Re: [PATCH] Add 'bytevector-slice'.

[Maxime Devos <maximedevos@telenet.be>]

[-- Attachment #1.1.1: Type: text/plain, Size: 1444 bytes --]



On 14-01-2023 00:48, Ludovic Courtès wrote:
> Ah yes, that’s right, I misunderstood the comment.
> 
> In the example above, where we’re only dealing with slices, we could
> “skip” the parent (i.e., have each slice’s parent point to the “root” of
> the hierarchy), but I don’t think we can assume this to be the case
> generally.
> 
> Ludo’.

Why not?  I.e., where would things go wrong it the parent is "skipped" 
in other cases?  Something about GC I guess, but I don't follow what 
this "something" would be.

My guess is that you are thinking of the interaction with weak key-value 
maps, e.g. a map

    root-bytevector    -> stuff
    [other entries]

where the user might expect a slice of the root bytevector to keep the 
root itself intact -- which would be the case with your patch, but not 
with my "skip the parent" proposal.

I suppose there might be use cases for such things, in which case I 
wouldn't mind the original behaviour (though I'll have to look over 
Scheme-GNUnet code to verify no long chains are constructed), but if 
this is supposed to be supported, it should be documented. E.g.:

‘The returned slice keeps a reference to @var{bv} and not only to the 
underlying bytes of the bytevector.  Usually, this is of no importance, 
but this information is relevant when using GC data structures such as 
guardians and weak hash tables.’

Greetings,
Maxime.

[-- Attachment #1.1.2: OpenPGP public key --]
[-- Type: application/pgp-keys, Size: 929 bytes --]

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 236 bytes --]