Olivier Dion <[email protected]> writes:

> On Sun, 24 May 2026, Tomas Volf <[email protected]> wrote:
> [...]
>> This was quite interesting read.  The problem of generating identifiers
>> describe above seems non-trivial.  I probably do not have anything novel
>> to add to the debate, but two questions popped up in my head.
>>
>> 1. What do other Schemes do?
>> 2. Would incorporating source location (if available) into the hash be
>>    valid?
>
> I've tried to hash the syntax instead of the datum in psyntax.  Looking
> at the C code for hashing, hashing syntax seems to also include some
> metadata information.  But it did not yield a fix.

I am surprised to hear this.  Since even the warning message correctly
points both lines, the source location was obviously recorded.  So I do
not understand why it did not help with the hashing.

> I think what we probably want is a strong collision resistant hash.  A
> strong crypto hash would do the job for this. So the current core hash
> would not be impacted.
>
>
> I had some discussions related to this, i.e. crypto hash, in core guile,
> with Ludovic and Rob.  I think that having core support for some
> the crypto hashing makes sense, especially if the compiler needs it.
>
> Thoughts?

I agree with other posters, that using crypto hash as default for
generic hashing might not be ideal, but I am far from expert in the
area.

One extra thought though.  While I do understand why (hash) must have
bounded time (albeit whether the current maximal depth is sufficient is
a separate question), does the same hold even for generating
identifiers?  Maybe it does, I am not sure.

Tomas

-- 
There are only two hard things in Computer Science:
cache invalidation, naming things and off-by-one errors.

Reply via email to