perhaps a pattern that matches only directories or non-directories
278
* Consider using Python logging library as well as/instead of
281
* Commands should give some progress indication by default.
283
- But quieten this with ``--silent``.
285
* Change to using gettext message localization.
287
* Make a clearer separation between internal and external bzrlib
288
interfaces. Make internal interfaces use protected names. Write at
289
least some documentation for those APIs, probably as docstrings.
291
Consider using ZopeInterface definitions for the external interface;
292
I think these are already used in PyBaz. They allow automatic
293
checking of the interface but may be unfamiliar to general Python
294
developers, so I'm not really keen.
296
* Commands to dump out all command help into a manpage or HTML file or
299
* Handle symlinks in the working directory; at the very least it
300
should be possible for them to be present and ignored/unknown
301
without causing assertion failures.
303
Eventually symlinks should be versioned.
305
* Allow init in a subdirectory to create a nested repository, but only
306
if the subdirectory is not already versioned. Perhaps also require
307
a ``--nested`` to protect against confusion.
311
* More test framework:
313
- Class that describes the state of a working tree so we can just
316
* There are too many methods on Branch() that really manipulate the
317
WorkingTree. They should be moved across.
319
Also there are some methods which are duplicated on Tree and
320
Inventory objects, and it should be made more clear which ones are
321
proxies and which ones behave differently, and how.
323
* Try using XSLT to add some formatting to REST-generated HTML. Or
324
maybe write a small Python program that specifies a header and foot
325
for the pages and calls into the docutils libraries.
327
* --format=xml for log, status and other commands.
329
* Attempting to explicitly add a file that's already added should give
330
a warning; however there should be no warning for directories (since
331
we scan for new children) or files encountered in a directory that's
334
* Better handling of possible collisions on case-losing filesystems;
335
make sure a single file does not get added twice under different
338
* Clean up XML inventory:
340
- Use nesting rather than parent_id pointers.
342
- Hold the ElementTree in memory in the Inventory object and work
343
directly on that, rather than converting into Python objects every
344
time it is read in. Probably still exposoe it through some kind of
345
object interface though, but perhaps that should just be a proxy
348
- Less special cases for the root directory.
350
* Perhaps inventories should remember the revision in which each file
351
was last changed, as well as its current state? This is a bit
352
redundant but might often be interested to know.
354
* stat cache should perhaps only stat files as necessary, rather than
355
doing them all up-front. On the other hand, that disallows the
356
opimization of stating them in inode order.
358
* It'd be nice to pipeline multiple HTTP requests. Often we can
359
predict what will be wanted in future: all revisions, or all texts
360
in a particular revision, etc.
362
urlgrabber's docs say they are working on batched downloads; we
363
could perhaps ride on that or just create a background thread (ew).
365
* Paranoid mode where we never trust SHA-1 matches.
367
* Don't commit if there are no changes unless forced.
369
* --dry-run mode for commit? (Or maybe just run with
370
check-command=false?)
372
* Generally, be a bit more verbose unless --silent is specified.
374
* Function that finds all changes to files under a given directory;
375
perhaps log should use this if a directory is given.
377
* XML attributes might have trouble with filenames containing \n and
378
\r. Do we really want to support this? I think perhaps not.
380
* Remember execute bits, so that exports will work OK.
382
* Unify smart_add and plain Branch.add(); perhaps smart_add should
383
just build a list of files to add and pass that to the regular add
386
* Function to list a directory, saying in which revision each file was
387
last modified. Useful for web and gui interfaces, and slow to
388
compute one file at a time.
390
This will be done when we track file texts by referring to the
391
version that created them.
393
* Check locking is correct during merge-related operations.
395
* Perhaps attempts to get locks should timeout after some period of
396
time, or at least display a progress message.
398
* Split out upgrade functionality from check command into a separate
401
* Don't pass around command classes but rather pass objects. This'd
402
make it cleaner to construct objects wrapping external commands.
404
* Track all merged-in revisions in a versioned add-only metafile.
406
* ``pull --clobber`` should discard any local changes not present
407
remotely. Not generally what you want, but possibly useful when
408
you're just mirroring another branch and want to keep tracking it
409
even when they e.g. uncommit or make similar non-forward movements.
410
Also for push I suppose. Clobber may not be the best name, maybe
413
* ``uncommit`` command that removes a revision from the end of the
414
revision-history; just doing this is enough to remove the commit,
415
and a new commit will automatically be made against the
416
predecessor. This can be repeated.
418
It only makes sense to delete from the tail of history, not from the
421
The revision, its inventory and texts remain floating in the store.
422
We should perhaps add the revision to a list of removed-commits, so
423
that it can be restored or at least accounted for when checking
424
consistency. This file would not be versioned, and probably should
425
not propagate when branched.
427
If we track merged revisions then we need to update this list too.
428
If the list is stored in a weave it's easy (implicit): the version
429
of the list can remain but it won't be referenced anymore. It's
430
probably best to just store this list in a weave in the first place
437
* Generate annotations from current file relative to previous
440
- Is it necessary to store any kind of annotation where data was
443
* Update revfile_ format and make it active:
445
- Texts should be identified by something keyed on the revision, not
446
an individual text-id. This is much more useful for annotate I
447
think; we want to map back to the revision that last changed it.
449
- Access revfile revisions through the Tree/Store classes.
451
- Check them from check commands.
455
.. _revfile: revfile.html
457
* Hooks for pre-commit, post-commit, etc.
459
Consider the security implications; probably should not enable hooks
460
for remotely-fetched branches by default.
462
* Pre-commit check. If this hook is defined, it needs to be handled
463
specially: create a temporary directory containing the tree as it
464
will be after the commit. This means excluding any ignored/unknown
465
files, and respecting selective commits. Run the pre-commit check
466
(e.g. compile and run test suite) in there.
468
Possibly this should be done by splitting the commit function into
469
several parts (under a single interface). It is already rather
470
large. Decomposition:
472
- find tree modifications and prepare in-memory inventory
474
- export that inventory to a temporary directory
476
- run the test in that temporary directory
478
- if that succeeded, continue to actually finish the commit
480
What should be done with the text of modified files while this is
481
underway? I don't think we want to count on holding them in memory
482
and we can't trust the working files to stay in one place so I
483
suppose we need to move them into the text store, or otherwise into
484
a temporary directory.
486
If the commit does not actually complete, we would rather the
487
content was not left behind in the stores.
491
* GUI (maybe in Python GTK+?)
493
* C library interface
495
* Expansion of $Id$ keywords within working files. Perhaps do this in
496
exports first as a simpler case because then we don't need to deal
497
with removing the tags on the way back in.
added
removed
TODO
