> I built something like this as a toy for *SL a while back, a throw-away > prototype. I decided to set up a different property, not to use the procedure > property. Any reason for the preference?
It's not an important preference. Using a separate structure property would have worked too for this example. I wanted a small, quick example that showed why I'd like to introduce those two syntax parameters (struct-field-ref, struct-field-set!) for use with define-struct definitions. Without it, it's too tempting to try to use datum->syntax or try to fiddle with the struct-type, with messy results that either involve unhygienic operations or the structure type inspector. In fact, the undergrad I'm working with during the summer hit several of these snags. We tried using datum->syntax, and ran into a wall with silly hygiene issues. We ended up resorting to the unsafe-struct-ref approach in my previous email. If there's an alternative approach to the task that's clear and obvious, I'd love to change ours to use it. My proposed additional syntax parameters will allow those who care to be able to access the bindings from the internal use of make-struct-type, and I think that's ok. But I want to see if this is something that other people agree is a nice thing to have. _________________________________________________ For list-related administrative tasks: http://lists.racket-lang.org/listinfo/dev
