|
krh@1
|
1 |
- keep history of installed packages/journal of package transaction,
|
|
krh@1
|
2 |
so we can roll back to yesterday, or see what got installed in the
|
|
krh@1
|
3 |
latest yum update.
|
|
krh@1
|
4 |
|
|
krh@1
|
5 |
- we build a cache of the currently installed set to service
|
|
krh@1
|
6 |
dependency inquiries fast:
|
|
krh@1
|
7 |
|
|
krh@1
|
8 |
map from property to pkg (as hash) providing it
|
|
krh@1
|
9 |
map from property to pkgs requiring it
|
|
krh@1
|
10 |
map from pkg name to manifest
|
|
krh@1
|
11 |
map from string to string pool index
|
|
krh@1
|
12 |
|
|
krh@1
|
13 |
no implicit provides? not even pkgname?
|
|
krh@1
|
14 |
|
|
krh@1
|
15 |
- properties are strings, stored in a string table
|
|
krh@1
|
16 |
|
|
krh@1
|
17 |
- on disk maps are binary files of (string table index, hash) pairs
|
|
krh@1
|
18 |
|
|
krh@1
|
19 |
- at run time, we mmap the map, and keep changes in memory in a splay
|
|
krh@1
|
20 |
tree or similar. if searching the splay tree fails we punt to the
|
|
krh@1
|
21 |
mmap. once the transaction is done, we merge the map and the splay
|
|
krh@1
|
22 |
tree and write it back out.
|
|
krh@1
|
23 |
|
|
krh@1
|
24 |
- the on-disk string pool is sorted and we keep a list of indices into
|
|
krh@1
|
25 |
the string pool in sorted order so we can bsearch the list with a
|
|
krh@1
|
26 |
string to get its string pool index. maybe a hash table is better,
|
|
krh@1
|
27 |
less I/O as we will expect to find the string within the block we
|
|
krh@1
|
28 |
look up with the hash function.
|
|
krh@1
|
29 |
|
|
krh@5
|
30 |
- represent all files as a breadth first traversal of the tree of all
|
|
krh@5
|
31 |
files. each entry has its name (string pool index), the number of
|
|
krh@5
|
32 |
immediate children, total number of children, and owning package.
|
|
krh@5
|
33 |
for files both these numbers are zero. a file is identified by its
|
|
krh@5
|
34 |
index in this flattened tree.
|
|
krh@5
|
35 |
|
|
krh@5
|
36 |
to get the file name from an index, we search through the list. by
|
|
krh@5
|
37 |
summing up the number of children, we know when to skip a directory
|
|
krh@5
|
38 |
and when to descend into one. as we go we accumulate the path
|
|
krh@5
|
39 |
elements.
|
|
krh@5
|
40 |
|
|
krh@5
|
41 |
hmm, dropping number of immediate children and using a sentinel drops
|
|
krh@5
|
42 |
a word from every entry.
|
|
krh@5
|
43 |
|
|
krh@1
|
44 |
- signed pkgs
|
|
krh@8
|
45 |
|
|
krh@8
|
46 |
- gzip repository of look-aside pkg xml files somehow?
|
|
krh@8
|
47 |
|
|
krh@8
|
48 |
- transactions, proper recovery, make sure we don't poop our package
|
|
krh@8
|
49 |
database (no more rm /var/lib/rpm/__cache*).
|
|
krh@8
|
50 |
|
|
krh@8
|
51 |
- no external dependencies, forget about bdb, sqlite. It's *simple*
|
|
krh@8
|
52 |
and we need to control the on-disk format for these tools.
|
|
krh@8
|
53 |
|
|
krh@18
|
54 |
- diff from one package set to another answers: "what changed in
|
|
krh@18
|
55 |
rawhide between since yesterday?"
|
|
krh@18
|
56 |
|
|
krh@18
|
57 |
- rewrite qsort and bsearch that doesn't require global context var
|
|
krh@18
|
58 |
and can output a map describing the permutaion.
|
|
krh@18
|
59 |
|
|
krh@18
|
60 |
- use hash table for package and property lists so we only store
|
|
krh@18
|
61 |
unique lists (like for string pool).
|
|
krh@40
|
62 |
|
|
krh@40
|
63 |
- use existing, running system as repo; eg
|
|
krh@40
|
64 |
|
|
krh@40
|
65 |
razor update razor://other-box.local evince
|
|
krh@40
|
66 |
|
|
krh@40
|
67 |
to pull eg the latest evince and dependencies from another box. We
|
|
krh@40
|
68 |
should be able to regenerate a rzr pkg from the system so we can
|
|
krh@40
|
69 |
reuse the signature from the originating repo.
|
|
krh@41
|
70 |
|
|
krh@41
|
71 |
- Ok, maybe the fastest package set merge method in the end is to use
|
|
krh@41
|
72 |
the razor_importer, but use a hash table for the properties. This
|
|
krh@41
|
73 |
way we can assign them unique IDs immediately (like tokenizing
|
|
krh@41
|
74 |
strings).
|