7 * Lintian based automated rejects
8 - Have a set of lintian tags each package *must* not have. If it does
10 - If a tag is overriden by the maintainer, do not reject, but put it
11 into NEW. If the override is ok note that in a table and dont act on
12 it for any future uploads of this package anymore.
13 - possibly have two classes of tags. one for "shouldnt happen by
14 accident" and one "shouldnt happen". the first gets ignored from us
15 if overwritten in the package, the second only us can overwrite.
16 - its a suite option in dak, not active for all at once.
18 * Throw away all DD uploaded .debs. (Depend on "Lintian based automated
20 - Depends on the lintian based automated rejects, we do want correct
21 .debs uploaded. Ubuntu nicely demonstrated the quality loss if people
22 can just upload anything without even having to get one build
24 - Need a way to define a build-architecture for arch_all debs. Some of
25 them can only be build on certain architectures.
26 A control file header build-architecture: YXY should do it.
27 - its a suite option, not active for all at once.
28 - should have all buildd machines under dsa control
30 * Implement autosigning, see ftpmaster_autosigning on ftp-master host in text/.
32 * Check TODO.old and move still-valid/useful entries over here.
34 * need a testsuite _badly_
36 * database table "binaries" contains a column 'type TEXT NOT
37 NULL'. This should be made a FK on override_type, as it only contains
40 - sql query to do the db work for it:
41 ALTER TABLE binaries ADD COLUMN new_type INT4 REFERENCES override_type(id);
42 UPDATE BINARIES SET new_type = 7 WHERE type = 'deb';
43 UPDATE BINARIES SET new_type = 8 WHERE type = 'udeb';
44 ALTER TABLE binaries DROP COLUMN type;
45 ALTER TABLE binaries RENAME COLUMN new_type TO type;
47 - needs updateX.py written and then the rest of the code changed to deal
50 * Checkout SQL Alchemy and probably use that for our database layer.
52 * reject on > or < in a version constraint
54 * use pythonX.Y-tarfile to check orig.tar.gz timestamps too.
56 * the .dak stuff is fundamentally braindamaged for various reasons, it
57 should DIE. If we want to cache information - use a ("temporary")
58 database table and let p-a clean it up, e.g. like contents does.
60 * security global mail overrides should special case buildd stuff so
61 that buildds get ACCEPTED mails. Or maybe send them at
62 new-security-install time. That way upload-security doesn't grow
65 * debianqueued sucks. Reimplement in a sane way.
69 * 'dak process-new' allows you to edit the section and change the
70 component, but really shouldn't allow the component change.
72 * 'dak process-new' doesn't do the right thing with -2 and -1 uploads,
73 as you can end up with the .orig.tar.gz not in the pool or belonging
76 * 'dak process-new' doesn't trap signals from 'dak examine-package' properly
78 * 'dak queue-report' should footnote the actual notes, and also * the
79 versions with notes so we can see new versions since being noted...
81 * <neuro> the orig was in NEW, the changes that caused it to be NEW
82 were pulled out in -2, and we end up with no orig in the archive :(
88 * 'dak check-overrides' should remove the src-only override when a
89 binary+source override exists
91 * override checks sucks; it needs to track changes made by the
92 maintainer and pass them onto ftpmaster instead of warning the maintainer.
94 * Fix component handling in overrides
99 * 'dak cruft-report' could do with overrides
101 * cruft-report could spot "half-dropped" binaries. Like if a package
102 used to build A and B, but B is no longer built for half the
105 * cruft-report's NVIU check doesn't catch cases where source package
106 changed name, should check binaries too. [debian-devel@l.d.o,
109 * 'dak cruft-report' doesn't look at debian-installer but should.