1 Towards replacing rpm + yum (0.1):
5 - rpm file parser, create repo command
9 - versions in depsolving
13 - merge file lists when merging package sets
15 - import filelist.xml.gz too in yum importer
19 - figure out how to canonically represent empty string... ~0?
23 - keep history of installed packages/journal of package transaction,
24 so we can roll back to yesterday, or see what got installed in the
27 - gzip repository of look-aside pkg xml files somehow?
29 - transactions, proper recovery, make sure we don't poop our package
30 database (no more rm /var/lib/rpm/__cache*).
32 - rewrite qsort and bsearch that doesn't require global context var
33 and can output a map describing the permutaion.
35 - use hash table for package and property lists so we only store
36 unique lists (like for string pool).
38 - use existing, running system as repo; eg
40 razor update razor://other-box.local evince
42 to pull eg the latest evince and dependencies from another box. We
43 should be able to regenerate a rzr pkg from the system so we can
44 reuse the signature from the originating repo.
46 - Ok, maybe the fastest package set merge method in the end is to use
47 the razor_importer, but use a hash table for the properties. This
48 way we can assign them unique IDs immediately (like tokenizing
51 - test suite should be easy, just keep .repo files around and test
52 different type of upgrades that way (obsoletes, conflicts, file
53 conflicts, file/dir problems etc). Or maybe just keep a simple file
54 format ad use a custom importer to create the .repo files.
56 - pipelined download and install; download is network bound, install
57 is disk bound. Start installing once we have self-contained set of
58 packages. Install in reverse topo-sort order. Interruptible
59 installation; stops at nearest checkpoint.
61 - split out hash table code from importer, make the merger use just
66 do { ... } while (((e++)->name & RAZOR_ENTRY_LAST) == 0);
68 idiom for iteration of directories.
70 - version the sections in the file, put the element size in the header
71 so we can add stuff to elements in a backwards compatible way.
73 - overlay package sets? mount a read-only /usr over nfs or from the
74 virt-host and have a local package set overlaid over the read-only
75 one. shouldn't need new features in the core package set data
76 structure, but should be just conventions on top. we have the base
77 package set from the r/o system, the overlay set from the local
78 system and we can have an effective package set which is the merge
79 of everything from the overlay into the base set. the effective set
80 is easy to compute and we could do it on the fly or cache it. or
81 maybe the effective set is the on-disk representation and the
82 overlay can be computed when needed, we just keep a link back to the