[Pkg-bazaar-commits] ./bzr/unstable r282: - move all TODO items into ./TODO

Martin Pool mbp at sourcefrog.net
Fri Apr 10 07:51:40 UTC 2009 

------------------------------------------------------------
revno: 282
committer: Martin Pool <mbp at sourcefrog.net>
timestamp: Tue 2005-04-26 15:20:17 +1000
message:
  - move all TODO items into ./TODO
removed:
  doc/roadmap.txt
  doc/testing.txt
  doc/work-order.txt
added:
  TODO
modified:
  doc/index.txt
  doc/random.txt
-------------- next part --------------
=== added file 'TODO'
--- a/TODO	1970-01-01 00:00:00 +0000
+++ b/TODO	2005-04-26 05:20:17 +0000
@@ -0,0 +1,78 @@
+(See also various low-level TODOs in the source code.)
+
+Small things
+------------
+
+* Add of a file that was present in the base revision should put back
+  the previous file-id.
+
+* Handle diff of files which do not have a trailing newline; probably
+  requires patching difflib to get it exactly right, or otherwise
+  calling out to GNU diff.
+
+* Import ElementTree update patch.
+
+
+Medium things
+-------------
+
+* Faster diff/status.  
+
+  Status should be handled differently because it needs to report on
+  deleted and unknown files.  diff only needs to deal with versioned
+  files.
+
+* Merge Aaron's merge code.
+
+* Merge revert patch.
+
+* Turn on stat cache code, and add optimization about avoiding
+  dangerous cache entries.
+
+* mv command?
+
+* More efficient diff of only selected files.
+
+* Fix up Inventory objects to represent root object as an entry.
+
+* Don't convert entire entry from 
+
+* Extract changes from one revision to the next to a text form
+  suitable for transmission over email.
+
+* More test cases.
+
+* Write a reproducible benchmark, perhaps importing various kernel versions.
+
+* Change test.sh from Bourne shell into something in pure Python so
+  that it can be more portable.
+
+* Directly import diffs!  It seems a bit redundant to need to rescan
+  the directory to work out what files diff added/deleted/changed when
+  all the information is there in the diff in the first place.
+  Getting the exact behaviour for added/deleted subdirectories etc
+  might be hard.
+
+  At the very least we could run diffstat over the diff, or perhaps
+  read the status output from patch.  Just knowing which files might
+  be modified would be enough to guide the add and commit.
+  
+  Given this we might be able to import patches at 1/second or better.
+
+* Get branch over http.
+
+* Pull pure updates over http.
+
+* revfile compression.
+
+* Split inventory into per-directory files.
+
+
+Large things
+------------
+
+* Web interface
+
+* GUI (maybe in Python GTK+?)
+
+* C library interface
\ No newline at end of file

=== modified file 'doc/index.txt'
--- a/doc/index.txt	2005-03-28 00:14:04 +0000
+++ b/doc/index.txt	2005-04-26 05:20:17 +0000
@@ -180,18 +180,10 @@
 Project management and organization
 -----------------------------------
 
-* `Development news <news.html>`__
-
 * `Notes on how to get a VCS adopted <adoption.html>`__
 
-* `Testing plan <testing.html>`__ -- very sketchy.
-
 * `Thanks <thanks.html>`__ to various people
 
-* `Roadmap <roadmap.html>`__: High-level order for implementing features.
-
-* `<work-order.html>`__: current tasks.
-
 * `Extra commands <extra-commands.html>`__ for
   internal/developer/debugger use.
 

=== modified file 'doc/random.txt'
--- a/doc/random.txt	2005-04-12 06:20:20 +0000
+++ b/doc/random.txt	2005-04-26 05:20:17 +0000
@@ -235,26 +235,6 @@
 
 -----
 
-
-* Fix up diffs for files without a trailing newline
-
------
-
-* Directly import diffs!  It seems a bit redundant to need to rescan
-  the directory to work out what files diff added/deleted/changed when
-  all the information is there in the diff in the first place.
-  Getting the exact behaviour for added/deleted subdirectories etc
-  might be hard.
-
-  At the very least we could run diffstat over the diff, or perhaps
-  read the status output from patch.  Just knowing which files might
-  be modified would be enough to guide the add and commit.
-  
-  Given this we might be able to import patches at 1/second or better.
-
-
-----
-
 When sending patches by email (optionall) send each one as a separate
 mail, with a sequence number [%03d/%03d] at the start of the Subject.
 See mail from linus 2005-04-07.

=== removed file 'doc/roadmap.txt'
--- a/doc/roadmap.txt	2005-03-09 04:51:05 +0000
+++ b/doc/roadmap.txt	1970-01-01 00:00:00 +0000
@@ -1,61 +0,0 @@
-Roadmap for Bazaar-NG
-*********************
-
-(This document contains only things still-to-do, in in approximate
-order.  For a list of things already done, see the `development
-news`__.)
-
-__ news.html
-
-* get doctest going again now that we have predictable-order status
-  commands
-
-* note at top of .bzr.log
-
-* basic local versioning working: init, add, remove, mv, status, log,
-  diff, commit, revert
-
-  * revert command
-  * mv command
-  
-* some commands should work on selected files, if any are given:
-
-  * diff
-  * status
-  * log
-  * commit
-
-* work properly when invoked from a subdirectory
-
-* give a file-id to tree root?
-
-* use stat information to detect unchanged files without reading the
-  full text (or establish that this is not safe).
-
-* basic find command: find versioned, deleted, unknown, etc.
-
-* branch command
-
-* get a branch from an http server (remote Tree/Store/Branch proxy
-  classes)
-
-* write experimental weave algorithm in Python to see how it works?
-
-* update command: pull in changes from another branch that is a strict
-  superset of the destination
-
-* merge command: reconcile changes in this branch with those in
-  another.  punt on structural changes at first and only merge text,
-  then work out a nice way to resolve structure.
-
-* ``vc-bzr.el`` emacs support
-
-* *go self-hosting at about this point*, with parallel commits to baz
-
-* ignore patterns from ``.bzrignore``
-
-
-
-longer-term things:
-
-* Some kind of `compressed storage <compression.html>`_.
\ No newline at end of file

=== removed file 'doc/testing.txt'
--- a/doc/testing.txt	2005-03-09 04:51:05 +0000
+++ b/doc/testing.txt	1970-01-01 00:00:00 +0000
@@ -1,27 +0,0 @@
-***********************
-Test plan for Patchflow
-***********************
-
-This should be amenable to good automated testing.  Here are some
-ideas.
-
-* Corrupt the branch in various ways and make sure it is detected.
-
-  - To start with test all cases checked by check()
-
-* Commit.
-
-
-I'm using doctest_ for some API tests, and it looks pretty nice.  It
-adds some documentation of the pre/post-conditions of various
-operations.  It also encourages creating a clean and friendly Python
-API.
-
-Testing through the Python API does not cover all of the external
-shell interface, but it is much easier to write, because we don't need
-to deal with serializing everything to/from text.
-
-We may need additional tests to this.
-
-
-.. _doctest: http://docs.python.org/lib/module-doctest.html
\ No newline at end of file

=== removed file 'doc/work-order.txt'
--- a/doc/work-order.txt	2005-03-09 04:51:05 +0000
+++ b/doc/work-order.txt	1970-01-01 00:00:00 +0000
@@ -1,64 +0,0 @@
-* 
-
-
-
-* Build a past revision from history:
-
-  - Working; now rebuild only up to a particular point.
-  
-  - A generic test that can be applied to any test working directory
-    that all previous revisions can be successfully recreated.
-
-* Ignore-file support: finalize design and then implement it.
-
-* Command-line options: should be passed down as arguments to ``cmd``
-  functions. 
-
-  - ``commit`` message should be an option, not an argument.
-
-* Class representing a tree (or working directory or branch)
-
-  - Should this really be a class or is it better to just have a set
-    of global functions?  What data is held in memory other than the
-    directory name?
-
-  - List status and inventory sorted by file name.
-
-  - Add/move/delete files.
-
-* Store patches in a zip file?
-
-* When building diffs, show only the relevant tail of the two
-  filenames in the index line. 
-
-* Store patch headers in XML?  What parser/generator to use?  Maybe
-  tupletree-format, or Zope xmlpickle, or ElementTree_?
-
-.. _ElementTree: http://effbot.org/zone/element-index.htm
-
-* Make immutable files (e.g. recorded patches) read-only after they're
-  written.  This would be a good idea, except that it means people
-  need to use ``rm -rf`` to remove branches, which may be dangerous.
-
-
-Guiding principles
-------------------
-
-* In the first instance, just cache everything; don't worry about
-  recreating them from simpler components.  So keep a copy around of
-  all previous revisions, their inventory and manifest, etc.
-
-* Start writing a test framework into which new development can fit.
-
-* No need to handle move/rename or subdirectories yet.
-
-* Do try to add assertions against invalid states. 
-
-* Keep it simple in interface and implementation.
-
-
-
-
-.. Local variables:
-.. mode: rst
-.. End:

More information about the Pkg-bazaar-commits mailing list