[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Scheme-reports] 6.3.7 blobs

Andy Wingo scripsit:

> `make-blob' needs a fill argument.  (Filling it with random memory is a
> potential security vulnerability in many contexts.)

Security-conscious systems can fill it with truly random data, or all zeros,
or whatever they like.

> Someone mentioned (Peter?) that partial-bloc and partial-blob-copy!
> could be rolled into blob-copy and blob-copy!, and I agree very much.  I
> like the language about copy direction; that was very carefully done.

Thank you!  I think I got the language from memcpy.  Ticket filed.

> There is no read syntax for blobs; I think this is an error.  What about
> good ol' #vu8().

This makes sense if they are truly byte vectors, not so much if they are
truly blobs.  The difference being that a byte vector is a vector of
small integers, whereas a blob is something that can be *seen* as a vector
of small integers, or in a bunch of other ways (for which WG2 will provide
procedures).  See http://trac.sacrideo.us/wg/wiki/BlobAPI for a proposal.

> What would you call a sub-blob which shares state with another blob?

We don't have such state-sharing objects now, except lists.

John Cowan   cowan@x    http://ccil.org/~cowan
You cannot enter here.  Go back to the abyss prepared for you!  Go back!
Fall into the nothingness that awaits you and your Master.  Go! --Gandalf

Scheme-reports mailing list