| 1 |
On Mon, 2021-09-13 at 12:08 +0200, Ulrich Mueller wrote: |
| 2 |
> > > > > > On Mon, 13 Sep 2021, Sheng Yu wrote: |
| 3 |
> |
| 4 |
> > -The archive contains a number of files, stored in a single |
| 5 |
> > directory |
| 6 |
> > -whose name should match the basename of the package file. However, |
| 7 |
> > -the implementation must be able to process an archive where |
| 8 |
> > -the directory name is mismatched. There should be no explicit |
| 9 |
> > archive |
| 10 |
> > -member entry for the directory. |
| 11 |
> > +The archive contains a number of files. All package-related files |
| 12 |
> > +should be stored in a single directory whose name matches the CPV |
| 13 |
> > of |
| 14 |
> > +the package file. However, the implementation must be able to |
| 15 |
> > process |
| 16 |
> > +an archive where the directory name is mismatched. There should be |
| 17 |
> > no |
| 18 |
> > +explicit archive member entry for the directory. |
| 19 |
> |
| 20 |
> I wonder about CPV here. That's ${CATEGORY}/${P} and contains a slash, |
| 21 |
> so it cannot be the name of a directory. Also, what about the package |
| 22 |
> revision? |
| 23 |
|
| 24 |
Please restore the previous wording. The GLEP deliberately did not |
| 25 |
enforce a specific filename because it's about internal format. |
| 26 |
|
| 27 |
> |
| 28 |
> > +6. The package manifest data file ``Manifest`` (required). |
| 29 |
> > + |
| 30 |
> > +7. A signature for the package Manifest file ``Manifest.sig`` |
| 31 |
> > + (optional). |
| 32 |
> |
| 33 |
> Given that the outer archive is uncompressed tar, every file will be |
| 34 |
> zero-padded to a full block which adds some amount of bloat. So, could |
| 35 |
> the signature be inlined in the Manifest file? That's also what GLEP |
| 36 |
> 74 |
| 37 |
> specifies. |
| 38 |
|
| 39 |
Using inline signature in Manifest makes sense. |
| 40 |
|
| 41 |
> |
| 42 |
> Also, IIRC one of the goals of the format was to allow partial |
| 43 |
> download |
| 44 |
> of metadata. That will only work if the Manifest file will be the |
| 45 |
> first |
| 46 |
> file in the archive (or at least appear before the image archive). |
| 47 |
|
| 48 |
I disagree. This is solved by having detached metadata signature -- you |
| 49 |
can do a partial fetch and verify the metadata directly. |
| 50 |
|
| 51 |
On the other hand, putting Manifest first would make it impossible to |
| 52 |
create the archive from data stream without using temporary files, |
| 53 |
effectively doubling the needed free space. Well, technically you could |
| 54 |
just reserve space and write Manifest later but that would strongly |
| 55 |
depend on the size of PGP signature and that's not something I'd feel |
| 56 |
comfortable relying on. |
| 57 |
|
| 58 |
-- |
| 59 |
Best regards, |
| 60 |
Michał Górny |
Archives