« back to all changes in this revision
Viewing changes to TODO
- browse files at revision 286
- compare with another revision
- download diff
- download tarball
- view history from revision 286
- Committer: Martin Pool
- Date: 2005-04-26 10:37:53 UTC
- Revision ID: mbp@sourcefrog.net-20050426103753-e022129bd6afd29b
- New bzr whoami --email option
- files added:
- bzrlib/tests.py
- files removed:
- HACKING
- bzrlib/plugins
- bzrlib/selftest
- bzrlib/util
- bzrlib/util/effbot
- bzrlib/util/effbot/org
- contrib
- contrib/bash
- contrib/emacs
- contrib/zsh
- patches
- tools
- files renamed:
- bzrlib/util/elementtree/ => elementtree/
- bzrlib/util/urlgrabber/ => urlgrabber/
- files modified:
- .bzrignore
added
removed
TODO
1
.. -*- mode: indented-text; compile-command: "make -C doc" -*-
2
3
4
*******************
5
Things to do in bzr
6
*******************
7
8
9
See also various low-level TODOs in the source code. Try looking in
1
(See also various low-level TODOs in the source code. Try looking in
10
2
the list archive or on gmane.org for previous discussion of these
11
issues.
12
13
These are classified by approximate size: an hour or less, a day or
14
less, and several days or more.
15
3
issues, or )
16
4
17
5
Small things
18
6
------------
19
7
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
23
--tiny``::
24
25
2005-11-05 mbp Flowers for stephane
26
27
* print a message at the end of running the tests telling them tht the
28
test log and output exists but can be removed
29
30
* tests for running the commit editor, and fix problem of not passing in
31
multiple arguments
32
33
* Merging add of a new file clashing with an existing file doesn't
34
work; add gets an error that it's already versioned and the merge
35
aborts.
36
37
* Merge should ignore the destination's working directory, otherwise
38
we get an error about the statcache when pulling from a remote
39
branch.
40
41
8
* Add of a file that was present in the base revision should put back
42
9
the previous file-id.
43
10
44
* Not sure I'm happy with needing to pass a root id to EmptyTree;
45
comparing anything against an EmptyTree with no root should have the
46
same effect(?)
47
48
11
* Handle diff of files which do not have a trailing newline; probably
49
12
requires patching difflib to get it exactly right, or otherwise
50
13
calling out to GNU diff.
51
14
52
* Should be able to copy files between branches to preserve their
53
file-id (and perhaps eventually parentage.)
54
55
* -r option should take a revision-id as well as a revno.
56
57
* allow ``bzr st -r 300`` to show a summary of changes since then.
58
59
* ``bzr info`` should count only people with distinct email addresses as
60
different committers. (Or perhaps only distinct userids?)
61
62
* On Windows, command-line arguments should be `glob-expanded`__,
63
because the shell doesn't do this. However, there are probably some
64
commands where this shouldn't be done, such as 'bzr ignore', because
65
we want to accept globs.
66
67
* ``bzr ignore`` command that just adds a line to the ``.bzrignore`` file
68
and makes it versioned. Fix this to break symlinks.
69
70
* Any useful sanity checks in 'bzr ignore'? Perhaps give a warning if
71
they try to add a single file which is already versioned, or if they
72
add a pattern which already exists, or if it looks like they gave an
73
unquoted glob.
74
75
__ http://mail.python.org/pipermail/python-list/2001-April/037847.html
76
77
* Separate read and write version checks?
78
79
* ``bzr status DIR`` should give status on all files under that
80
directory.
81
82
* ``bzr log DIR`` should give changes to any files within DIR; at the
83
moment it only lists things which modify the specific named file
84
(and not its contents)
85
86
* ``bzr inventory -r REV`` and perhaps unify this with ``bzr ls``,
87
giving options to display ids, types, etc.
88
89
* Split BzrError into various more specific subclasses for different
90
errors people might want to catch.
91
92
* If the export destination ends in '.tar', '.tar.gz', etc then create
93
a tarball instead of a directory. (Need to actually make a
94
temporary directory and then tar that up.)
95
96
http://www.gelato.unsw.edu.au/archives/git/0504/2194.html
97
98
* RemoteBranch could maintain a cache either in memory or on disk. We
99
know more than an external cache might about which files are
100
immutable and which can vary. On the other hand, it's much simpler
101
to just use an external proxy cache.
102
103
Perhaps ~/.bzr/http-cache. Baz has a fairly simple cache under
104
~/.arch-cache, containing revision information encoded almost as a
105
bunch of archives. Perhaps we could simply store full paths.
106
107
* Maybe also store directories in the statcache so that we can quickly
108
identify that they still exist.
109
110
* Diff should show timestamps; for files from the working directory we
111
can use the file itself; for files from a revision we should use the
112
commit time of the revision.
113
114
* Perhaps split command infrastructure from the actual command
115
definitions.
116
117
* Cleaner support for negative boolean options like --no-recurse.
118
119
* Statcache should possibly map all file paths to / separators
120
121
* quotefn doubles all backslashes on Windows; this is probably not the
122
best thing to do. What would be a better way to safely represent
123
filenames? Perhaps we could doublequote things containing spaces,
124
on the principle that filenames containing quotes are unlikely?
125
Nice for humans; less good for machine parsing.
126
127
* Patches should probably use only forward slashes, even on Windows,
128
otherwise Unix patch can't apply them. (?)
129
130
* Branch.update_revisions() inefficiently fetches revisions from the
131
remote server twice; once to find out what text and inventory they
132
need and then again to actually get the thing. This is a bit
133
inefficient.
134
135
One complicating factor here is that we don't really want to have
136
revisions present in the revision-store until all their constituent
137
parts are also stored.
138
139
The basic problem is that RemoteBranch.get_revision() and similar
140
methods return object, but what we really want is the raw XML, which
141
can be popped into our own store. That needs to be refactored.
142
143
* ``bzr status FOO`` where foo is ignored should say so.
144
145
* ``bzr mkdir A...`` should just create and add A.
146
147
* Guard against repeatedly merging any particular patch.
148
149
* More options for diff:
150
151
- diff two revisions of the same tree
152
153
- diff two different branches, optionally at different revisions
154
155
- diff a particular file in another tree against the corresponding
156
version in this tree (which should be the default if the second
157
parameter is a tree root)
158
159
- diff everything under a particular directory, in any of the above
160
ways
161
162
- diff two files inside the same tree, even if they have different
163
ids
164
165
- and, of course, tests for all this
166
167
* ``bzr add`` with no arguments should probably be the same as ``bzr add .``
168
169
* Reproducible performance benchmark to measure whether performance is
170
getting better or worse.
171
172
* ``bzr log -m foo`` should perhaps error if nothing matches?
173
174
* ``bzr diff -r 30 -r 40 foo.c`` or ``bzr diff -r30..40 foo.c``
175
176
If diffing between two branches then we probably want two -r
177
options, since the revisions don't form a range that can be
178
evaluated on either one.
179
180
* bzr diff shouldn't diff binary files
181
182
* setup.py install when run from a bzr tree should freeze the tree
183
revision-id into the installed bzr.
184
185
* bzr script should trap ImportError and perhaps give a better error
186
message?
187
188
* revert after a merge should possibly remove all the BASE/THIS/OTHER
189
files to get you back to where you were.
190
191
* files that are added and then deleted are still reported as added
192
193
* stores should raise KeyError, not IndexError
194
195
* merging from a remote branch seems to sometimes raise errors not
196
present locally
197
198
* should be possible to give a related branch when pulling from a
199
remote branch to make things faster
200
201
* sometimes gives "conflicting add" even when the contents are in fact
202
the same???
203
15
* Import ElementTree update patch.
16
17
* Syntax should be "bzr export -r REV".
18
19
* "cat -rREV FILE"
204
20
205
21
Medium things
206
22
-------------
207
23
208
* ~/.bzr.log is not written anymore for some reason.
209
210
* merge should add all revision and inventory XML to the local store.
211
212
* check should give a warning for revisions that are named in the
213
chain but not actually present in the store.
214
215
* remove anything outside of the branch implementation that directly
216
accesses the stores.
217
218
* More efficient diff of only selected files. We should be able to
219
just get the id for the selected files, look up their location and
220
diff just those files. No need to traverse the entire inventories.
221
222
* ``bzr status DIR`` or ``bzr diff DIR`` should report on all changes
223
under that directory.
24
* Faster diff/status.
25
26
Status should be handled differently because it needs to report on
27
deleted and unknown files. diff only needs to deal with versioned
28
files.
29
30
* Merge Aaron's merge code.
31
32
* Merge revert patch.
33
34
* Turn on stat cache code, and add optimization about avoiding
35
dangerous cache entries.
36
37
* mv command?
38
39
* More efficient diff of only selected files.
224
40
225
41
* Fix up Inventory objects to represent root object as an entry.
226
42
227
* Don't convert entire entry from ElementTree to an object when it is
228
read in, but rather wait until the program actually wants to know
229
about that node.
43
* Don't convert entire entry from
230
44
231
45
* Extract changes from one revision to the next to a text form
232
46
suitable for transmission over email.
233
47
234
48
* More test cases.
235
49
236
- ``missing`` command
237
238
- Selected-file commit
239
240
- Impossible selected-file commit: adding things in non-versioned
241
directories, crossing renames, etc.
242
243
50
* Write a reproducible benchmark, perhaps importing various kernel versions.
244
51
52
* Change test.sh from Bourne shell into something in pure Python so
53
that it can be more portable.
54
245
55
* Directly import diffs! It seems a bit redundant to need to rescan
246
56
the directory to work out what files diff added/deleted/changed when
247
57
all the information is there in the diff in the first place.
251
61
At the very least we could run diffstat over the diff, or perhaps
252
62
read the status output from patch. Just knowing which files might
253
63
be modified would be enough to guide the add and commit.
254
64
255
65
Given this we might be able to import patches at 1/second or better.
256
66
67
* Get branch over http.
68
69
* Pull pure updates over http.
70
257
71
* revfile compression.
258
72
259
* Split inventory into per-directory files?
73
* Split inventory into per-directory files.
260
74
261
75
* Fix ignore file parsing:
262
76
266
80
267
81
- perhaps a pattern that matches only directories or non-directories
268
82
269
* Consider using Python logging library as well as/instead of
270
bzrlib.trace.
271
272
* Commands should give some progress indication by default.
273
274
- But quieten this with ``--silent``.
275
276
* Change to using gettext message localization.
277
278
* Make a clearer separation between internal and external bzrlib
279
interfaces. Make internal interfaces use protected names. Write at
280
least some documentation for those APIs, probably as docstrings.
281
282
Consider using ZopeInterface definitions for the external interface;
283
I think these are already used in PyBaz. They allow automatic
284
checking of the interface but may be unfamiliar to general Python
285
developers, so I'm not really keen.
286
287
* Commands to dump out all command help into a manpage or HTML file or
288
whatever.
289
290
* Handle symlinks in the working directory; at the very least it
291
should be possible for them to be present and ignored/unknown
292
without causing assertion failures.
293
294
Eventually symlinks should be versioned.
295
296
* Allow init in a subdirectory to create a nested repository, but only
297
if the subdirectory is not already versioned. Perhaps also require
298
a ``--nested`` to protect against confusion.
299
300
* Branch names?
301
302
* More test framework:
303
304
- Class that describes the state of a working tree so we can just
305
assert it's equal.
306
307
* There are too many methods on Branch() that really manipulate the
308
WorkingTree. They should be moved across.
309
310
Also there are some methods which are duplicated on Tree and
311
Inventory objects, and it should be made more clear which ones are
312
proxies and which ones behave differently, and how.
313
314
* Try using XSLT to add some formatting to REST-generated HTML. Or
315
maybe write a small Python program that specifies a header and foot
316
for the pages and calls into the docutils libraries.
317
318
* --format=xml for log, status and other commands.
319
320
* Attempting to explicitly add a file that's already added should give
321
a warning; however there should be no warning for directories (since
322
we scan for new children) or files encountered in a directory that's
323
being scanned.
324
325
* Better handling of possible collisions on case-losing filesystems;
326
make sure a single file does not get added twice under different
327
names.
328
329
* Clean up XML inventory:
330
331
- Use nesting rather than parent_id pointers.
332
333
- Hold the ElementTree in memory in the Inventory object and work
334
directly on that, rather than converting into Python objects every
335
time it is read in. Probably still exposoe it through some kind of
336
object interface though, but perhaps that should just be a proxy
337
for the elements.
338
339
- Less special cases for the root directory.
340
341
* Perhaps inventories should remember the revision in which each file
342
was last changed, as well as its current state? This is a bit
343
redundant but might often be interested to know.
344
345
* stat cache should perhaps only stat files as necessary, rather than
346
doing them all up-front. On the other hand, that disallows the
347
opimization of stating them in inode order.
348
349
* It'd be nice to pipeline multiple HTTP requests. Often we can
350
predict what will be wanted in future: all revisions, or all texts
351
in a particular revision, etc.
352
353
urlgrabber's docs say they are working on batched downloads; we
354
could perhaps ride on that or just create a background thread (ew).
355
356
* Paranoid mode where we never trust SHA-1 matches.
357
358
* Don't commit if there are no changes unless forced.
359
360
* --dry-run mode for commit? (Or maybe just run with
361
check-command=false?)
362
363
* Generally, be a bit more verbose unless --silent is specified.
364
365
* Function that finds all changes to files under a given directory;
366
perhaps log should use this if a directory is given.
367
368
* XML attributes might have trouble with filenames containing \n and
369
\r. Do we really want to support this? I think perhaps not.
370
371
* Remember execute bits, so that exports will work OK.
372
373
* Unify smart_add and plain Branch.add(); perhaps smart_add should
374
just build a list of files to add and pass that to the regular add
375
function.
376
377
* Function to list a directory, saying in which revision each file was
378
last modified. Useful for web and gui interfaces, and slow to
379
compute one file at a time.
380
381
This will be done when we track file texts by referring to the
382
version that created them.
383
384
* Check locking is correct during merge-related operations.
385
386
* Perhaps attempts to get locks should timeout after some period of
387
time, or at least display a progress message.
388
389
* Split out upgrade functionality from check command into a separate
390
``bzr upgrade``.
391
392
* Don't pass around command classes but rather pass objects. This'd
393
make it cleaner to construct objects wrapping external commands.
394
395
* Track all merged-in revisions in a versioned add-only metafile.
396
397
* ``pull --clobber`` should discard any local changes not present
398
remotely. Not generally what you want, but possibly useful when
399
you're just mirroring another branch and want to keep tracking it
400
even when they e.g. uncommit or make similar non-forward movements.
401
Also for push I suppose. Clobber may not be the best name, maybe
402
``--destroy``?
403
404
* ``uncommit`` command that removes a revision from the end of the
405
revision-history; just doing this is enough to remove the commit,
406
and a new commit will automatically be made against the
407
predecessor. This can be repeated.
408
409
It only makes sense to delete from the tail of history, not from the
410
end.
411
412
The revision, its inventory and texts remain floating in the store.
413
We should perhaps add the revision to a list of removed-commits, so
414
that it can be restored or at least accounted for when checking
415
consistency. This file would not be versioned, and probably should
416
not propagate when branched.
417
418
If we track merged revisions then we need to update this list too.
419
If the list is stored in a weave it's easy (implicit): the version
420
of the list can remain but it won't be referenced anymore. It's
421
probably best to just store this list in a weave in the first place
422
and be done.
83
* Expansion of $Id$ tags within working files. Perhaps do this in
84
exports first as a simpler case because then we don't need to deal
85
with removing the tags on the way back in.
423
86
424
87
425
88
Large things
426
89
------------
427
90
428
* Generate annotations from current file relative to previous
429
annotations.
430
431
- Is it necessary to store any kind of annotation where data was
432
deleted?
433
434
* Update revfile_ format and make it active:
435
436
- Texts should be identified by something keyed on the revision, not
437
an individual text-id. This is much more useful for annotate I
438
think; we want to map back to the revision that last changed it.
439
440
- Access revfile revisions through the Tree/Store classes.
441
442
- Check them from check commands.
443
444
- Store annotations.
445
446
.. _revfile: revfile.html
447
448
* Hooks for pre-commit, post-commit, etc.
449
450
Consider the security implications; probably should not enable hooks
451
for remotely-fetched branches by default.
452
453
* Pre-commit check. If this hook is defined, it needs to be handled
454
specially: create a temporary directory containing the tree as it
455
will be after the commit. This means excluding any ignored/unknown
456
files, and respecting selective commits. Run the pre-commit check
457
(e.g. compile and run test suite) in there.
458
459
Possibly this should be done by splitting the commit function into
460
several parts (under a single interface). It is already rather
461
large. Decomposition:
462
463
- find tree modifications and prepare in-memory inventory
464
465
- export that inventory to a temporary directory
466
467
- run the test in that temporary directory
468
469
- if that succeeded, continue to actually finish the commit
470
471
What should be done with the text of modified files while this is
472
underway? I don't think we want to count on holding them in memory
473
and we can't trust the working files to stay in one place so I
474
suppose we need to move them into the text store, or otherwise into
475
a temporary directory.
476
477
If the commit does not actually complete, we would rather the
478
content was not left behind in the stores.
479
480
91
* Web interface
481
92
482
93
* GUI (maybe in Python GTK+?)
483
94
484
95
* C library interface
485
486
* Expansion of $Id$ keywords within working files. Perhaps do this in
487
exports first as a simpler case because then we don't need to deal
488
with removing the tags on the way back in.
489
490
* ``bzr find``
Loggerhead is a web-based interface for Breezy
Version: 2.0.1