Re: intrinsic vs extrinsic identifier: toward more robustness?

[Simon Tournier]

Hi,

On Mon, 06 Mar 2023 at 13:22, Maxime Devos <maximedevos@telenet.be> wrote:

>>> For git-fetch, the value of the 'commit' field is intrinsic (except when
>>> it's a tag instead).
>> 
>> No, it is imprecise.  The exception is *not* label tag as value for the
>> ’commit’ field but the exception is Git commit hash as value.
>
> Are you referring to the fact that currently, the 'commit' field usually 
> contains a tag name, and that it containing a commit is the exception?

Yes.

> If so, that doesn't contradict my claim.

There is no contradiction but imprecision.


> I do not see how making a list of all identifiers helps with robustness 
> -- you need the object the identifiers point to, not the identifier itself.

If you have the identifiers, you have a chance to find again the
content.  For example, in addition to NAR+SHA256, we could also store
Git+SHA1 or plain SHA256 or something else.  It would help in exploiting
other content-address systems.  For instance, SWH stores,

        "checksums": {
            "sha1": "3a48fbd0a69c7875dc18bd48a16da04d1512ed47",
            "sha1_git": "69cb76019a474330e99666f147ecb85e44de1ce6",
            "sha256": 
"e62e0f13f9025642a52f9fcb12ca0c31d5e05f78e97224f55b3d70d47c73b549"
        },

and maybe ’sha256_nar’ soon.  Somehow, we have a list of mirrors so why
not similarly having a list of intrinsic identifier.


> You are hashing the 'hello-2.12.1' directory

Thanks!  Having the noise too close and I missed the obvious. :-)


Cheers,
simon