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

*To*: Sebastian Egner <sebastian.egner@xxxxxxxxxxx>*Subject*: Re: hash-by-identity guarantees of change consistency?*From*: Panu Kalliokoski <atehwa@xxxxxxxx>*Date*: Fri, 21 Apr 2006 15:17:25 +0300*Cc*: srfi-69@xxxxxxxxxxxxxxxxx*Delivered-to*: srfi-69@xxxxxxxxxxxxxxxxx*In-reply-to*: <OF6FB0B803.DBE07944-ONC1257156.00267393-C1257156.002A2FD4@xxxxxxxxxxx>*References*: <hG07BD.A.ETY.hlqREB@rotkohl> <OF6FB0B803.DBE07944-ONC1257156.00267393-C1257156.002A2FD4@xxxxxxxxxxx>*User-agent*: Mutt/1.5.9i

On Thu, Apr 20, 2006 at 09:39:45AM +0200, Sebastian Egner wrote: > > Any ideas / suggestions on how to best approach this? > The usual approach is to turn a blind eye to the problem, > because most approaches that really solve it (like keeping > track of modifications) have inacceptable performance. > My approach would be to clearly state a general precondition > on hash tables: Hash tables assume that the values stored in > the table are not modified. That is already stated, the problem being just that it severes the usefulness of eq? hash tables. > Another solution, e.g. in the Map type of the LEDA library > for C++, is to use the /pointer/ to the heap-allocated object > as a hash value. This is indeed the most efficient hash > function you can imagine. Unfortunately, this imposes the > severe constraint that the underlying garbage collector must > support this, e.g. by not moving objects in the heap. ... and you can't do it in standard RnRS Scheme, AFAIK. Panu -- personal contact: panu.kalliokoski@xxxxxxxxxxx, +35841 5323835 technical contact: atehwa@xxxxxx, http://www.iki.fi/atehwa/ PGP fingerprint: 0EA5 9D33 6590 FFD4 921C 5A5F BE85 08F1 3169 70EC

**Follow-Ups**:**Re: hash-by-identity guarantees of change consistency?***From:*Per Bothner

**References**:**Re: hash-by-identity guarantees of change consistency?***From:*Sebastian Egner

- Prev by Date:
**Re: hash-by-identity guarantees of change consistency?** - Next by Date:
**Re: hash-by-identity guarantees of change consistency?** - Previous by thread:
**Re: hash-by-identity guarantees of change consistency?** - Next by thread:
**Re: hash-by-identity guarantees of change consistency?** - Index(es):