1 |
On Mon, 2018-11-26 at 20:17 +0100, Ulrich Mueller wrote: |
2 |
> > > > > > On Mon, 26 Nov 2018, Michał Górny wrote: |
3 |
> > Specification |
4 |
> > ============= |
5 |
> > The container format |
6 |
> > -------------------- |
7 |
> > The gpkg package container is an uncompressed .tar achive whose filename |
8 |
> > should use ``.gpkg.tar`` suffix. This archive contains the following |
9 |
> > members, all placed in a single directory whose name matches |
10 |
> > the basename of the package file, in order: |
11 |
> |
12 |
> I see no value in adding another directory indirection, and it will add |
13 |
> more overhead. |
14 |
|
15 |
Tar bomb is not a good design. Given tar padding, there will be no |
16 |
overhead unless the full path exceeds ustar limits which is unlikely. |
17 |
|
18 |
> Also, AFAICS the tar|tar pipeline that you previously |
19 |
> suggested won't work any more (or would at least require additional |
20 |
> arguments). |
21 |
|
22 |
I'm pretty sure the tar pipeline was actually written with account for |
23 |
the directory. |
24 |
|
25 |
> |
26 |
> > 1. The package identifier file ``gpkg-1.txt`` (required). |
27 |
> > [...] |
28 |
> > The implementations must include a package identifier file named |
29 |
> > ``gpkg-1.txt``. The filename includes package format version; |
30 |
> > implementations should reject packages which do not contain this file |
31 |
> > as unsupported format. |
32 |
> > The file can have any contents. Normally, it should be empty. |
33 |
> |
34 |
> If the file is empty, why is it named gpkg-1.txt (instead of just |
35 |
> gpkg-1)? |
36 |
> |
37 |
|
38 |
*shrug*. I can make it 'gpkg-1' or 'gpkg.1' or whatever you want ;-). |
39 |
|
40 |
-- |
41 |
Best regards, |
42 |
Michał Górny |