1 |
On Mon, 2018-11-12 at 18:33 +0100, Ulrich Mueller wrote: |
2 |
> > > > > > On Mon, 12 Nov 2018, Michał Górny wrote: |
3 |
> > On Mon, 2018-11-12 at 17:51 +0100, Fabian Groffen wrote: |
4 |
> > > I'm wondering here, how much sense does it make to compress 2., 3. |
5 |
> > > and/or 4. if you compress the whole gpkg? I have the impression |
6 |
> > > compression on compression isn't beneficial here. Shouldn't just |
7 |
> > > compressing of the gpkg tar be sufficient? |
8 |
> > Please read the spec again. It explicitly says it's not compressed. |
9 |
> |
10 |
> Isn't that the wrong way around? The tar format contains a lot of |
11 |
> padding, so using uncompressed tar for the outer archive would be |
12 |
> somewhat wasteful. Why not leave the inner tar files uncompressed, but |
13 |
> compress the whole binpkg instead? |
14 |
|
15 |
Uncompressed tar is mostly suitable for random access. Compressed tar |
16 |
isn't suitable for random access at all. |
17 |
|
18 |
With uncompressed tar, it's trivial to access one of the members. With |
19 |
compressed tar, you always end up decompressing everything. |
20 |
|
21 |
With uncompressed tar, it's easy to rewrite the metadata (read: apply |
22 |
package updates) without updating the rest. With compressed tar, you'd |
23 |
have to recompress all the huge packages in order to apply updates. |
24 |
|
25 |
> Also, what would be wrong with ar? It's a standard POSIX tool, and |
26 |
> should be available everywhere. |
27 |
> |
28 |
|
29 |
The original post says what's wrong with ar. Please be more specific if |
30 |
you disagree with it. |
31 |
|
32 |
-- |
33 |
Best regards, |
34 |
Michał Górny |