[hpke] Re: A few PRs for review

[Stephen Farrell <[e-mail]>]

On 24/06/2026 18:35, Richard Barnes wrote:
> Hi Stephen,
> 
> I think the question you raise was addressed by this PR from Martin which I
> merged on my own (seemed minor enough):
> 
> https://github.com/hpkewg/hpke/pull/66/changes

Ah sorry, missed that and yes it unpicks the nit:-)

> 
> Re NUL bytes - I think you are in the rough 

Absent a pile of others jumping in: me too.

> (chairs feel free to correct
> me).  In addition to the list discussion here, OpenSSL appears to be the
> only implementation that made this misinterpretation.  I would actually
> argue that the footgun points the other way, as others have on the list --
> having a NUL-clean API at lower layers removes possibilities for
> misinterpretation of inputs.  Every function that relies on strlen() is an
> out-of-bounds read in waiting.  If you really want to supply the psk_id
> from some string-like type, there are various types of escaping available.

Yep, there are arguments both ways. My conclusion is that for
anything commonly entered via CLI then allowing embedded NULs
is worse. YMMV of course.

S.

> --Richard
> 
> On Wed, Jun 24, 2026 at 7:18 AM Stephen Farrell <[e-mail]>
> wrote:
> 
>>
>> Hiya,
>>
>> Apologies, forgot to ask a nit-picky thing about this:
>>
>> The info, psk and psk_id fields are described as having
>> a default value of "", i.e. the empty string. If those
>> values can include embedded NUL bytes (I agree info and
>> psk ought support that), is an empty string the right
>> way to describe the default value?
>>
>> I guess the potential ambiguity is between a zero-length
>> default vs a default of length 1 containing a single NUL
>> byte.
>>
>> I doubt anyone would be confused by this but wondered if
>> there's a better way to describe the default having a
>> length of zero.
>>
>> Ta,
>> S.
>>
>> On 24/06/2026 14:28, Stephen Farrell wrote:
>>>
>>> Hiya,
>>>
>>> On 24/06/2026 04:20, Richard Barnes wrote:
>>>> Hi HPKE WG folks,
>>>>
>>>> As our main spec is strolling down the path to finalization, a few more
>>>> minor PRs have come along:
>>>>
>>>> https://github.com/hpkewg/hpke/pulls
>>>>
>>>> None are major, but a few have technical content, so review would be
>>>> appreciated.  If I don't hear anything by the end of the week, I'll go
>>>> ahead and merge them and cut a new draft for the IESG to review.
>>>
>>> WRT PR#70, I guess I'll be in the rough, but I'm still
>>> not keen on psk_id values with embedded NUL bytes as I
>>> think that's more a foot-gun than useful, e.g. if such
>>> a thing is supposed to be entered via a CLI or via some
>>> script or into a form, unexpected things may happen, or
>>> it may not be possible to enter the value. (I think the
>>> rest of the PR is likely fine.)
>>>
>>> Cheers,
>>> S.
>>>
>>>>
>>>> Thanks,
>>>> --Richard
>>>>
>>>>
>>>> _______________________________________________
>>>> hpke mailing list -- [e-mail]
>>>> To unsubscribe send an email to [e-mail]
>>>
>>>
>>> _______________________________________________
>>> hpke mailing list -- [e-mail]
>>> To unsubscribe send an email to [e-mail]
>>
>>
>