root/releases/pkgcore/0.2.10/dev-notes/TODO.rst @ masterdriverz%2540gentoo.org-20070309221744-6u30xw5rb8n13cjj

Revision masterdriverz%2540gentoo.org-20070309221744-6u30xw5rb8n13cjj, 3.0 KB (checked in by Charlie Shepherd <masterdriverz@…>, 22 months ago)

Whitespace fixes

Line 
1==========
2Rough TODO
3==========
4
5- rip out use.* code from pkgcore_checks.addons.UseAddon.__init__, and
6  generalize it into pkgcore.ebuild.repository
7
8- not hugely important, but... make a cpython version of SlottedDict from
9  pkgcore.util.obj; 3% reduction for full repo walk, thus not a real huge
10  concern atm.
11
12- userpriv for pebuild misbehaves..
13
14- http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/491285
15  check into, probably better then my curfty itersort; need to see how
16  well heapqu's nlargest pop behaves (looks funky)
17
18- look into converting MULTILIB_STRICT* crap over to a trigger
19
20- install-sources trigger
21
22- recreate verify-rdepends also
23
24- observer objects for reporting back events from merging/unmerging
25  cpython 'tee' is needed, contact harring for details.
26  basic form of it is in now, but need something more powerful for
27  parallelization
28  elog is bound to this also
29
30- Possibly convert to cpython:
31
32  - flat_hash.database._parse_data
33  - metadata.database._parse_data
34  - posixpath (os.path)
35
36- get the tree clean of direct /var/db/pkg access
37
38- vdb2 format (ask harring for details).
39
40- pkgcore.fs.ops.merge_contents; doesn't rewite the contents set when a file
41  it's mergeing is relying on symlinked directories for the full path; eg,
42  /usr/share/X11/xkb/compiled -> /var/blah, it records the former instead of
43  recording the true absolute path.
44
45- pmerge mods; [ --skip-set SET ] , [ --skip atom ], use similar restriction
46  to --replace to prefer vdb for matching atoms
47
48- refactor pkgcore.ebuild.cpv.ver_cmp usage to avoid full cpv parsing when
49  _cpv is in use;
50  'nuff said, look in pkgcore.ebuild.cpv.cpy_ver_cmp
51
52- testing of fakeroot integration
53
54  it was working back in the ebd branch days; things have changed since then
55  (heavily), enabling/disabling should work fine, but will need to take a look
56  at the contentset generation to ensure perms/gid leaks through correctly.
57
58- modify repository.prototype.tree.match to take an optional comparison
59
60  reasoning being that if we're just going to do a max, pass in the max so it
61  has the option of doing the initial sorting without passing through
62  visibility filters (which will trigger metadata lookups)
63
64- 'app bundles'.  Reliant on serious overhauling of deps to do 'locked deps',
65  but think of it as rpath based app stacks, a full apache stack compiled to
66  run from /opt/blah for example.
67
68- pkgcore.ebuild.gpgtree
69
70  derivative of pkgcore.ebuild.ebuild_repository, this overloads
71  ebuild_factory and eclass_cache so that gpg checks are done.
72  This requires some hackery, partially dependant on config.central changes
73  (see above).  Need a way to specify the trust ring to use, 'severity' level
74  (different class targets works for me).
75  Anyone who implements this deserves massive cookies.
76
77- pkgcore.ebuild.gpgprofile:
78  Same as above.
79
80- reintroduce locking of certain high level components using read/write;
81  mainly, use it as a way to block sync'ing a repo that's being used to build,
82  lock the vdb for updates, etc.
Note: See TracBrowser for help on using the browser.